diff --git a/.gitattributes b/.gitattributes index a6344aac8c09253b3b630fb776ae94478aa0275b..52373fe24473b1aa44333d318f578ae6bf04b49b 100644 --- a/.gitattributes +++ b/.gitattributes @@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text *.zip filter=lfs diff=lfs merge=lfs -text *.zst filter=lfs diff=lfs merge=lfs -text *tfevents* filter=lfs diff=lfs merge=lfs -text +tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/mlc-chat-config.json b/mlc-chat-config.json new file mode 100644 index 0000000000000000000000000000000000000000..f3dd8df0d97f49a73019f134a56da0a66faa3135 --- /dev/null +++ b/mlc-chat-config.json @@ -0,0 +1,83 @@ +{ + "version": "0.1.0", + "model_type": "gemma2", + "quantization": "q0f16", + "model_config": { + "hidden_size": 4608, + "intermediate_size": 36864, + "attention_bias": false, + "num_attention_heads": 32, + "num_key_value_heads": 16, + "head_dim": 128, + "num_hidden_layers": 46, + "rms_norm_eps": 1e-06, + "vocab_size": 256000, + "hidden_activation": "gelu_pytorch_tanh", + "position_embedding_base": 10000.0, + "context_window_size": 4096, + "prefill_chunk_size": 2048, + "tensor_parallel_shards": 1, + "max_batch_size": 80, + "attn_logit_softcapping": 50.0, + "final_logit_softcapping": 30.0, + "query_pre_attn_scalar": 144, + "sliding_window": 4096 + }, + "vocab_size": 256000, + "context_window_size": 4096, + "sliding_window_size": -1, + "prefill_chunk_size": 2048, + "attention_sink_size": -1, + "tensor_parallel_shards": 1, + "temperature": 1.0, + "presence_penalty": 0.0, + "frequency_penalty": 0.0, + "repetition_penalty": 1.0, + "top_p": 1.0, + "tokenizer_files": [ + "tokenizer.model", + "tokenizer.json", + "tokenizer_config.json" + ], + "tokenizer_info": { + "token_postproc_method": "byte_fallback", + "prepend_space_in_encode": false, + "strip_space_in_decode": false + }, + "conv_template": { + "name": "gemma_instruction", + "system_template": "{system_message}", + "system_message": "", + "system_prefix_token_ids": [ + 2 + ], + "add_role_after_system_message": true, + "roles": { + "user": "user", + "assistant": "model" + }, + "role_templates": { + "user": "{user_message}", + "assistant": "{assistant_message}", + "tool": "{tool_message}" + }, + "messages": [], + "seps": [ + "\n" + ], + "role_content_sep": "\n", + "role_empty_sep": "\n", + "stop_str": [ + "" + ], + "stop_token_ids": [ + 1, + 107 + ], + "function_string": "", + "use_function_calling": false + }, + "pad_token_id": 0, + "bos_token_id": 2, + "eos_token_id": 1 +} \ No newline at end of file diff --git a/ndarray-cache.json b/ndarray-cache.json new file mode 100644 index 0000000000000000000000000000000000000000..86e97266f22a48040d84544dfc1bcad78c9fb57e --- /dev/null +++ b/ndarray-cache.json @@ -0,0 +1,5382 @@ +{ + "metadata": { + "ParamSize": 370, + "ParamBytes": 54454256640.0, + "BitsPerParam": 14.686606395927337 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 2359296000, + "records": [ + { + "name": "model.embed_tokens.weight", + "shape": [ + 256000, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2359296000, + "byteOffset": 0 + } + ], + "md5sum": "dc36e04174ceac3d7d23463b5d0d5181" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.0.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "4f0e853727e47613806ffa9c4c482ff2" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.0.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "21fc9a247c7e078674e21f55791089e3" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.0.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "7b6c03f580fe4119384785ef11ca4387" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.0.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "0a4c23fe136646c69827d68a51d94fff" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.1.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "6b485dacf260f4fa5690ad5f9c056cdb" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.1.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "72d93a6fa7ce8ef45f7e92291de659a7" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.1.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "faebe2c0b094d27995cea1fb726b162a" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.1.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "b7486ddadec0d8c5b2cdfbd015070d84" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.2.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "a29efe9166112e96288dbca63318d224" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.2.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "d6c07888f84c4d7b9207c8af8812c8c8" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.10.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "cc641d45e6a9f09da1abeb52f359a23a" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.11.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "f04105260c6b8b5cea9222a47bc2eed2" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.11.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "c88be0bc9d4893fcd7ec78bc4190a4b5" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.11.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "96bc024d0029894355a306ed1ca2c93a" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.11.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "eea52126cae035d5862119a6130f4b95" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.12.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "c35f2943c085189db06e8fa5fdefcac2" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.12.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "e2f728f041fea40538f97df2a74b6a61" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.12.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "53977f25830b1d83063393cf6e06e11b" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.12.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "1d7c1b03ed5c22fd10e07f222ce40343" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.13.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "2a9e7c74a08e7d4a740ef5cc1d16e1e4" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.13.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "60f7265756557068a21f4a74a00e5065" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.13.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "2f08ac47e48250c696cf73e100725deb" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.13.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "0f343ca21f549dfe46fc541759f83444" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.14.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "3c573cb4a07c2dbf1ba61c2bbdbe7da5" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.14.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "f953f21956b50b4bc9e2d1b584883141" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.14.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "ce383a398f02af492b0bc7b8eef97aaa" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.14.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "06b9fa6b136e7cdc917af3a48d5f696a" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.15.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "fb73d3cf5113e78ecc793631c659d8d4" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.15.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "6f0b10d5db85c567deb5d966759af0cd" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.10.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "fe0ae2bbb18c6e0179cbe97833c5cd6d" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.10.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "e8bc0f6ed5fdff9f85e6ba0284f66b01" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.10.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "8c2719dbc5fd7513a948c92de098e927" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.6.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "2380c8423932cc6b542b0574b83200c2" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.6.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "196e535102df782e482ca3147cc1c127" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.7.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "d047b7e99b3dd4be9c7c7faa6798d5d0" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.7.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "9d4564c6ca52ada14b4630ca4be84c3b" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.7.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "874c1a06468f373519867ffcefb370cc" + }, + { + "dataPath": "params_shard_38.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.7.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "e4d9c2ee9c5e0103a51777f02ce8eb0c" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.8.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "0cae21a4e33d59c1ec4539eb13568d37" + }, + { + "dataPath": "params_shard_40.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.8.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "7f50d6caf5e55da261262ef68a0d793d" + }, + { + "dataPath": "params_shard_41.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.8.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "a76dce23ab85b669e0f2928bbbf0867c" + }, + { + "dataPath": "params_shard_42.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.8.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "c8480558f1f90dad78885e50b97d6c48" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.9.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "1b885b2200798b7cfd1879572c59352e" + }, + { + "dataPath": "params_shard_44.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.9.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "77f9abf4964a372bc69d2d3796e2589b" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.9.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "c59b3117c62bda38b842653d6ceb375d" + }, + { + "dataPath": "params_shard_46.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.9.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "02d1044d98322c2a291a82cebe0e7857" + }, + { + "dataPath": "params_shard_47.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.15.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "906edd6dcbe6278695ca0e848c69e583" + }, + { + "dataPath": "params_shard_48.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.15.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "5e9554edf16196c92858917406ea400a" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.16.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "08942848f84502d3a9872e335ca3fec2" + }, + { + "dataPath": "params_shard_50.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.16.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "dc4db64c4ae40646745079d1a6c45b97" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.16.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "4f77afd133ed31ce97d662c303318f8b" + }, + { + "dataPath": "params_shard_52.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.16.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "29aaabfda8a389edea789a19b10efd58" + }, + { + "dataPath": "params_shard_53.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.17.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "876710b458576180c5bdd37723a12cff" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.17.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "edf947b8a8028c4e7278f496746a3d69" + }, + { + "dataPath": "params_shard_55.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.17.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "65a0be861a8dc6fdcf2e7c2353637e3f" + }, + { + "dataPath": "params_shard_56.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.17.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "f7c84cdbd8fcbb6624d9bb17252cd1e2" + }, + { + "dataPath": "params_shard_57.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.18.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "8a1f6df7d173c58177b6ff49e44b6f60" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.18.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "2f7c9d760ed57a0a963a975739bceba8" + }, + { + "dataPath": "params_shard_59.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.18.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "50c66672f6e350aa32133092f67e2325" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.18.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "81798ecdb577f945f643e4e41d674c2c" + }, + { + "dataPath": "params_shard_61.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.19.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "ce6aaf2fd04fdf82e550b9c3b3b0a986" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.19.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "697236882302cf7618eb6a231156af19" + }, + { + "dataPath": "params_shard_63.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.19.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "ccd5b13f6da68b770003bae7acd90081" + }, + { + "dataPath": "params_shard_64.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.19.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "87f8c1eb1c0635f3c6a45fdedfdab1c0" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.20.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "0839d1f19482eaea044426792f2abad2" + }, + { + "dataPath": "params_shard_66.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.20.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "b1f469791c6f561d7ab5378b3c573b68" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.20.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "ad5e4eacb1548fdc4b2fa0c884cd2b5b" + }, + { + "dataPath": "params_shard_68.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.20.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "056e0b725496bb0a1dee4375e1f60900" + }, + { + "dataPath": "params_shard_69.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.21.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "00e7aea26a88ccb6e27fc43ee0b596c6" + }, + { + "dataPath": "params_shard_70.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.21.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "632e2a9805f500bd67b9c8b46075ec96" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.21.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "a19d25593152024db2cdb58653521249" + }, + { + "dataPath": "params_shard_72.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.21.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "e1521585a47d49646f32f00cfeb28f13" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.22.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "f7b5cf942e2ad27cb31b3d86bb934740" + }, + { + "dataPath": "params_shard_74.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.22.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "bc20a1b9ce0d50def914fa3283740755" + }, + { + "dataPath": "params_shard_75.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.22.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "181b6f0f812244ed28f845443394c2c3" + }, + { + "dataPath": "params_shard_76.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.22.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "ccc07720caf4878f8ab28a00247574eb" + }, + { + "dataPath": "params_shard_77.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.23.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "6e880ad3a769101bf5d42b3b2bd30428" + }, + { + "dataPath": "params_shard_78.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.23.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "ae6d7000e8302c78c975881680afb875" + }, + { + "dataPath": "params_shard_79.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.23.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "f0f8c0b3791755ffcb3e2bab3629b32e" + }, + { + "dataPath": "params_shard_80.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.2.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "c6a1621ea851755cc1c676c40ba8396c" + }, + { + "dataPath": "params_shard_81.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.2.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "8ed9f65c0b647b2794a456560a2a22e6" + }, + { + "dataPath": "params_shard_82.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.3.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "58e2d4f9838fd2ddab6ce01a60a1c025" + }, + { + "dataPath": "params_shard_83.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.3.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "a6b70b5bafedb22e5e6355741eb50fa3" + }, + { + "dataPath": "params_shard_84.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.3.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "3093327e75aa189e4c1d62964f796cfb" + }, + { + "dataPath": "params_shard_85.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.3.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "fc87c5e22378d6e9057ffeee62b139ff" + }, + { + "dataPath": "params_shard_86.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.4.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "2d7be78a34ea444d640f252063277c3e" + }, + { + "dataPath": "params_shard_87.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.4.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "7d86ef778cd55b79b4df50d4c34e8deb" + }, + { + "dataPath": "params_shard_88.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.4.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "3a65d46a8dca4899ac05f0a2e382c3dd" + }, + { + "dataPath": "params_shard_89.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.4.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "a60ab57c6d1ce3da7030047c7b491e70" + }, + { + "dataPath": "params_shard_90.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.5.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "f27c1dc563fd7050411badbbd667e1c5" + }, + { + "dataPath": "params_shard_91.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.5.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "73b24ad60ab31462446ce6602fa1a44e" + }, + { + "dataPath": "params_shard_92.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.5.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "e831fef8e15fdf98ea271c1f80748476" + }, + { + "dataPath": "params_shard_93.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.5.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "f1d2b99785dc024c66e656607780a26d" + }, + { + "dataPath": "params_shard_94.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.6.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "471f65297ef8e62ff152af3fbbcb29af" + }, + { + "dataPath": "params_shard_95.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.6.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "9f02c1eff8688c1446ed046a4e94d294" + }, + { + "dataPath": "params_shard_96.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.23.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "299eec1f29838a1619f651b8ac2c2d2a" + }, + { + "dataPath": "params_shard_97.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.24.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "48ee8aa14b6bedbf74ea626ccc6b66a9" + }, + { + "dataPath": "params_shard_98.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.24.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "1d8c70a2fc2694d951a55a7614f2ee8f" + }, + { + "dataPath": "params_shard_99.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.24.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "f2c72b1c908d4fe944fdab3339eaf781" + }, + { + "dataPath": "params_shard_100.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.24.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "15c479aacfa31f1227b563789686152d" + }, + { + "dataPath": "params_shard_101.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.25.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "24de14c07f42cb9d7142f9085470c02a" + }, + { + "dataPath": "params_shard_102.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.25.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "561db5e2c4e9b20a00497d459e65fb0f" + }, + { + "dataPath": "params_shard_103.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.25.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "1a8a3aec61c5bef5e6e7942fc2ba31b0" + }, + { + "dataPath": "params_shard_104.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.25.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "d4b800d82bb141c8a3f0531e7dc8362f" + }, + { + "dataPath": "params_shard_105.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.26.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "5f1d312f5a8aff7f31d45f49dba6b47a" + }, + { + "dataPath": "params_shard_106.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.26.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "a68eb722f2505b2775c3b27591c531d3" + }, + { + "dataPath": "params_shard_107.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.26.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "2e82c8779b7d5498268a303a1fe33c42" + }, + { + "dataPath": "params_shard_108.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.26.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "d196249f35fa960330bf61163eaf14e6" + }, + { + "dataPath": "params_shard_109.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.27.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "4915ebff7e72c5278a55d0525ae23bcf" + }, + { + "dataPath": "params_shard_110.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.27.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "804a60e642f3dad08c7fcafd2637a225" + }, + { + "dataPath": "params_shard_111.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.27.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "55b2212a596e3f28e9ef49f4e1d58f85" + }, + { + "dataPath": "params_shard_112.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.27.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "04ce72e3c3f6f7aa3771e2c9941f9092" + }, + { + "dataPath": "params_shard_113.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.28.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "99482aea32b7ab638410a48d1e389f61" + }, + { + "dataPath": "params_shard_114.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.28.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "b0e38e30df4691211db0690a94039541" + }, + { + "dataPath": "params_shard_115.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.28.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "628902c3c3c8e71990dccd546e73079b" + }, + { + "dataPath": "params_shard_116.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.28.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "0249fa2584c4e66e6217d6b4f2aa673c" + }, + { + "dataPath": "params_shard_117.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.29.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "7e7ded135f939f346bdf016d8457982e" + }, + { + "dataPath": "params_shard_118.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.29.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "d33319299cd0ce8206cb31d488a0d57f" + }, + { + "dataPath": "params_shard_119.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.29.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "69703b0f869124b81aa722f5455e240a" + }, + { + "dataPath": "params_shard_120.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.29.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "535ca2d671d58d7416a81e4843331b09" + }, + { + "dataPath": "params_shard_121.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.30.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "1b2c8133ed20173bab22eedac23723c2" + }, + { + "dataPath": "params_shard_122.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.30.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "b47ca8b0c1b07aa6165cb3a6f9aee351" + }, + { + "dataPath": "params_shard_123.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.30.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "05d454a922c916dadf49c4c393f330fd" + }, + { + "dataPath": "params_shard_124.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.30.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "63856a22d3fe94c5e7b826a664d8fe16" + }, + { + "dataPath": "params_shard_125.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.31.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "f189245c12adcdbabb594dfd63004834" + }, + { + "dataPath": "params_shard_126.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.31.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "d783d0c7f7ce0d9ab67ca8cc88683a72" + }, + { + "dataPath": "params_shard_127.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.31.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "37b47716e6cab80d51d2ef22da02423a" + }, + { + "dataPath": "params_shard_128.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.31.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "93136414d590df68e40cd9b7c8d54642" + }, + { + "dataPath": "params_shard_129.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.32.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "8cebf3bf42c0be0fe4615e091045e5f1" + }, + { + "dataPath": "params_shard_130.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.32.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "d4a7221be2b3875d63bc943db6c7e9ea" + }, + { + "dataPath": "params_shard_131.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.32.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "51482b34c93adf5591e4d20eb2fe9cc5" + }, + { + "dataPath": "params_shard_132.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.32.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "a938aa19af2ac5e1a7be4a9d62afbb4a" + }, + { + "dataPath": "params_shard_133.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.33.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "48de06f22b7d9f7e1062969b08f59bee" + }, + { + "dataPath": "params_shard_134.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.33.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "c3cd145b1d52e5a789bacd5852d757cf" + }, + { + "dataPath": "params_shard_135.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.33.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "c5e75c19618091f7cc441e82cacdb178" + }, + { + "dataPath": "params_shard_136.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.33.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "a9328660ee78726d415ecedb73a05595" + }, + { + "dataPath": "params_shard_137.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.34.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "b5400774e11eb7479e59682fcf48ca2f" + }, + { + "dataPath": "params_shard_138.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.34.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "a592b06aeec1f5c5e9b8fed80352647e" + }, + { + "dataPath": "params_shard_139.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.34.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "550804e4f5d0aa2c564f8a6d370914af" + }, + { + "dataPath": "params_shard_140.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.34.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "a535254722180c0a0e340615721be681" + }, + { + "dataPath": "params_shard_141.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.35.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "d19ca4e5552327dbf359da82fff07851" + }, + { + "dataPath": "params_shard_142.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.35.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "d84a53918f7caa3fc1d55e90c5ff3063" + }, + { + "dataPath": "params_shard_143.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.35.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "75ff61a4d93ae52a38cee962586e09d4" + }, + { + "dataPath": "params_shard_144.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.35.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "4b3d7f04e78cdea4fd912782085059fe" + }, + { + "dataPath": "params_shard_145.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.36.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "2f77cea9990c9b182caaf01fc65731aa" + }, + { + "dataPath": "params_shard_146.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.36.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "bf682f566b1f0f005d98b2165f706985" + }, + { + "dataPath": "params_shard_147.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.36.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "f6e31dad96223a2203e3abc4aeab63b6" + }, + { + "dataPath": "params_shard_148.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.36.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "f57d8dc2aeace971196913bdb3bb20a7" + }, + { + "dataPath": "params_shard_149.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.37.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "43f05564b30a41d16b74a9a2a58461c5" + }, + { + "dataPath": "params_shard_150.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.37.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "15ef6e151987113b53eb6b8bf553cb93" + }, + { + "dataPath": "params_shard_151.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.37.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "ba6c983dcd13cb7ebe67393fddb18ebd" + }, + { + "dataPath": "params_shard_152.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.37.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "89b326f0bbc270c5d848211b33f75282" + }, + { + "dataPath": "params_shard_153.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.38.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "6d6c98526eeecee414755d1b9eb218c2" + }, + { + "dataPath": "params_shard_154.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.38.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "efc27d1bfeeae5a55d187023d814fda7" + }, + { + "dataPath": "params_shard_155.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.38.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "89d55c72ca5f2455acdf2a47c3c206fd" + }, + { + "dataPath": "params_shard_156.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.38.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "53463953a04a04caaa844b11a083b559" + }, + { + "dataPath": "params_shard_157.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.39.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "e2f625fccb46df5fe8783092cf1f38b8" + }, + { + "dataPath": "params_shard_158.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.39.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "4e18e425f8893842443b76e0f065048f" + }, + { + "dataPath": "params_shard_159.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.39.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "2d936649571d6961692ddc69d0a5db70" + }, + { + "dataPath": "params_shard_160.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.39.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "bc34cc6fe4a37feac389b36fe18dd2d0" + }, + { + "dataPath": "params_shard_161.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.40.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "03831ad4f8f9ae173b2f5c7d07dafa1a" + }, + { + "dataPath": "params_shard_162.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.40.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "c8e9e9fc07c29838d7c8f03606f6867b" + }, + { + "dataPath": "params_shard_163.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.40.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "1e11442984dc3747db6f25e2bfbfd5d2" + }, + { + "dataPath": "params_shard_164.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.40.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "6b4b772980ae3f6c831e1cdbcb29d787" + }, + { + "dataPath": "params_shard_165.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.41.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "1ad9afaadf62bf00e595cdd08e203696" + }, + { + "dataPath": "params_shard_166.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.41.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "892bc0ed228f6ffee8747838ff18ef18" + }, + { + "dataPath": "params_shard_167.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.41.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "b7eff25eec698c74beb6c2e815a5c979" + }, + { + "dataPath": "params_shard_168.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.41.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "80d03d9a7dd3d5cdc45175f715abc76f" + }, + { + "dataPath": "params_shard_169.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.42.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "35370a4adb178622ed43e4d56c5c934e" + }, + { + "dataPath": "params_shard_170.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.42.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "368303116313404fd25eed9653c855d7" + }, + { + "dataPath": "params_shard_171.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.42.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "47d4437856896ed12b829a78bed5a6eb" + }, + { + "dataPath": "params_shard_172.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.42.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "5adebc49547de2e6f7bb6851b03f028a" + }, + { + "dataPath": "params_shard_173.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.43.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "2e6dcc7e7418af3d7fb92d5aa42c4d08" + }, + { + "dataPath": "params_shard_174.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.43.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "432b8c18203509aacdd770908d566e7e" + }, + { + "dataPath": "params_shard_175.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.43.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "6f0c116284eb0dc45ef46dad7746aa1c" + }, + { + "dataPath": "params_shard_176.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.43.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "f6ba3f7b643bec7f8446f04217afb0b9" + }, + { + "dataPath": "params_shard_177.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.44.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "e398958a8be1e10a8478d1d2e15fb14c" + }, + { + "dataPath": "params_shard_178.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.44.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "0c38211e4f037b8d63443123ec8d8496" + }, + { + "dataPath": "params_shard_179.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.44.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "c58f41bfdbd0c5d432dd75ae2225c817" + }, + { + "dataPath": "params_shard_180.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.44.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "1fa5a61b58c9f02732d4e1463d092599" + }, + { + "dataPath": "params_shard_181.bin", + "format": "raw-shard", + "nbytes": 679477248, + "records": [ + { + "name": "model.layers.45.mlp.gate_up_proj.weight", + "shape": [ + 73728, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 679477248, + "byteOffset": 0 + } + ], + "md5sum": "68b213f87a3df7334432a82dfd68ca8b" + }, + { + "dataPath": "params_shard_182.bin", + "format": "raw-shard", + "nbytes": 75497472, + "records": [ + { + "name": "model.layers.45.self_attn.qkv_proj.weight", + "shape": [ + 8192, + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 75497472, + "byteOffset": 0 + } + ], + "md5sum": "2b47a8861441bdf53f9dac6ce02980e3" + }, + { + "dataPath": "params_shard_183.bin", + "format": "raw-shard", + "nbytes": 37748736, + "records": [ + { + "name": "model.layers.45.self_attn.o_proj.weight", + "shape": [ + 4608, + 4096 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 37748736, + "byteOffset": 0 + } + ], + "md5sum": "21bf969aca87c0a3b40035d5cddf4b68" + }, + { + "dataPath": "params_shard_184.bin", + "format": "raw-shard", + "nbytes": 339738624, + "records": [ + { + "name": "model.layers.45.mlp.down_proj.weight", + "shape": [ + 4608, + 36864 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 339738624, + "byteOffset": 0 + } + ], + "md5sum": "dbd18d3a38c3bad9a0b9198a90fa9760" + }, + { + "dataPath": "params_shard_185.bin", + "format": "raw-shard", + "nbytes": 1704960, + "records": [ + { + "name": "model.layers.0.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 0 + }, + { + "name": "model.layers.0.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 9216 + }, + { + "name": "model.layers.0.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 18432 + }, + { + "name": "model.layers.0.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 27648 + }, + { + "name": "model.layers.1.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 36864 + }, + { + "name": "model.layers.1.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 46080 + }, + { + "name": "model.layers.1.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 55296 + }, + { + "name": "model.layers.1.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 64512 + }, + { + "name": "model.layers.10.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 73728 + }, + { + "name": "model.layers.10.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 82944 + }, + { + "name": "model.layers.10.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 92160 + }, + { + "name": "model.layers.10.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 101376 + }, + { + "name": "model.layers.11.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 110592 + }, + { + "name": "model.layers.11.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 119808 + }, + { + "name": "model.layers.11.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 129024 + }, + { + "name": "model.layers.11.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 138240 + }, + { + "name": "model.layers.12.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 147456 + }, + { + "name": "model.layers.12.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 156672 + }, + { + "name": "model.layers.12.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 165888 + }, + { + "name": "model.layers.12.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 175104 + }, + { + "name": "model.layers.13.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 184320 + }, + { + "name": "model.layers.13.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 193536 + }, + { + "name": "model.layers.13.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 202752 + }, + { + "name": "model.layers.13.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 211968 + }, + { + "name": "model.layers.14.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 221184 + }, + { + "name": "model.layers.14.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 230400 + }, + { + "name": "model.layers.14.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 239616 + }, + { + "name": "model.layers.14.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 248832 + }, + { + "name": "model.layers.6.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 258048 + }, + { + "name": "model.layers.6.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 267264 + }, + { + "name": "model.layers.6.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 276480 + }, + { + "name": "model.layers.6.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 285696 + }, + { + "name": "model.layers.7.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 294912 + }, + { + "name": "model.layers.7.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 304128 + }, + { + "name": "model.layers.7.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 313344 + }, + { + "name": "model.layers.7.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 322560 + }, + { + "name": "model.layers.8.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 331776 + }, + { + "name": "model.layers.8.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 340992 + }, + { + "name": "model.layers.8.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 350208 + }, + { + "name": "model.layers.8.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 359424 + }, + { + "name": "model.layers.9.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 368640 + }, + { + "name": "model.layers.9.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 377856 + }, + { + "name": "model.layers.9.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 387072 + }, + { + "name": "model.layers.9.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 396288 + }, + { + "name": "model.layers.15.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 405504 + }, + { + "name": "model.layers.15.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 414720 + }, + { + "name": "model.layers.15.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 423936 + }, + { + "name": "model.layers.15.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 433152 + }, + { + "name": "model.layers.16.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 442368 + }, + { + "name": "model.layers.16.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 451584 + }, + { + "name": "model.layers.16.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 460800 + }, + { + "name": "model.layers.16.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 470016 + }, + { + "name": "model.layers.17.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 479232 + }, + { + "name": "model.layers.17.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 488448 + }, + { + "name": "model.layers.17.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 497664 + }, + { + "name": "model.layers.17.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 506880 + }, + { + "name": "model.layers.18.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 516096 + }, + { + "name": "model.layers.18.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 525312 + }, + { + "name": "model.layers.18.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 534528 + }, + { + "name": "model.layers.18.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 543744 + }, + { + "name": "model.layers.19.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 552960 + }, + { + "name": "model.layers.19.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 562176 + }, + { + "name": "model.layers.19.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 571392 + }, + { + "name": "model.layers.19.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 580608 + }, + { + "name": "model.layers.20.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 589824 + }, + { + "name": "model.layers.20.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 599040 + }, + { + "name": "model.layers.20.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 608256 + }, + { + "name": "model.layers.20.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 617472 + }, + { + "name": "model.layers.21.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 626688 + }, + { + "name": "model.layers.21.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 635904 + }, + { + "name": "model.layers.21.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 645120 + }, + { + "name": "model.layers.21.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 654336 + }, + { + "name": "model.layers.22.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 663552 + }, + { + "name": "model.layers.22.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 672768 + }, + { + "name": "model.layers.22.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 681984 + }, + { + "name": "model.layers.22.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 691200 + }, + { + "name": "model.layers.2.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 700416 + }, + { + "name": "model.layers.2.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 709632 + }, + { + "name": "model.layers.2.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 718848 + }, + { + "name": "model.layers.2.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 728064 + }, + { + "name": "model.layers.3.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 737280 + }, + { + "name": "model.layers.3.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 746496 + }, + { + "name": "model.layers.3.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 755712 + }, + { + "name": "model.layers.3.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 764928 + }, + { + "name": "model.layers.4.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 774144 + }, + { + "name": "model.layers.4.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 783360 + }, + { + "name": "model.layers.4.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 792576 + }, + { + "name": "model.layers.4.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 801792 + }, + { + "name": "model.layers.5.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 811008 + }, + { + "name": "model.layers.5.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 820224 + }, + { + "name": "model.layers.5.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 829440 + }, + { + "name": "model.layers.5.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 838656 + }, + { + "name": "model.layers.23.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 847872 + }, + { + "name": "model.layers.23.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 857088 + }, + { + "name": "model.layers.23.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 866304 + }, + { + "name": "model.layers.23.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 875520 + }, + { + "name": "model.layers.24.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 884736 + }, + { + "name": "model.layers.24.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 893952 + }, + { + "name": "model.layers.24.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 903168 + }, + { + "name": "model.layers.24.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 912384 + }, + { + "name": "model.layers.25.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 921600 + }, + { + "name": "model.layers.25.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 930816 + }, + { + "name": "model.layers.25.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 940032 + }, + { + "name": "model.layers.25.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 949248 + }, + { + "name": "model.layers.26.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 958464 + }, + { + "name": "model.layers.26.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 967680 + }, + { + "name": "model.layers.26.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 976896 + }, + { + "name": "model.layers.26.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 986112 + }, + { + "name": "model.layers.27.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 995328 + }, + { + "name": "model.layers.27.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1004544 + }, + { + "name": "model.layers.27.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1013760 + }, + { + "name": "model.layers.27.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1022976 + }, + { + "name": "model.layers.28.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1032192 + }, + { + "name": "model.layers.28.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1041408 + }, + { + "name": "model.layers.28.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1050624 + }, + { + "name": "model.layers.28.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1059840 + }, + { + "name": "model.layers.29.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1069056 + }, + { + "name": "model.layers.29.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1078272 + }, + { + "name": "model.layers.29.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1087488 + }, + { + "name": "model.layers.29.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1096704 + }, + { + "name": "model.layers.30.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1105920 + }, + { + "name": "model.layers.30.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1115136 + }, + { + "name": "model.layers.30.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1124352 + }, + { + "name": "model.layers.30.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1133568 + }, + { + "name": "model.layers.31.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1142784 + }, + { + "name": "model.layers.31.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1152000 + }, + { + "name": "model.layers.31.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1161216 + }, + { + "name": "model.layers.31.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1170432 + }, + { + "name": "model.layers.32.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1179648 + }, + { + "name": "model.layers.32.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1188864 + }, + { + "name": "model.layers.32.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1198080 + }, + { + "name": "model.layers.32.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1207296 + }, + { + "name": "model.layers.33.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1216512 + }, + { + "name": "model.layers.33.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1225728 + }, + { + "name": "model.layers.33.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1234944 + }, + { + "name": "model.layers.33.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1244160 + }, + { + "name": "model.layers.34.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1253376 + }, + { + "name": "model.layers.34.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1262592 + }, + { + "name": "model.layers.34.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1271808 + }, + { + "name": "model.layers.34.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1281024 + }, + { + "name": "model.layers.35.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1290240 + }, + { + "name": "model.layers.35.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1299456 + }, + { + "name": "model.layers.35.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1308672 + }, + { + "name": "model.layers.35.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1317888 + }, + { + "name": "model.layers.36.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1327104 + }, + { + "name": "model.layers.36.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1336320 + }, + { + "name": "model.layers.36.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1345536 + }, + { + "name": "model.layers.36.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1354752 + }, + { + "name": "model.layers.37.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1363968 + }, + { + "name": "model.layers.37.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1373184 + }, + { + "name": "model.layers.37.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1382400 + }, + { + "name": "model.layers.37.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1391616 + }, + { + "name": "model.layers.38.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1400832 + }, + { + "name": "model.layers.38.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1410048 + }, + { + "name": "model.layers.38.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1419264 + }, + { + "name": "model.layers.38.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1428480 + }, + { + "name": "model.layers.39.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1437696 + }, + { + "name": "model.layers.39.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1446912 + }, + { + "name": "model.layers.39.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1456128 + }, + { + "name": "model.layers.39.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1465344 + }, + { + "name": "model.layers.40.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1474560 + }, + { + "name": "model.layers.40.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1483776 + }, + { + "name": "model.layers.40.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1492992 + }, + { + "name": "model.layers.40.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1502208 + }, + { + "name": "model.layers.41.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1511424 + }, + { + "name": "model.layers.41.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1520640 + }, + { + "name": "model.layers.41.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1529856 + }, + { + "name": "model.layers.41.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1539072 + }, + { + "name": "model.layers.42.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1548288 + }, + { + "name": "model.layers.42.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1557504 + }, + { + "name": "model.layers.42.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1566720 + }, + { + "name": "model.layers.42.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1575936 + }, + { + "name": "model.layers.43.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1585152 + }, + { + "name": "model.layers.43.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1594368 + }, + { + "name": "model.layers.43.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1603584 + }, + { + "name": "model.layers.43.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1612800 + }, + { + "name": "model.layers.44.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1622016 + }, + { + "name": "model.layers.44.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1631232 + }, + { + "name": "model.layers.44.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1640448 + }, + { + "name": "model.layers.44.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1649664 + }, + { + "name": "model.layers.45.input_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1658880 + }, + { + "name": "model.layers.45.post_attention_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1668096 + }, + { + "name": "model.layers.45.post_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1677312 + }, + { + "name": "model.layers.45.pre_feedforward_layernorm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1686528 + }, + { + "name": "model.norm.weight", + "shape": [ + 4608 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9216, + "byteOffset": 1695744 + } + ], + "md5sum": "ac76064680a74df4c5cfd5b812c392d0" + } + ] +} \ No newline at end of file diff --git a/params_shard_0.bin b/params_shard_0.bin new file mode 100644 index 0000000000000000000000000000000000000000..a053ff18b84a7c38c396c0fa6b1e87cd2d05e9cf --- /dev/null +++ b/params_shard_0.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ecaa755837342c5069038bb5054525e02744d3ecc35b5109f2be6ec3d152df1f +size 2359296000 diff --git a/params_shard_1.bin b/params_shard_1.bin new file mode 100644 index 0000000000000000000000000000000000000000..cb199fa04bb8ce3d9d90ccff5e5a225a71f4f419 --- /dev/null +++ b/params_shard_1.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:14d671c57e249ef013bb2136f6accdc8742e7367ecfa7c8b9cca83bce7708998 +size 339738624 diff --git a/params_shard_10.bin b/params_shard_10.bin new file mode 100644 index 0000000000000000000000000000000000000000..d2476f3decb92fc64b48d8fb03496f24a1e47ed8 --- /dev/null +++ b/params_shard_10.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bb54ac3fe9d341753e0cbcdc9c2528fc54652f80517ae6c4000cbd456f68f797 +size 37748736 diff --git a/params_shard_100.bin b/params_shard_100.bin new file mode 100644 index 0000000000000000000000000000000000000000..67ffe5be5d1b0503c984d9458bb9dcdd70ead506 --- /dev/null +++ b/params_shard_100.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:076d647c503ae058200793c5e32daa9083874ead355f0fe14f6f8b016a37d953 +size 37748736 diff --git a/params_shard_101.bin b/params_shard_101.bin new file mode 100644 index 0000000000000000000000000000000000000000..d369d706712fd636322db7fa70bb7a5f4cee0b14 --- /dev/null +++ b/params_shard_101.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62aa92ed97cb853ed03827dc749df47cdebcdb55ab0aa9a0047f0eafad75c3a6 +size 339738624 diff --git a/params_shard_102.bin b/params_shard_102.bin new file mode 100644 index 0000000000000000000000000000000000000000..85b42a90650d856bd9861e721c46252b4d8a02f1 --- /dev/null +++ b/params_shard_102.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a8efe171b6009827ecda19248f8401ce81301772dce8cccf44e6a91f40e62664 +size 679477248 diff --git a/params_shard_103.bin b/params_shard_103.bin new file mode 100644 index 0000000000000000000000000000000000000000..6cc4d75fdd8c4c3ff9bf82a58edb1d90dc0d328b --- /dev/null +++ b/params_shard_103.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4dd429f2c7befe252c4bd828f02ad65f841910e757e729748467b02e6f006305 +size 75497472 diff --git a/params_shard_104.bin b/params_shard_104.bin new file mode 100644 index 0000000000000000000000000000000000000000..1517e16c2d6350b7124d6749974b61d9d6b8afbd --- /dev/null +++ b/params_shard_104.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5881e82b721ff399ad908d854acf53539c648171e6720a3a7eb28fb256aacb6f +size 37748736 diff --git a/params_shard_105.bin b/params_shard_105.bin new file mode 100644 index 0000000000000000000000000000000000000000..7a641ff7e0886e2c51484473188d59acd1856ac7 --- /dev/null +++ b/params_shard_105.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:85deca34e0f233eed6cb439415ee549f2ea2c3c7e6cb45553e0d9d7f49271cb0 +size 339738624 diff --git a/params_shard_106.bin b/params_shard_106.bin new file mode 100644 index 0000000000000000000000000000000000000000..7a5f17aa385bab8d13f1df66aff032155efd2434 --- /dev/null +++ b/params_shard_106.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8366b8d0314da4d8a8519993076bba671d18479b65cfa1a73cf7969c60ea4ffa +size 679477248 diff --git a/params_shard_107.bin b/params_shard_107.bin new file mode 100644 index 0000000000000000000000000000000000000000..90aaaa634d99706fb5a06aae02865ed845a488e6 --- /dev/null +++ b/params_shard_107.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f53bf4a48f761e9fe0f4da8b24dfef5eb03105a4a5b3df1e6a4a07aa8c39d866 +size 75497472 diff --git a/params_shard_108.bin b/params_shard_108.bin new file mode 100644 index 0000000000000000000000000000000000000000..640f2dc4d2a637b929a8f45b1333ebadf1ee8d35 --- /dev/null +++ b/params_shard_108.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9f3356eca8e15400fdab4abb1689a8800e1626d59430f3c8fcd7062d834cb3e8 +size 37748736 diff --git a/params_shard_109.bin b/params_shard_109.bin new file mode 100644 index 0000000000000000000000000000000000000000..7333886b48e614f881fdc94ef69fb1de2c9342eb --- /dev/null +++ b/params_shard_109.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b851241961ed31bd317ffb8d15eaa101c026b5c0ffe7748473e4ff23150861ea +size 339738624 diff --git a/params_shard_11.bin b/params_shard_11.bin new file mode 100644 index 0000000000000000000000000000000000000000..c8b373bb76ac60b636fb4ac7b24c01d908498aef --- /dev/null +++ b/params_shard_11.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6a73643dd56d9f775aab47dabb8c293a0030bc5b8160e9e67d78db516d72a5ce +size 339738624 diff --git a/params_shard_110.bin b/params_shard_110.bin new file mode 100644 index 0000000000000000000000000000000000000000..60c2a362d56e0a6df2036f3009c0735cc085569e --- /dev/null +++ b/params_shard_110.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:27138550c3d349472b3971d230aff44cce5ed4378fa1c3adafccee199982c8cc +size 679477248 diff --git a/params_shard_111.bin b/params_shard_111.bin new file mode 100644 index 0000000000000000000000000000000000000000..d0e6c26943485538305fd1161f7a851247acb566 --- /dev/null +++ b/params_shard_111.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:be7903fc48dfc16e5763c392d941097a2fa18547ce783cbf6a930e53a2660226 +size 75497472 diff --git a/params_shard_112.bin b/params_shard_112.bin new file mode 100644 index 0000000000000000000000000000000000000000..507aeedc7b083a3fccc519a535ded299c0cd2a09 --- /dev/null +++ b/params_shard_112.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:123854a67f40a398121c4baf172b95337945a4b917e5a7018b8e00b99c19eb2e +size 37748736 diff --git a/params_shard_113.bin b/params_shard_113.bin new file mode 100644 index 0000000000000000000000000000000000000000..5281968ad4f806e1dce2503ca1576b117dc4c9fd --- /dev/null +++ b/params_shard_113.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:99269b8152eeb6b36175297e2e3603a517fc57d307a7e85658a308074cafafde +size 75497472 diff --git a/params_shard_114.bin b/params_shard_114.bin new file mode 100644 index 0000000000000000000000000000000000000000..5961e93f4cf816579e2c3529d02aa43135bc6722 --- /dev/null +++ b/params_shard_114.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29361e843cb11a33ea8e587cc46677536831bf4ca932c559b65bcada4f9c8f59 +size 37748736 diff --git a/params_shard_115.bin b/params_shard_115.bin new file mode 100644 index 0000000000000000000000000000000000000000..d190b981a6f7f4393d73ab4e66f03970bc740062 --- /dev/null +++ b/params_shard_115.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6677f0a55a895af5246354bf61936e077d155dae0d5cdc3fbe75e4718131bcc2 +size 339738624 diff --git a/params_shard_116.bin b/params_shard_116.bin new file mode 100644 index 0000000000000000000000000000000000000000..b8b0a95f8bdd38cfc9b8bf7e48f3eb0d5174a9bd --- /dev/null +++ b/params_shard_116.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2ff826c0245003fdd46e2cbbbfe7980c0ad7d5a11adeadd5e7bc8c83e4ac5d7f +size 679477248 diff --git a/params_shard_117.bin b/params_shard_117.bin new file mode 100644 index 0000000000000000000000000000000000000000..ffcbec14f5bf5bc6a051ab0d94173a058b8a1eaa --- /dev/null +++ b/params_shard_117.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eba4e94dc47467252d84fa115c081ac7a71f173e7a17de900fc22dfe6db065fe +size 339738624 diff --git a/params_shard_118.bin b/params_shard_118.bin new file mode 100644 index 0000000000000000000000000000000000000000..f5e81f0b5f27e2445a609c0ff996cffc824b2074 --- /dev/null +++ b/params_shard_118.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:46939bf5e252db56ced273bcc51e166957617290d078331e22ce7b2e2a4b7276 +size 679477248 diff --git a/params_shard_119.bin b/params_shard_119.bin new file mode 100644 index 0000000000000000000000000000000000000000..406eb9f6e3154718bd20f2995f31fab9aed02cc5 --- /dev/null +++ b/params_shard_119.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:133d66efc7353a9d6dae8f51fed094db7b1ffc3409f87ec1fd0a2d60e92b894e +size 75497472 diff --git a/params_shard_12.bin b/params_shard_12.bin new file mode 100644 index 0000000000000000000000000000000000000000..51e8ff684bd11fb1fc9081d4469d59f8e45d0d69 --- /dev/null +++ b/params_shard_12.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:941f2874ab37ab9b2ec2880deeb1b44290ae1db0be83daa9a6f72803c3c93061 +size 339738624 diff --git a/params_shard_120.bin b/params_shard_120.bin new file mode 100644 index 0000000000000000000000000000000000000000..27fd9181d1e961f235b2a2bc5d8b5d3c14554ffb --- /dev/null +++ b/params_shard_120.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:283512493abe797d500dfdb8e0cf024dbe68e14f8f7473b1215466001e41aadf +size 37748736 diff --git a/params_shard_121.bin b/params_shard_121.bin new file mode 100644 index 0000000000000000000000000000000000000000..f3067e28742acf9201d1407d46a7dc9c0cf885c4 --- /dev/null +++ b/params_shard_121.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:91f05a00ea740d9ff1883c5442ddc2990e05760399ad2d5d17d29a06da50ddea +size 339738624 diff --git a/params_shard_122.bin b/params_shard_122.bin new file mode 100644 index 0000000000000000000000000000000000000000..c431ed7de148da19c5b6b0dd1e5a63183bfbfdd9 --- /dev/null +++ b/params_shard_122.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:473641ba16f348f1d459490b4c33e89776adc6461a71d550e42ffb6cc39ed5bc +size 679477248 diff --git a/params_shard_123.bin b/params_shard_123.bin new file mode 100644 index 0000000000000000000000000000000000000000..03b462558476f82c9f9512d549ef0fc23c6b296a --- /dev/null +++ b/params_shard_123.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4c6cf4aafe569d0216519598cbd728deb94c36106dcab95e2c0f65708922cc7b +size 75497472 diff --git a/params_shard_124.bin b/params_shard_124.bin new file mode 100644 index 0000000000000000000000000000000000000000..0391820291950a2939dd462a7e9eb6c6ff27649e --- /dev/null +++ b/params_shard_124.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:38a77fb0e2002dadf6e1e5343eac729771945c47a9f4895adfc53eef59e227e2 +size 37748736 diff --git a/params_shard_125.bin b/params_shard_125.bin new file mode 100644 index 0000000000000000000000000000000000000000..c0c01862d3edfb707daba60b411c3375842103b9 --- /dev/null +++ b/params_shard_125.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2d3399d324c7e0985ff4256379bd1cab7ce8114b1b48acf01dd57e6c9a51d015 +size 339738624 diff --git a/params_shard_126.bin b/params_shard_126.bin new file mode 100644 index 0000000000000000000000000000000000000000..d432d039bec6929d584e5492d385884cd1369edf --- /dev/null +++ b/params_shard_126.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3094b8c41bc6132bf3f7532886d6a9ba0f5d5ab04f383d0a6ee3e263c7d06bf8 +size 679477248 diff --git a/params_shard_127.bin b/params_shard_127.bin new file mode 100644 index 0000000000000000000000000000000000000000..4ec9d6810c31829e418b317356f7f34611766f9b --- /dev/null +++ b/params_shard_127.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d45df021550c221753d1d0628bdc31ec4260fb928ec2f91748c9a879872f14ab +size 75497472 diff --git a/params_shard_128.bin b/params_shard_128.bin new file mode 100644 index 0000000000000000000000000000000000000000..bd35ed3fc6e22f552c4463863ec7b5c1ffd1698e --- /dev/null +++ b/params_shard_128.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c2b838dd62f14dd5a05c7bb824666da2a88edec3c9369ef4381ae73ef077d046 +size 37748736 diff --git a/params_shard_129.bin b/params_shard_129.bin new file mode 100644 index 0000000000000000000000000000000000000000..29a47089aec3b6764bab5dbd1b6277399df8c8db --- /dev/null +++ b/params_shard_129.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dba8c4754964cbc59fa8d2a71d672140dc5371ab1446d159e3567bcc67c234e2 +size 679477248 diff --git a/params_shard_13.bin b/params_shard_13.bin new file mode 100644 index 0000000000000000000000000000000000000000..06c9c97388df9a42185b3fd0c9f06cbae8459226 --- /dev/null +++ b/params_shard_13.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7eadc8293e643c9b9406c4efdcc9203fa1d34ca1ca43cb3879786bb16322b1ff +size 679477248 diff --git a/params_shard_130.bin b/params_shard_130.bin new file mode 100644 index 0000000000000000000000000000000000000000..3d907ea1f252b1a81c880e91e9c7dcc7f6f44ceb --- /dev/null +++ b/params_shard_130.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:efca8ee2862c483f29bac6b05ae9397a73e3739d5d5589fe38379f9ef574d918 +size 75497472 diff --git a/params_shard_131.bin b/params_shard_131.bin new file mode 100644 index 0000000000000000000000000000000000000000..e2a2b8f293530508a6a64fcc0906f194a169b88a --- /dev/null +++ b/params_shard_131.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:98f78d32035b441a1a32af41d633456837cd39d81f78034077a0f3348d361e43 +size 37748736 diff --git a/params_shard_132.bin b/params_shard_132.bin new file mode 100644 index 0000000000000000000000000000000000000000..c1d5388617177a1bb31585482983580a7ec3eb95 --- /dev/null +++ b/params_shard_132.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1d6d5339e8be94409ec67479876b823deba6898ba0f8138f7e8b8971e5293638 +size 339738624 diff --git a/params_shard_133.bin b/params_shard_133.bin new file mode 100644 index 0000000000000000000000000000000000000000..b43142b3344c9608892311342e10a8d00000ab8d --- /dev/null +++ b/params_shard_133.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a520ecbcf82789087b8df3c33e1ce4766aa5f4b159081e0cb959db336319a51 +size 339738624 diff --git a/params_shard_134.bin b/params_shard_134.bin new file mode 100644 index 0000000000000000000000000000000000000000..8f4c0c0c2a38f66cfc49a72f8da2188e72a90eff --- /dev/null +++ b/params_shard_134.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d65a166b8ac7a93171b6b9893fe6805a106e78a6968167d25653115e478b7432 +size 679477248 diff --git a/params_shard_135.bin b/params_shard_135.bin new file mode 100644 index 0000000000000000000000000000000000000000..68cc1c73136ffe94b3d2477596d9bace258fe1d4 --- /dev/null +++ b/params_shard_135.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cf66535f3e36f427f73d68064551ca26513417f49ef59f54a9cfb539978e4690 +size 75497472 diff --git a/params_shard_136.bin b/params_shard_136.bin new file mode 100644 index 0000000000000000000000000000000000000000..927e6b6e608b9353ef4fddab56f74a6be8a3b378 --- /dev/null +++ b/params_shard_136.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0891e3b6752084b80c3391aeb065d91c20ed31d0314cc0430281f1f17bc74fbd +size 37748736 diff --git a/params_shard_137.bin b/params_shard_137.bin new file mode 100644 index 0000000000000000000000000000000000000000..67b228bfad9def2d3d27296433a024adcb7482ac --- /dev/null +++ b/params_shard_137.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8dd1d57b1a92b0863baa747133d50c1e53ec6e67e165aea37ab0fc287b16f16c +size 339738624 diff --git a/params_shard_138.bin b/params_shard_138.bin new file mode 100644 index 0000000000000000000000000000000000000000..a3e99bb007d5c1670368c02fe1836858c4ca84c2 --- /dev/null +++ b/params_shard_138.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62ff833d2cecf7557bf9ec2a8c5c9a0cf6f734e564ec6cc9ac5f9c33ef743a46 +size 679477248 diff --git a/params_shard_139.bin b/params_shard_139.bin new file mode 100644 index 0000000000000000000000000000000000000000..f7965752b07840bffa97bb2e8a9f1a125886fb67 --- /dev/null +++ b/params_shard_139.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2849307db6413641f27e967726dddeee41913f015f64694b4610c97ed0cb46f9 +size 75497472 diff --git a/params_shard_14.bin b/params_shard_14.bin new file mode 100644 index 0000000000000000000000000000000000000000..63471d85ea89e64224e236b146ef9b8bcf508836 --- /dev/null +++ b/params_shard_14.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29efc3743754155f6abc6ac884c813bd76b15d6ea0d58293c1050a8187960def +size 75497472 diff --git a/params_shard_140.bin b/params_shard_140.bin new file mode 100644 index 0000000000000000000000000000000000000000..42ba27d0c58fec09a7cfccc2fff9af416a9f13de --- /dev/null +++ b/params_shard_140.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d19f999c0ee5b6e1b6d948afa7d22bd43713b534d1691b6720d61bf6b22b7aee +size 37748736 diff --git a/params_shard_141.bin b/params_shard_141.bin new file mode 100644 index 0000000000000000000000000000000000000000..34dc0d568482eadf58c50ff2d79651cea4d7157f --- /dev/null +++ b/params_shard_141.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d15148d0c91a03af4029646390f5af283411829ba13184730fce913b938ec9ce +size 339738624 diff --git a/params_shard_142.bin b/params_shard_142.bin new file mode 100644 index 0000000000000000000000000000000000000000..20c0f695f541cc1e2c702e83ff201032359f73c0 --- /dev/null +++ b/params_shard_142.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:311b3faa8ecb3fed3ee2e84ad21e038b45148b4f027de071e4e2240a59dcfd33 +size 679477248 diff --git a/params_shard_143.bin b/params_shard_143.bin new file mode 100644 index 0000000000000000000000000000000000000000..0e3e0c47cf46e9e60349eb01846ffcb7ac37ffba --- /dev/null +++ b/params_shard_143.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3a830a569d98841c6cd30318dea68d27a1290895749a6eafa70b09e220c43d60 +size 75497472 diff --git a/params_shard_144.bin b/params_shard_144.bin new file mode 100644 index 0000000000000000000000000000000000000000..9930ca4a6a8baee12f24c56924f56a6be879629c --- /dev/null +++ b/params_shard_144.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:19a09a3d79619f3cebca58d56d23e2e5732c5a30d5fbca5da052b3c27d78d646 +size 37748736 diff --git a/params_shard_145.bin b/params_shard_145.bin new file mode 100644 index 0000000000000000000000000000000000000000..df7cb67be5205db544f75c09db2a7c3dc6a924f4 --- /dev/null +++ b/params_shard_145.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d21a5bf387b8630a21fdad46ab8d2ce63e7b4c32ecaaedfc1c8586458507902f +size 679477248 diff --git a/params_shard_146.bin b/params_shard_146.bin new file mode 100644 index 0000000000000000000000000000000000000000..5a7a2417222f4117cf9ad54a8d4df9744b071f38 --- /dev/null +++ b/params_shard_146.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0e3cbe732e375de0ca8968290c7681e843989116cd11c4fa4a3fdb2fef5ddd77 +size 75497472 diff --git a/params_shard_147.bin b/params_shard_147.bin new file mode 100644 index 0000000000000000000000000000000000000000..bed0ee96d66d29e36c0a0bcfa380f27c0fa69044 --- /dev/null +++ b/params_shard_147.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b67c757e5bbcd1af6500b27224e55957cd63bb6d796f01a7e384531e7693398b +size 37748736 diff --git a/params_shard_148.bin b/params_shard_148.bin new file mode 100644 index 0000000000000000000000000000000000000000..28715d5d926ee620e1f89a63eddedc21b1aae5fa --- /dev/null +++ b/params_shard_148.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:84ec573ab0c833df3770f179919db29f2e668c2e01f5ba82f5a2b1a07d6fc1d5 +size 339738624 diff --git a/params_shard_149.bin b/params_shard_149.bin new file mode 100644 index 0000000000000000000000000000000000000000..e6b311852e4aab13c2ee71f1897e482c4d10a362 --- /dev/null +++ b/params_shard_149.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:12988d94138b40ad02452f3123a880374fb1998bfe27da18039bc21c39783d1b +size 339738624 diff --git a/params_shard_15.bin b/params_shard_15.bin new file mode 100644 index 0000000000000000000000000000000000000000..8805caa9d9619b24ffe8050bdbde7ed778372318 --- /dev/null +++ b/params_shard_15.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:51e401192066ed31b43d8cb6765ffa4f63313d719f50019a50d6a74ee2cd762b +size 37748736 diff --git a/params_shard_150.bin b/params_shard_150.bin new file mode 100644 index 0000000000000000000000000000000000000000..bd68391626d3a6b6ecbaefb45446fa4f189346d0 --- /dev/null +++ b/params_shard_150.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dbc7d03d5357339b40be8b21993450feeb18bc2a583ebc5b4c234df7a25c0511 +size 679477248 diff --git a/params_shard_151.bin b/params_shard_151.bin new file mode 100644 index 0000000000000000000000000000000000000000..fd9d1ee5f6d4277fe7508e901007efba4b9f8451 --- /dev/null +++ b/params_shard_151.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:22da1db5a0ac843490f32f857551bf9345df2a7d193abafa5d2bee8fbe7c53f8 +size 75497472 diff --git a/params_shard_152.bin b/params_shard_152.bin new file mode 100644 index 0000000000000000000000000000000000000000..d45169015b2ed15aca096cae387e61aec49c6c24 --- /dev/null +++ b/params_shard_152.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e9faaa9c3308e7fce5ec41154ffba51d68c9f8f37ceec112f6c040276835009 +size 37748736 diff --git a/params_shard_153.bin b/params_shard_153.bin new file mode 100644 index 0000000000000000000000000000000000000000..2bd30a189933255d45dbe933f4cb561855ac1d37 --- /dev/null +++ b/params_shard_153.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e03be8071a3005155e6320a6c6076ac307eb8c90a0a9ad38acd59d3793ec724a +size 339738624 diff --git a/params_shard_154.bin b/params_shard_154.bin new file mode 100644 index 0000000000000000000000000000000000000000..602d66b709aea39f778e9075220de7665a558d1d --- /dev/null +++ b/params_shard_154.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6cbc5f5b2a34c26defce1aa328ab79e6de4623b50a3d51205e6ccde28e7cb8a8 +size 679477248 diff --git a/params_shard_155.bin b/params_shard_155.bin new file mode 100644 index 0000000000000000000000000000000000000000..557768ef97cd3a33ec82e40c849330af7f8a4fb9 --- /dev/null +++ b/params_shard_155.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ecccd81d11abb49b7ef7e30e030c4c2ca2464260fd210ff8a71ef001c059376a +size 75497472 diff --git a/params_shard_156.bin b/params_shard_156.bin new file mode 100644 index 0000000000000000000000000000000000000000..2f48400454885c4b2e83bb6e941037b42e4cbee7 --- /dev/null +++ b/params_shard_156.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7f3a9d450e71e7be20f4b7f6557c55ededc89a3f3479271f5edb664294a6c7ce +size 37748736 diff --git a/params_shard_157.bin b/params_shard_157.bin new file mode 100644 index 0000000000000000000000000000000000000000..2787bdfca9992004d5cdcabe9a87b0c3c1d86955 --- /dev/null +++ b/params_shard_157.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:db5d3553ddeb32d8b0261206f5b172d7f91f00ab8862ab2a072beab9c7ebf1e4 +size 339738624 diff --git a/params_shard_158.bin b/params_shard_158.bin new file mode 100644 index 0000000000000000000000000000000000000000..2caa02c3fab953f0f77b7d79a10087edf5ceee41 --- /dev/null +++ b/params_shard_158.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:80724a4eef08ea5594bff039bef2a5bc4f49a20c8ae912b84b9aa33499ec3e7a +size 679477248 diff --git a/params_shard_159.bin b/params_shard_159.bin new file mode 100644 index 0000000000000000000000000000000000000000..5723772c5c30e87bc5443eda0d6fe0c681daa6d6 --- /dev/null +++ b/params_shard_159.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d6008683263e7ca7f9885849d5405e6e1b0334bbb735c12f49525e7096cf8726 +size 75497472 diff --git a/params_shard_16.bin b/params_shard_16.bin new file mode 100644 index 0000000000000000000000000000000000000000..c1a52d650d6d705ab921a18d955ed5058a033c3c --- /dev/null +++ b/params_shard_16.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff81591ebee7635e615f31daa9ae613a6199941adcde365d1ac22eca15055574 +size 339738624 diff --git a/params_shard_160.bin b/params_shard_160.bin new file mode 100644 index 0000000000000000000000000000000000000000..617f07c3d39b6044cbf457d9a7d5f3b3176fc260 --- /dev/null +++ b/params_shard_160.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cef56f725e719a95aef6db2bd18f84d00268a9955efad9b5017341eeb581a083 +size 37748736 diff --git a/params_shard_161.bin b/params_shard_161.bin new file mode 100644 index 0000000000000000000000000000000000000000..20ec6f0ae0a28cf38884b8fb21ce6a90bf92d88d --- /dev/null +++ b/params_shard_161.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5493a563922ee55701b41537081217109b5ce14f7545e1c099b2a7b743ee4847 +size 339738624 diff --git a/params_shard_162.bin b/params_shard_162.bin new file mode 100644 index 0000000000000000000000000000000000000000..2e715a03493d64466b5b73c37f679aa76b23743e --- /dev/null +++ b/params_shard_162.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:439bcc6ebe4d473795e1bd9857055d86986831fb34119b1e2b097a8c65147d48 +size 679477248 diff --git a/params_shard_163.bin b/params_shard_163.bin new file mode 100644 index 0000000000000000000000000000000000000000..e74a457cc5e5e1ef51e6bca9bcd75ee8b9afb4d0 --- /dev/null +++ b/params_shard_163.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4c9f8b878026e5bd62e15049150b8b56ff0d899c91b772cfba4b071c4da68a90 +size 75497472 diff --git a/params_shard_164.bin b/params_shard_164.bin new file mode 100644 index 0000000000000000000000000000000000000000..4cb44c3d5cb75e1feda2434399276792e0177a1d --- /dev/null +++ b/params_shard_164.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3800ca1a35ccf88e49284f4f5e6b8b6d922432a23824d63a7d812725c02938c2 +size 37748736 diff --git a/params_shard_165.bin b/params_shard_165.bin new file mode 100644 index 0000000000000000000000000000000000000000..8dd6d7a47df9ab0526f4df465dcde8a19930230d --- /dev/null +++ b/params_shard_165.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:daa1fc98c3e0f86608d52bbb3d7b2fab0a78824e9fda823eba44994ef119889f +size 75497472 diff --git a/params_shard_166.bin b/params_shard_166.bin new file mode 100644 index 0000000000000000000000000000000000000000..3bb364849e4fbb9dbe3fa05ee2ae1e155847b76d --- /dev/null +++ b/params_shard_166.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cf0e66c534f8e374228d36dc778839fbf238f4ccfd59a5b1f16c03370985c711 +size 37748736 diff --git a/params_shard_167.bin b/params_shard_167.bin new file mode 100644 index 0000000000000000000000000000000000000000..5311ff8c916101e85b8e3e3a0c362dd6034e560b --- /dev/null +++ b/params_shard_167.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e144d3867940ab2944dd2b17073d9ca3870a6086394afac4889d0219c05d6d4c +size 339738624 diff --git a/params_shard_168.bin b/params_shard_168.bin new file mode 100644 index 0000000000000000000000000000000000000000..f0a9fa5aa198a2ac3e6218fc0f61f84f018a197c --- /dev/null +++ b/params_shard_168.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8c430b009d481e80fe6e925c2f44e4c958eeef0c6b07caf60e2d388f79e297fc +size 679477248 diff --git a/params_shard_169.bin b/params_shard_169.bin new file mode 100644 index 0000000000000000000000000000000000000000..12ff51a5dd0d1df0c1dbfe6f556e5d8f85f516b0 --- /dev/null +++ b/params_shard_169.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:458c4e31bef46b71cc7e3522ebf39da9bcfe1d2b76fcba6a39708c617df3288e +size 339738624 diff --git a/params_shard_17.bin b/params_shard_17.bin new file mode 100644 index 0000000000000000000000000000000000000000..3ef23af8e85fa793fe73e9bc174ec73704c65d7b --- /dev/null +++ b/params_shard_17.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:73bb259bdbe8c86316d7a389b283c61f5f1951956ddcd1d5e47431cf83c8a2a9 +size 679477248 diff --git a/params_shard_170.bin b/params_shard_170.bin new file mode 100644 index 0000000000000000000000000000000000000000..3401e908b34942b596f6c08a361bed23b1b9715f --- /dev/null +++ b/params_shard_170.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e837fcd5b2f479be6c67651c94eed8eefd11f03ca870182beec825a8eb4f4f0a +size 679477248 diff --git a/params_shard_171.bin b/params_shard_171.bin new file mode 100644 index 0000000000000000000000000000000000000000..05a385cb357d5475d58c783c16ad20480d462ab5 --- /dev/null +++ b/params_shard_171.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:56fdb14ab11c21f1131dd3b0f57af4b23c882e3f8f52d4c5a8ce15cea4902909 +size 75497472 diff --git a/params_shard_172.bin b/params_shard_172.bin new file mode 100644 index 0000000000000000000000000000000000000000..bd25d54082da9b172ffbbef1cc812d92b0a27b38 --- /dev/null +++ b/params_shard_172.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:674fb01a152e9159fcee40854119ce8453f37748e1861e32f765cf70f52ce272 +size 37748736 diff --git a/params_shard_173.bin b/params_shard_173.bin new file mode 100644 index 0000000000000000000000000000000000000000..26bc010e6b68849f42d59f86dd00c7f83770cf6e --- /dev/null +++ b/params_shard_173.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ebde6486ef5820b5b9f40a6b9b8a8128e0038ac9fd1b8804596496c4dfaaa0da +size 339738624 diff --git a/params_shard_174.bin b/params_shard_174.bin new file mode 100644 index 0000000000000000000000000000000000000000..ef35352bbcf9e4fd8716b4b4a210e22f4e38edd2 --- /dev/null +++ b/params_shard_174.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:78253f87f18d7d9d68b5a36e97bcdf9c08af9d33651969a16c4bc8c2a417c007 +size 679477248 diff --git a/params_shard_175.bin b/params_shard_175.bin new file mode 100644 index 0000000000000000000000000000000000000000..8ec30241ac2d433305921b2cc6de3ea8ec0e6ff5 --- /dev/null +++ b/params_shard_175.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:175620dc24af1edcc6a707e881e6d2bc557708d0c005314f902fbdcea314a0e6 +size 75497472 diff --git a/params_shard_176.bin b/params_shard_176.bin new file mode 100644 index 0000000000000000000000000000000000000000..71699aa2b1d550eb564e5404f8e73e6030b8cbb3 --- /dev/null +++ b/params_shard_176.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:365b9407fdcdff74d617edf00a7cfa5748db8620b4b0a7729477cdb276bb0031 +size 37748736 diff --git a/params_shard_177.bin b/params_shard_177.bin new file mode 100644 index 0000000000000000000000000000000000000000..b1035514a4ae1528e97c4d918ff684bff4af2900 --- /dev/null +++ b/params_shard_177.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ab510fe292c1e19c20f09a651f7b053f76df632a9e8f8ef320eb1b301192db8 +size 339738624 diff --git a/params_shard_178.bin b/params_shard_178.bin new file mode 100644 index 0000000000000000000000000000000000000000..6e374d3a447b05ff548b5fb6d0ea54d42113166a --- /dev/null +++ b/params_shard_178.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a4c46c345e5fc7b0989d802c7b624f658f9a511f81a105347e6c525eb5dbcaa9 +size 679477248 diff --git a/params_shard_179.bin b/params_shard_179.bin new file mode 100644 index 0000000000000000000000000000000000000000..5ab32dfb2eca7fafd03ff8fec603416601a5f39e --- /dev/null +++ b/params_shard_179.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:651b7b188fd5cd08ecd388363fce7097de56a9f64e1e5af38eaa7513f1b2a186 +size 75497472 diff --git a/params_shard_18.bin b/params_shard_18.bin new file mode 100644 index 0000000000000000000000000000000000000000..d63bd6a470cbc757d2fe44259f1d68f92a499a22 --- /dev/null +++ b/params_shard_18.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:28c664331be83972ab97f428736e1d0bea356cc4f7c3c1e48d0ffb8f90d78f00 +size 75497472 diff --git a/params_shard_180.bin b/params_shard_180.bin new file mode 100644 index 0000000000000000000000000000000000000000..c068fb3b44d55e1d1648e650ba4b37eddb3b55be --- /dev/null +++ b/params_shard_180.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7f9d1dc6bd7dee6cbf0859ca4398ba9c7d4514f93526e7ae4d95d523acabe2d3 +size 37748736 diff --git a/params_shard_181.bin b/params_shard_181.bin new file mode 100644 index 0000000000000000000000000000000000000000..71c3d7ea7a1081016ce905c58aca4ce5dca68725 --- /dev/null +++ b/params_shard_181.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3f368c05cc09466805cef5e9c10023af7cc3178d1ab0d2a1263dbe5b99b4cfef +size 679477248 diff --git a/params_shard_182.bin b/params_shard_182.bin new file mode 100644 index 0000000000000000000000000000000000000000..4b1698ee188dbee4412ac04d468b8a87ce24c1ff --- /dev/null +++ b/params_shard_182.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3abec30e45a457327609b0324d9d05156572c76e4ad2869cd550ddb6d6ae69b8 +size 75497472 diff --git a/params_shard_183.bin b/params_shard_183.bin new file mode 100644 index 0000000000000000000000000000000000000000..6532071fc14e14f061f044522d37547275b07b63 --- /dev/null +++ b/params_shard_183.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c7750a1441129dba68a80f645f9525d67f5a1472552f70a82019ed8a84eb860 +size 37748736 diff --git a/params_shard_184.bin b/params_shard_184.bin new file mode 100644 index 0000000000000000000000000000000000000000..47d6cbb760e0f6165edeeeecedd782acb1cda992 --- /dev/null +++ b/params_shard_184.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e8dc8d115a01b594edb5a434b62a1a66577a4c35283e2a05f2e851f4de2e42f7 +size 339738624 diff --git a/params_shard_185.bin b/params_shard_185.bin new file mode 100644 index 0000000000000000000000000000000000000000..c41739e37a800633af6242b4cc18ae90277023f5 --- /dev/null +++ b/params_shard_185.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:db1f0fbabf57bc6dddfb45bbc51886ae8481cca86209ce90b1847eb65a4347d0 +size 1704960 diff --git a/params_shard_19.bin b/params_shard_19.bin new file mode 100644 index 0000000000000000000000000000000000000000..624d8bf2a765f82a6c76f38a8347720b2e476b72 --- /dev/null +++ b/params_shard_19.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:031d3b880fd48e08810f3a61fdc2ac0d451748a40b986a77222edf4ea718cd81 +size 37748736 diff --git a/params_shard_2.bin b/params_shard_2.bin new file mode 100644 index 0000000000000000000000000000000000000000..ede88d5d6f3f8924619d942e888cd9b701368343 --- /dev/null +++ b/params_shard_2.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:346697bdccdcabe2487c8941c55464ce4dccf76421ab3b67d39559e44b65f767 +size 679477248 diff --git a/params_shard_20.bin b/params_shard_20.bin new file mode 100644 index 0000000000000000000000000000000000000000..667b34bf8c077e4aacd1823dfcd2a629ed4360dd --- /dev/null +++ b/params_shard_20.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f354437b04952c28223b4764a7368f56b0f685fd16a9c2e5b38ac863a0ecb833 +size 339738624 diff --git a/params_shard_21.bin b/params_shard_21.bin new file mode 100644 index 0000000000000000000000000000000000000000..2ddeadcfbc6e1dd17347ae0e4637d3a79e368ff6 --- /dev/null +++ b/params_shard_21.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61b191acb6f9424e95021eb0d19caabb3869aaf4e3b7a9e5b538aa7d806eea92 +size 679477248 diff --git a/params_shard_22.bin b/params_shard_22.bin new file mode 100644 index 0000000000000000000000000000000000000000..d3373223a9128abfe6a7b6ba4c4e838cbcd6875c --- /dev/null +++ b/params_shard_22.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aeba668669f1b0ad73b39586a545c0aaab6f064ba9c4e6f465657237bab507ba +size 75497472 diff --git a/params_shard_23.bin b/params_shard_23.bin new file mode 100644 index 0000000000000000000000000000000000000000..aa081d85d9b4cb044514874de92e58df7c577dbb --- /dev/null +++ b/params_shard_23.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0faaa59ef77bcda35164008a44b845f2c1f893a607d0bf6af794794301b62c24 +size 37748736 diff --git a/params_shard_24.bin b/params_shard_24.bin new file mode 100644 index 0000000000000000000000000000000000000000..9b359ee9ef9cc640c1cfb6717f2c6f5f60d47233 --- /dev/null +++ b/params_shard_24.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:57b289be7190eaf61bf735056d5a0fbb706442939e5fb7ceaa983a21138e5ac6 +size 339738624 diff --git a/params_shard_25.bin b/params_shard_25.bin new file mode 100644 index 0000000000000000000000000000000000000000..ba9695b0b4a7f63f4f99cc2922307ca0e4782173 --- /dev/null +++ b/params_shard_25.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e52e09e962ad377f96b8e9fb55ddc1a09a0ea8b771de1edb3dcfffee1882976d +size 679477248 diff --git a/params_shard_26.bin b/params_shard_26.bin new file mode 100644 index 0000000000000000000000000000000000000000..988d24457802316b106eccc1e30b93def03e4dc2 --- /dev/null +++ b/params_shard_26.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:934a554f78edbea0e4fc807699a4d577f2b43eafe7970343240b2b1964500ef0 +size 75497472 diff --git a/params_shard_27.bin b/params_shard_27.bin new file mode 100644 index 0000000000000000000000000000000000000000..f0afad01de487d34ccceaa28de500df045f97735 --- /dev/null +++ b/params_shard_27.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9f12e331a1cb8cb1cd62637d2a08c3148b654591f96a63b88a683ee78150611b +size 37748736 diff --git a/params_shard_28.bin b/params_shard_28.bin new file mode 100644 index 0000000000000000000000000000000000000000..75be6c00c387d7086182fde7a66183223162a10f --- /dev/null +++ b/params_shard_28.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0f6e76bff18ad3bb045a5be3660b93bedeac3efb8f886ec69e881a63212215f3 +size 75497472 diff --git a/params_shard_29.bin b/params_shard_29.bin new file mode 100644 index 0000000000000000000000000000000000000000..9d11ab475a9d73ac1c2ea0164869dcb69af0a849 --- /dev/null +++ b/params_shard_29.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:de2310429dc17f9abbe1a5cc0cc494048d8ae4dae2e6ab8952a178ac50a5ddb9 +size 37748736 diff --git a/params_shard_3.bin b/params_shard_3.bin new file mode 100644 index 0000000000000000000000000000000000000000..6475fd998330652e276f8e66796bcaf1fb492100 --- /dev/null +++ b/params_shard_3.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a76d3f4038f7e9c88a934cadacd9c69ad13fc582467d0c20dda4f1fb77108b9c +size 75497472 diff --git a/params_shard_30.bin b/params_shard_30.bin new file mode 100644 index 0000000000000000000000000000000000000000..f5ecbece2e54306b831587b886fa875bf55fbf6c --- /dev/null +++ b/params_shard_30.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad4110bbc509085f04186be09464427263aff78888612cbdf4cbe58644dab0cb +size 679477248 diff --git a/params_shard_31.bin b/params_shard_31.bin new file mode 100644 index 0000000000000000000000000000000000000000..89bcf19f1b617933440946a661ea5b80507d3fab --- /dev/null +++ b/params_shard_31.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a8b1a0ec5e53414731cfe7a31f4dfe2d7ddfb533cfb50bf0c857403cf126a764 +size 75497472 diff --git a/params_shard_32.bin b/params_shard_32.bin new file mode 100644 index 0000000000000000000000000000000000000000..e371c7241d358f33e628a0e84ac0f41fbdc43b65 --- /dev/null +++ b/params_shard_32.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a460d444173480639c40be5cc4b2403c9932fe99454ace9c2e426f0543b91b85 +size 37748736 diff --git a/params_shard_33.bin b/params_shard_33.bin new file mode 100644 index 0000000000000000000000000000000000000000..71097ba5a4058715d2cd75fc442adc06bb239acd --- /dev/null +++ b/params_shard_33.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4c306676c045c3750000dd0420488cff7b0f084950419933c3a56936d2d0005b +size 339738624 diff --git a/params_shard_34.bin b/params_shard_34.bin new file mode 100644 index 0000000000000000000000000000000000000000..f67e822f1102fe841160c0e397fda8987017ceaa --- /dev/null +++ b/params_shard_34.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8ab04580911075cb245e45a60978555d2421a2497ec085f4c411a95b9390e958 +size 679477248 diff --git a/params_shard_35.bin b/params_shard_35.bin new file mode 100644 index 0000000000000000000000000000000000000000..7352f55b013cacaaf3e1a298e492a8ff2ee0c11a --- /dev/null +++ b/params_shard_35.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c619b444a5c0bb43bf66d3b1664f48168a4f525b81718c7aeaa3d18c0c746b44 +size 339738624 diff --git a/params_shard_36.bin b/params_shard_36.bin new file mode 100644 index 0000000000000000000000000000000000000000..eb2c683b02070f074b7076ee1a5bf263cd3d031d --- /dev/null +++ b/params_shard_36.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f6439f133c341e9c13bad41f8649ff04202f6eab911b067d74a19cc1abac9e8f +size 679477248 diff --git a/params_shard_37.bin b/params_shard_37.bin new file mode 100644 index 0000000000000000000000000000000000000000..38dddf8ce85e012c46a84dac651f7fd03c2245b5 --- /dev/null +++ b/params_shard_37.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5e2302a8085102e8f832858324146f5660ad070f118113f4ebb8e107dcbe6df2 +size 75497472 diff --git a/params_shard_38.bin b/params_shard_38.bin new file mode 100644 index 0000000000000000000000000000000000000000..4ac7ec2902673c78d9478938b977236081baf652 --- /dev/null +++ b/params_shard_38.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:449ee485c5ff5f801eeb3150937bee89ff43b2c78f30ab8f7814eb4fda493750 +size 37748736 diff --git a/params_shard_39.bin b/params_shard_39.bin new file mode 100644 index 0000000000000000000000000000000000000000..1e50a963197b8454720c0dfc1e650b161282383e --- /dev/null +++ b/params_shard_39.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e7c5c41136dcce09c6ea887e1a891bc2ea7d1883c6529cd064f6744b710f2379 +size 339738624 diff --git a/params_shard_4.bin b/params_shard_4.bin new file mode 100644 index 0000000000000000000000000000000000000000..865bf6769be33a91bc721eb7a05a1783da403ba6 --- /dev/null +++ b/params_shard_4.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7c791a21a7d2ad3ea946ee5385d3f26cafc849ad57032a335626b9dde039ca9e +size 37748736 diff --git a/params_shard_40.bin b/params_shard_40.bin new file mode 100644 index 0000000000000000000000000000000000000000..ef8c1239c26c8eaeda3515716d0fcd181f2fd507 --- /dev/null +++ b/params_shard_40.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5479ed699bf9c3f203218e02e28d6551e8923cec83aed99645348d3e23f6b5f5 +size 679477248 diff --git a/params_shard_41.bin b/params_shard_41.bin new file mode 100644 index 0000000000000000000000000000000000000000..e84ff3d87e4388e8a68bf7e3e6ba4d1ece7b0a95 --- /dev/null +++ b/params_shard_41.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:faf98733dd07b46389ec5e3104c7c71b36104fda915725c390afe9d8a4601877 +size 75497472 diff --git a/params_shard_42.bin b/params_shard_42.bin new file mode 100644 index 0000000000000000000000000000000000000000..a89dcc9744d5525e6edf9cafbf0e14ed9595f22d --- /dev/null +++ b/params_shard_42.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c63092c3c6ad4ac71210ad8946567153a24805a661124188f7701a55a088d13 +size 37748736 diff --git a/params_shard_43.bin b/params_shard_43.bin new file mode 100644 index 0000000000000000000000000000000000000000..df78bcdd0d85d1fb38d0e226961070e1a0dcb3df --- /dev/null +++ b/params_shard_43.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:09d4ff55fada825d279e7f11e5c3d4c1aa9f9075cf35cd1f1efdbc3fedafe98d +size 339738624 diff --git a/params_shard_44.bin b/params_shard_44.bin new file mode 100644 index 0000000000000000000000000000000000000000..e8b1f7bc1427d9b235cb26eeb2493ac521689393 --- /dev/null +++ b/params_shard_44.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a348f3dc529d3714804ffb5acbfaefca597736cad7d17243631008ce5836ec14 +size 679477248 diff --git a/params_shard_45.bin b/params_shard_45.bin new file mode 100644 index 0000000000000000000000000000000000000000..bc98062b4d403fbd759e3aac92c40e70c4992a59 --- /dev/null +++ b/params_shard_45.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8885d99e8b7e91e100438f6a867d85372380ba2da52ab09ac2826c61cc2446f7 +size 75497472 diff --git a/params_shard_46.bin b/params_shard_46.bin new file mode 100644 index 0000000000000000000000000000000000000000..a296eeec961adad6453b28e5e25228e4a934b996 --- /dev/null +++ b/params_shard_46.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62c9bfbefc09ba8d0cb55d65355b46515ea9f13fdbfe3bb65da21db4b16bf922 +size 37748736 diff --git a/params_shard_47.bin b/params_shard_47.bin new file mode 100644 index 0000000000000000000000000000000000000000..b733a17e200a8c9e7673375f2f69445b3aa9ccd6 --- /dev/null +++ b/params_shard_47.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5544a3ec2f03bd02482f7109e2b56cafc5a69ed71046ac5e339e5deb9aa26c9a +size 339738624 diff --git a/params_shard_48.bin b/params_shard_48.bin new file mode 100644 index 0000000000000000000000000000000000000000..5c717bb76a2064ecafc72997f6fc66d9cb1aaaa6 --- /dev/null +++ b/params_shard_48.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a97dd33ba05c39156f315e06580404ea5b9e1d0849331c0a35196409655917a6 +size 679477248 diff --git a/params_shard_49.bin b/params_shard_49.bin new file mode 100644 index 0000000000000000000000000000000000000000..5ba64d4c42968a79e8e62783ca28329357b46965 --- /dev/null +++ b/params_shard_49.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a1cda48e529470059032d726130e4701153220ced9ba825a61de691a300a8848 +size 339738624 diff --git a/params_shard_5.bin b/params_shard_5.bin new file mode 100644 index 0000000000000000000000000000000000000000..34b40363dc4822b031d374f89f1ef763031def2a --- /dev/null +++ b/params_shard_5.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:407f41cec06b526d33b27e02257a91a62fe71fcc8a14f92956a5774de3c5b726 +size 339738624 diff --git a/params_shard_50.bin b/params_shard_50.bin new file mode 100644 index 0000000000000000000000000000000000000000..4d754ce4a15e1fc33baa04c1125b3c06e6fdf5c5 --- /dev/null +++ b/params_shard_50.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:67227930cd5341880cad9f233f88d22f9f0f7dff1ae90a2b644ceabc86f22248 +size 679477248 diff --git a/params_shard_51.bin b/params_shard_51.bin new file mode 100644 index 0000000000000000000000000000000000000000..7b5929b46f0a313b24df046aac76965af88859f5 --- /dev/null +++ b/params_shard_51.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:13d2db4a9f121966a0bb377c8342410e7b92974f3546e9411e6c20dfa2d839dd +size 75497472 diff --git a/params_shard_52.bin b/params_shard_52.bin new file mode 100644 index 0000000000000000000000000000000000000000..85fbf31f5d4d02a370658f7b29573190e30db27c --- /dev/null +++ b/params_shard_52.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6addb1f87f88d6d6f62bb7789d677de4a21046aeb5a2876f81376c46dd8c32f1 +size 37748736 diff --git a/params_shard_53.bin b/params_shard_53.bin new file mode 100644 index 0000000000000000000000000000000000000000..0bb07f0c9eaf4522e1d12c3b8cd36896038a3eed --- /dev/null +++ b/params_shard_53.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:65a956e87a7b9dfca09b90acc22113d7ad2c92f95e10fb18ae7a72cdd3bb83d2 +size 339738624 diff --git a/params_shard_54.bin b/params_shard_54.bin new file mode 100644 index 0000000000000000000000000000000000000000..e9b9631d9c0eb387f8962bb4b53fe0b9db054d4d --- /dev/null +++ b/params_shard_54.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9310930a146c7c737ec8d951f3eec2dc6dc77c81e6fc4033d1bdb541b53ecc41 +size 679477248 diff --git a/params_shard_55.bin b/params_shard_55.bin new file mode 100644 index 0000000000000000000000000000000000000000..65eacb0520b8be2e868d93b0d31c735a052b3a1b --- /dev/null +++ b/params_shard_55.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86b34b5fa3bfb6e2bf903b5cb0c17a7479607a6e0993932c12d4001c49cdcf8e +size 75497472 diff --git a/params_shard_56.bin b/params_shard_56.bin new file mode 100644 index 0000000000000000000000000000000000000000..42535fedaeb6557ec505158ea178e444436d8318 --- /dev/null +++ b/params_shard_56.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b20f1ce1a0e7a6ab694d0a359038d5f0628a6fc4675e2dc2775640cd988dc4c8 +size 37748736 diff --git a/params_shard_57.bin b/params_shard_57.bin new file mode 100644 index 0000000000000000000000000000000000000000..f116855b056be17f24e5e7467d5b235be2df721a --- /dev/null +++ b/params_shard_57.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e5b3430bb5908154296b90ce0d4138910989427d743a8820fe3b890deb4dff68 +size 339738624 diff --git a/params_shard_58.bin b/params_shard_58.bin new file mode 100644 index 0000000000000000000000000000000000000000..8e4a1307d5472411798f647e5007d34f08700994 --- /dev/null +++ b/params_shard_58.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8051b53f4492b87bc9e01ef765cb3a7bc7cb939e24e5dd18f30fb4c507d82d55 +size 679477248 diff --git a/params_shard_59.bin b/params_shard_59.bin new file mode 100644 index 0000000000000000000000000000000000000000..5fa778eb6677db532ebe80dd95a477442c51b6bf --- /dev/null +++ b/params_shard_59.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:77f451b8761705484a4c3ccddeb8deb62fcd10088a52444814ba363df1463877 +size 75497472 diff --git a/params_shard_6.bin b/params_shard_6.bin new file mode 100644 index 0000000000000000000000000000000000000000..91f727c5f55017a04d5650472b5532e42faab25b --- /dev/null +++ b/params_shard_6.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:016a4c2e52acd370add451ff36362e51ed73c9d2eb490247f78f68e22f1ff268 +size 679477248 diff --git a/params_shard_60.bin b/params_shard_60.bin new file mode 100644 index 0000000000000000000000000000000000000000..51db4c3f3c45f7d9b0641337b612ea9428ff827d --- /dev/null +++ b/params_shard_60.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:812f2c143e6115152c103226a0bea945d39b060d97f184af3c764c7640cdbd9e +size 37748736 diff --git a/params_shard_61.bin b/params_shard_61.bin new file mode 100644 index 0000000000000000000000000000000000000000..dfa51d41931ce4421d526f7b78010d00d4779218 --- /dev/null +++ b/params_shard_61.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6e34e3d856a01e9b3a9e62e6b9f5555d47d8a1297105e86bec2bf4118b6f1bd3 +size 679477248 diff --git a/params_shard_62.bin b/params_shard_62.bin new file mode 100644 index 0000000000000000000000000000000000000000..4bdf44c5fcead4e3c423bb444f01a8d8351e289c --- /dev/null +++ b/params_shard_62.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7af80b758dd110ea28c7b370a9d9c3acfa9dd2f3a30e9962d1d601ab16654db5 +size 75497472 diff --git a/params_shard_63.bin b/params_shard_63.bin new file mode 100644 index 0000000000000000000000000000000000000000..7a040f1a96ce1c3446192f71fb08904ccc31b7ee --- /dev/null +++ b/params_shard_63.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3e5c896d30e05d68cb5f1bb899438ff20bbaa0159e21a3ad921432c81998ffe1 +size 37748736 diff --git a/params_shard_64.bin b/params_shard_64.bin new file mode 100644 index 0000000000000000000000000000000000000000..fe3def972e1c5b266e52291c6b4d0c329586033a --- /dev/null +++ b/params_shard_64.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fa6a96e67f0c5218b99e028d8c1071d5b08db9d39fa6021b974e5e05a4af405b +size 339738624 diff --git a/params_shard_65.bin b/params_shard_65.bin new file mode 100644 index 0000000000000000000000000000000000000000..58a739e03051ae765fe3282d4241527b3fad6e9f --- /dev/null +++ b/params_shard_65.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:763d353babaa78298205abbc6d5ddfe636699d6d0b28101dafcf5fabcac938a2 +size 339738624 diff --git a/params_shard_66.bin b/params_shard_66.bin new file mode 100644 index 0000000000000000000000000000000000000000..58e79f05dbd59efa4080213603e8c2aea98a7573 --- /dev/null +++ b/params_shard_66.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6636700951a05d2886919440c100efe7066629d5fb7212c85dd7f90a3f92eb8b +size 679477248 diff --git a/params_shard_67.bin b/params_shard_67.bin new file mode 100644 index 0000000000000000000000000000000000000000..6b6cd96cfc48be906a75bdb196ba1902402b2a67 --- /dev/null +++ b/params_shard_67.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fd9cd44b1775e781a40207b122692c3c06d7a1df76af3d12afd09522eb975ad6 +size 75497472 diff --git a/params_shard_68.bin b/params_shard_68.bin new file mode 100644 index 0000000000000000000000000000000000000000..087f9a378a203f1202c2c52effdab37e202ca65b --- /dev/null +++ b/params_shard_68.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ab41d79a894ad7cc065cb0d6c1323a6e2479c55cede32805fd563c41247eed1f +size 37748736 diff --git a/params_shard_69.bin b/params_shard_69.bin new file mode 100644 index 0000000000000000000000000000000000000000..5ef86cd1e78e65870d460ff6874795e5714b5b6d --- /dev/null +++ b/params_shard_69.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e5dadf0b9c56d26aab2a0736a6d9d52a9df65777f09e533005c27b4b7de0b46c +size 339738624 diff --git a/params_shard_7.bin b/params_shard_7.bin new file mode 100644 index 0000000000000000000000000000000000000000..a34c8140eb5c9ea910c403f880035fa9e4e364d7 --- /dev/null +++ b/params_shard_7.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:48f330b149f60f86b24aa9443320cb34a7252c4fc45bc7bf6891639871e77b2e +size 75497472 diff --git a/params_shard_70.bin b/params_shard_70.bin new file mode 100644 index 0000000000000000000000000000000000000000..f8c39995bac546ccdf0f067ec1a9cc765463dad4 --- /dev/null +++ b/params_shard_70.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:53b6b177681bb99a3ced660679151fcfa1b07b6388941429f0bf6d70cad70cb3 +size 679477248 diff --git a/params_shard_71.bin b/params_shard_71.bin new file mode 100644 index 0000000000000000000000000000000000000000..11c2580d2a17e76dac1280bc49464b8db40bc5b1 --- /dev/null +++ b/params_shard_71.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:513d9daac8a43480f5422b6365fccffd16302d0c88d5be17ddd1a710101a4a2f +size 75497472 diff --git a/params_shard_72.bin b/params_shard_72.bin new file mode 100644 index 0000000000000000000000000000000000000000..d3ee4ea3c07f14cac655d2df1a2462e218fcc36b --- /dev/null +++ b/params_shard_72.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e831c06a497151a45c36f7de016c3e935f73a0a04d8f507d3bee39bb59db8783 +size 37748736 diff --git a/params_shard_73.bin b/params_shard_73.bin new file mode 100644 index 0000000000000000000000000000000000000000..ef629ef1dca6122ca536aa5b9a56312c39428e47 --- /dev/null +++ b/params_shard_73.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ee2ae29dd273f83b28ad82c97370890b8a53deaffd781efe1366ba6786f4b98 +size 339738624 diff --git a/params_shard_74.bin b/params_shard_74.bin new file mode 100644 index 0000000000000000000000000000000000000000..0a3abbdfa3e2398658e74a55eff6a7cb7bed1f95 --- /dev/null +++ b/params_shard_74.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:36cdb3a21a2d6e7258178a9262aaec4757ccec0f0da1abe79d1b227ac4dd5fb7 +size 679477248 diff --git a/params_shard_75.bin b/params_shard_75.bin new file mode 100644 index 0000000000000000000000000000000000000000..6a1e280ed4dbd043db76d915dc1e9535d93bd7eb --- /dev/null +++ b/params_shard_75.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0474691ccebef01ebed235895eb1518bfd9ccf86d536b7015c0a165d3f1ae1f8 +size 75497472 diff --git a/params_shard_76.bin b/params_shard_76.bin new file mode 100644 index 0000000000000000000000000000000000000000..9a14d4b0a31d4ac1ea053d4c9025f04113ee7538 --- /dev/null +++ b/params_shard_76.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7de94d720e0ba52c72282da72a606dfce198d4a2bb76d3fabc360c2196181f3b +size 37748736 diff --git a/params_shard_77.bin b/params_shard_77.bin new file mode 100644 index 0000000000000000000000000000000000000000..3f3851bbcf39e7d01ef73a62dde4bc20fcfe4902 --- /dev/null +++ b/params_shard_77.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d78bc81ff03188be8d48a32b854785246b0b5fd8869cddca942030ee1b5ef663 +size 679477248 diff --git a/params_shard_78.bin b/params_shard_78.bin new file mode 100644 index 0000000000000000000000000000000000000000..264cd26222eaf1716be9dc3a96e5b77340221312 --- /dev/null +++ b/params_shard_78.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:57d839824b5f79b468f26dfd5d7a589799824722248887413d670724e2bd158f +size 75497472 diff --git a/params_shard_79.bin b/params_shard_79.bin new file mode 100644 index 0000000000000000000000000000000000000000..f22d6aa419992a8b7530aae9954d1392ee5f9b1f --- /dev/null +++ b/params_shard_79.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d05989e555416cd7c13d40079cd477ec6080270f63850cce1eb68ec878867b22 +size 37748736 diff --git a/params_shard_8.bin b/params_shard_8.bin new file mode 100644 index 0000000000000000000000000000000000000000..619cffe99d591ec649a8185ea6a7e10b4b4780af --- /dev/null +++ b/params_shard_8.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2f21912709e8ed7a7d9e71e79bb7d5d7678e9511a5744a5d7cc1ac778ea103c7 +size 37748736 diff --git a/params_shard_80.bin b/params_shard_80.bin new file mode 100644 index 0000000000000000000000000000000000000000..1202fa89b68e3caae3190d3760d0ad1eb71e7f84 --- /dev/null +++ b/params_shard_80.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:895881d286481b5c03b6a2cadff7b080e3ca57d00eef9a3d38d1ee6b681c16b4 +size 339738624 diff --git a/params_shard_81.bin b/params_shard_81.bin new file mode 100644 index 0000000000000000000000000000000000000000..3800b7af04c5f550983101cc123a1ca07126909f --- /dev/null +++ b/params_shard_81.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8ce7fdd22b7c6ac618207fd47fe60a1f933292c1ce447adbf9e39497b13f20d8 +size 679477248 diff --git a/params_shard_82.bin b/params_shard_82.bin new file mode 100644 index 0000000000000000000000000000000000000000..c40fa4e1318b94edb86fc7bc39d070570e1227f9 --- /dev/null +++ b/params_shard_82.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:896e57aae67bb40202903246b6065a005766997abc6f6248f6495300d5423095 +size 339738624 diff --git a/params_shard_83.bin b/params_shard_83.bin new file mode 100644 index 0000000000000000000000000000000000000000..e3e3008bdbdcbfae25dba9d4eefad4178a31b32d --- /dev/null +++ b/params_shard_83.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:92b8f8388ce12b4b9f60bf9ce883f5a79e97b30b3b54a1375c96e88cd73981a5 +size 679477248 diff --git a/params_shard_84.bin b/params_shard_84.bin new file mode 100644 index 0000000000000000000000000000000000000000..53d5597793a26c81fea057853a9f5e908228fc3a --- /dev/null +++ b/params_shard_84.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9c3d85650c23624215322d32b5cb54a4d7f9e77aa9138bbfcb8fd2277ccfd27e +size 75497472 diff --git a/params_shard_85.bin b/params_shard_85.bin new file mode 100644 index 0000000000000000000000000000000000000000..0a744ad4c9dfe6c7309ccb4999091289dfed805c --- /dev/null +++ b/params_shard_85.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8942972e379a9f3c50e69cb36037ae9bcf17ccb318f095f9193d75da22d611db +size 37748736 diff --git a/params_shard_86.bin b/params_shard_86.bin new file mode 100644 index 0000000000000000000000000000000000000000..8929c13e9a28cf30fb25fb0dfdeab6d64b458d79 --- /dev/null +++ b/params_shard_86.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9642cd0a264eb7b2ef95ab5a04c0f5d473f848eaacc393f80161699054766357 +size 339738624 diff --git a/params_shard_87.bin b/params_shard_87.bin new file mode 100644 index 0000000000000000000000000000000000000000..a9d6df832b2de4c597d310929d9a04a58a8d526f --- /dev/null +++ b/params_shard_87.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e1ee69c4496c94bc43e812a7ef2af1d2ea9896950ff6fdcdd22115c3dec215db +size 679477248 diff --git a/params_shard_88.bin b/params_shard_88.bin new file mode 100644 index 0000000000000000000000000000000000000000..8ffbbc87fa8d43291894a7dd552f4dfaa0fe309d --- /dev/null +++ b/params_shard_88.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:429ae97f3d67d948aa37e55457c3612cfe860b6638944edd1d2d4b06cab0e3be +size 75497472 diff --git a/params_shard_89.bin b/params_shard_89.bin new file mode 100644 index 0000000000000000000000000000000000000000..e6a45dcd6dc7c2ec2c6269dce43aa72faf013bfc --- /dev/null +++ b/params_shard_89.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:841a8b664a313a4da7d7012ca3ac835176926d78a89215c3e4eedaef7866f9c5 +size 37748736 diff --git a/params_shard_9.bin b/params_shard_9.bin new file mode 100644 index 0000000000000000000000000000000000000000..37016541c8b13626a0d1c87b6354b7116768e6ae --- /dev/null +++ b/params_shard_9.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86aeeedb7294a3a09d1c8bb4c7e2fefa8879e0f0d06f0c98fe8e838219abf827 +size 75497472 diff --git a/params_shard_90.bin b/params_shard_90.bin new file mode 100644 index 0000000000000000000000000000000000000000..38e41e23f9524480a712c23849b03bf44de3176b --- /dev/null +++ b/params_shard_90.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:11c89d045aab9b82a72d9aa37255dc54665d6ac3f3ea05c9688e4b6043bc9d99 +size 339738624 diff --git a/params_shard_91.bin b/params_shard_91.bin new file mode 100644 index 0000000000000000000000000000000000000000..fe474a4db92cb5127f40fd89cfd204fc953f0e8f --- /dev/null +++ b/params_shard_91.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:afb2ff23f31b55b420e21746ed1d8c256fb467542a336710e36ae5d094257b3e +size 679477248 diff --git a/params_shard_92.bin b/params_shard_92.bin new file mode 100644 index 0000000000000000000000000000000000000000..abe367831d5f30a46b0ec39b53c9dd49a1e558dc --- /dev/null +++ b/params_shard_92.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a218ef16f92e09ffe61f0fc83db69e975b985515c646da34ace95c8cf33d6b3d +size 75497472 diff --git a/params_shard_93.bin b/params_shard_93.bin new file mode 100644 index 0000000000000000000000000000000000000000..a77c87e83c11b00af7b95f1ea633af5779413318 --- /dev/null +++ b/params_shard_93.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0c89954c7fd5ecf79371eef555f69cf41e8dcc7493e4aa4a9e047c7dfe0bd380 +size 37748736 diff --git a/params_shard_94.bin b/params_shard_94.bin new file mode 100644 index 0000000000000000000000000000000000000000..37dfe712781c12c0fb609a43b2989f39e9ff72fd --- /dev/null +++ b/params_shard_94.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:46dfa22ff906681e7e557c9277fbc1125737aa4d6a94680a01f3da84c84bd4db +size 75497472 diff --git a/params_shard_95.bin b/params_shard_95.bin new file mode 100644 index 0000000000000000000000000000000000000000..41d5eb446b597e23982e7c2c2701ec9d1692156d --- /dev/null +++ b/params_shard_95.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8025fd2c4082f6aff1282ac889e648072343d3ea3aec7741baa74ee4f1d4998f +size 37748736 diff --git a/params_shard_96.bin b/params_shard_96.bin new file mode 100644 index 0000000000000000000000000000000000000000..81204abdaf9766c9344009e18777eadf753a5d53 --- /dev/null +++ b/params_shard_96.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c23749253222daabba438df29051fd886726607d10e92bcc39d02bfab3cc992e +size 339738624 diff --git a/params_shard_97.bin b/params_shard_97.bin new file mode 100644 index 0000000000000000000000000000000000000000..4916b3f6a0dfaafde0621bc0c2cc749e279bee08 --- /dev/null +++ b/params_shard_97.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d24f2734eef66740b144cbd1e3ea6ceda5ecf6d5fa23e7bde80dbacd07e2313c +size 339738624 diff --git a/params_shard_98.bin b/params_shard_98.bin new file mode 100644 index 0000000000000000000000000000000000000000..c739a9b9c80b84d5702e0298e45074dd76517dc9 --- /dev/null +++ b/params_shard_98.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:68a6676ca5fbefe536fedbc719efde061b0bed595c11f41eb5b6d1cd6d73bb24 +size 679477248 diff --git a/params_shard_99.bin b/params_shard_99.bin new file mode 100644 index 0000000000000000000000000000000000000000..63d6f0c5de0a8fd3155b7b2d192f85f88c61701c --- /dev/null +++ b/params_shard_99.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8f0e5d7d5e8fcd51181118e06347a03f1644526de46ee9db22b3d8d2e5ccd83f +size 75497472 diff --git a/tokenizer.json b/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..5771f48b1e9b53a3865929ed27275c483186c9d7 --- /dev/null +++ b/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7da53ca29fb16f6b2489482fc0bc6a394162cdab14d12764a1755ebc583fea79 +size 17518525 diff --git a/tokenizer.model b/tokenizer.model new file mode 100644 index 0000000000000000000000000000000000000000..796efe9ab515c15e146ce7588e6d7b9b8134dbf8 --- /dev/null +++ b/tokenizer.model @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61a7b147390c64585d6c3543dd6fc636906c9af3865a5548f27f31aee1d4c8e2 +size 4241003 diff --git a/tokenizer_config.json b/tokenizer_config.json new file mode 100644 index 0000000000000000000000000000000000000000..0e9edce2d1a5e6a5b002865aeade7f475ecf087d --- /dev/null +++ b/tokenizer_config.json @@ -0,0 +1,1757 @@ +{ + "add_bos_token": true, + "add_eos_token": false, + "added_tokens_decoder": { + "0": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "1": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "2": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "3": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "4": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "5": { + "content": "<2mass>", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "6": { + "content": "[@BOS@]", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "7": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "8": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "9": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "10": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "11": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "12": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "13": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "14": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "15": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "16": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "17": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "18": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "19": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "20": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "21": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "22": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "23": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "24": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "25": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "26": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "27": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "28": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "29": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "30": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "31": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "32": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "33": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "34": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "35": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "36": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "37": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "38": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "39": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "40": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "41": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "42": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "43": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "44": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "45": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "46": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "47": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "48": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "49": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "50": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "51": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "52": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "53": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "54": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "55": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "56": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "57": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "58": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "59": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "60": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "61": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "62": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "63": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "64": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "65": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "66": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "67": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "68": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "69": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "70": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "71": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "72": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "73": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "74": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "75": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "76": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "77": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "78": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "79": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "80": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "81": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "82": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "83": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "84": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "85": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "86": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "87": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "88": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "89": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "90": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "91": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "92": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "93": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "94": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "95": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "96": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "97": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "98": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "99": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "100": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "101": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "102": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "103": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "104": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "105": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "106": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "107": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": true + }, + "108": { + "content": "\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "109": { + "content": "\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "110": { + "content": "\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "111": { + "content": "\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "112": { + "content": "\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "113": { + "content": "\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "114": { + "content": "\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "115": { + "content": "\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "116": { + "content": "\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "117": { + "content": "\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "118": { + "content": "\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "119": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "120": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "121": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "122": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "123": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "124": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "125": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "126": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "127": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "128": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "129": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "130": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "131": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "132": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "133": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "134": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "135": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "136": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "137": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "138": { + "content": "\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "139": { + "content": "▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "140": { + "content": "▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "141": { + "content": "▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "142": { + "content": "▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "143": { + "content": "▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "144": { + "content": "▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "145": { + "content": "▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "146": { + "content": "▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "147": { + "content": "▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "148": { + "content": "▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "149": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "150": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "151": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "152": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "153": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "154": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "155": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "156": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "157": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "158": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "159": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "160": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "161": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "162": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "163": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "164": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "165": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "166": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "167": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "168": { + "content": "▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "169": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "170": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "172": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "173": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "174": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "175": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "171": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "176": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "177": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "178": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "179": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "180": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "181": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "182": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "183": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "184": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "185": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "186": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "187": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "188": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "189": { + "content": "

", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "190": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "191": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "192": { + "content": "
", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "193": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "194": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "195": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "196": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "197": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "198": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "199": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "200": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "201": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "202": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "203": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "204": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "205": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "206": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "207": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "208": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "209": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "210": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "211": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "212": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "213": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "214": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "215": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + }, + "216": { + "content": "", + "lstrip": false, + "normalized": false, + "rstrip": false, + "single_word": false, + "special": false + } + }, + "additional_special_tokens": [ + "", + "" + ], + "bos_token": "", + "chat_template": "{{ bos_token }}{% if messages[0]['role'] == 'system' %}{{ raise_exception('System role not supported') }}{% endif %}{% for message in messages %}{% if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}{{ raise_exception('Conversation roles must alternate user/assistant/user/assistant/...') }}{% endif %}{% if (message['role'] == 'assistant') %}{% set role = 'model' %}{% else %}{% set role = message['role'] %}{% endif %}{{ '' + role + '\n' + message['content'] | trim + '\n' }}{% endfor %}{% if add_generation_prompt %}{{'model\n'}}{% endif %}", + "clean_up_tokenization_spaces": false, + "eos_token": "", + "model_max_length": 1000000000000000019884624838656, + "pad_token": "", + "sp_model_kwargs": {}, + "spaces_between_special_tokens": false, + "tokenizer_class": "GemmaTokenizer", + "unk_token": "", + "use_default_system_prompt": false +}