diff --git "a/ndarray-cache.json" "b/ndarray-cache.json" new file mode 100644--- /dev/null +++ "b/ndarray-cache.json" @@ -0,0 +1,10252 @@ +{ + "metadata": { + "ParamSize": 898, + "ParamBytes": 2952110912.0, + "BitsPerParam": 7.485917575658407 + }, + "records": [ + { + "dataPath": "params_shard_0.bin", + "format": "raw-shard", + "nbytes": 28214784, + "records": [ + { + "name": "llm.model.layers.36.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.36.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 4608 + }, + { + "name": "llm.model.layers.36.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6640128 + }, + { + "name": "llm.model.layers.36.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7469568 + }, + { + "name": "llm.model.layers.36.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8299008 + }, + { + "name": "llm.model.layers.36.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21570048 + }, + { + "name": "llm.model.layers.36.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23228928 + }, + { + "name": "llm.model.layers.36.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24887808 + }, + { + "name": "llm.model.layers.36.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 24892416 + }, + { + "name": "llm.model.layers.36.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 27546624 + }, + { + "name": "llm.model.layers.36.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 27878400 + }, + { + "name": "llm.model.layers.37.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 28210176 + } + ], + "md5sum": "35e881a0efe648d25103b9fa7e33067b" + }, + { + "dataPath": "params_shard_1.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.37.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.37.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.37.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.37.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.37.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.37.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.37.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.37.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "21f1eebdca43e049d6b3aa5e63e15e5e" + }, + { + "dataPath": "params_shard_2.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.37.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.37.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.37.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.37.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.37.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.38.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.38.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.38.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.38.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.38.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.38.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.38.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.38.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "fb73ea3fb64c204a020b307341493942" + }, + { + "dataPath": "params_shard_3.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.38.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.38.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.38.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.38.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.38.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.38.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.39.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.39.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.39.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.39.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "860d07237df0044ddcc09fafb2a8f22c" + }, + { + "dataPath": "params_shard_4.bin", + "format": "raw-shard", + "nbytes": 29882880, + "records": [ + { + "name": "llm.model.layers.39.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.39.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.39.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.39.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.39.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.39.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.39.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.39.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.39.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.39.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.norm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + }, + { + "name": "resampler.attn.q_proj.bias", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29869056 + }, + { + "name": "resampler.attn.k_proj.bias", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29873664 + }, + { + "name": "resampler.attn.v_proj.bias", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29878272 + } + ], + "md5sum": "613cf80845cbabf67bad7f564a8022a7" + }, + { + "dataPath": "params_shard_5.bin", + "format": "raw-shard", + "nbytes": 31855104, + "records": [ + { + "name": "resampler.attn.q_proj.weight", + "shape": [ + 2304, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10616832, + "byteOffset": 0 + }, + { + "name": "resampler.attn.k_proj.weight", + "shape": [ + 2304, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10616832, + "byteOffset": 10616832 + }, + { + "name": "resampler.attn.v_proj.weight", + "shape": [ + 2304, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10616832, + "byteOffset": 21233664 + }, + { + "name": "resampler.attn.out_proj.bias", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 31850496 + } + ], + "md5sum": "b70f08ce50cf05ab62a298950a385886" + }, + { + "dataPath": "params_shard_6.bin", + "format": "raw-shard", + "nbytes": 31880448, + "records": [ + { + "name": "resampler.attn.out_proj.weight", + "shape": [ + 2304, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10616832, + "byteOffset": 0 + }, + { + "name": "resampler.kv_proj.weight", + "shape": [ + 2304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 5308416, + "byteOffset": 10616832 + }, + { + "name": "resampler.ln_kv.bias", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 15925248 + }, + { + "name": "resampler.ln_kv.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 15929856 + }, + { + "name": "resampler.ln_post.bias", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 15934464 + }, + { + "name": "resampler.ln_post.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 15939072 + }, + { + "name": "resampler.ln_q.bias", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 15943680 + }, + { + "name": "resampler.ln_q.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 15948288 + }, + { + "name": "resampler.pos_embed_k", + "shape": [ + 1024, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4718592, + "byteOffset": 15952896 + }, + { + "name": "resampler.pos_embed", + "shape": [ + 64, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 294912, + "byteOffset": 20671488 + }, + { + "name": "resampler.proj", + "shape": [ + 2304, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 10616832, + "byteOffset": 20966400 + }, + { + "name": "resampler.query", + "shape": [ + 64, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 294912, + "byteOffset": 31583232 + }, + { + "name": "vpm.blocks.0.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 31878144 + } + ], + "md5sum": "c6c0bf99d5f816bd2e58612ea6e5a7c3" + }, + { + "dataPath": "params_shard_7.bin", + "format": "raw-shard", + "nbytes": 33140128, + "records": [ + { + "name": "vpm.blocks.0.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.0.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 2654208 + }, + { + "name": "vpm.blocks.0.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 2661120 + }, + { + "name": "vpm.blocks.0.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 10623744 + }, + { + "name": "vpm.blocks.0.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 10632352 + }, + { + "name": "vpm.blocks.0.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 20548768 + }, + { + "name": "vpm.blocks.0.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 20551072 + }, + { + "name": "vpm.blocks.0.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 30467488 + }, + { + "name": "vpm.blocks.0.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 30469792 + }, + { + "name": "vpm.blocks.0.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 30472096 + }, + { + "name": "vpm.blocks.0.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 30474400 + }, + { + "name": "vpm.blocks.1.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 30476704 + }, + { + "name": "vpm.blocks.1.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 30479008 + }, + { + "name": "vpm.blocks.1.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 33133216 + } + ], + "md5sum": "c3ad13cb76623ef5d880b2b9311dc535" + }, + { + "dataPath": "params_shard_8.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.1.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.1.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.1.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.1.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.1.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.1.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.1.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.1.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.1.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.10.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.10.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.10.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "45a5b2b91fa99add2a1743647e5dfcb0" + }, + { + "dataPath": "params_shard_9.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.10.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.10.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.10.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.10.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.10.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.10.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.10.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.10.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.10.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.11.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.11.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.11.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "7fcb2887fc163eddabf5bdcdc534b3ad" + }, + { + "dataPath": "params_shard_10.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.11.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.11.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.11.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.11.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.11.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.11.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.11.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.11.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.11.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.12.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.12.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.12.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "68ed719e489bf53061f9f1b0ec99b580" + }, + { + "dataPath": "params_shard_11.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.12.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.12.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.12.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.12.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.12.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.12.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.12.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.12.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.12.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.13.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.13.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.13.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "8edcd48dd209bd2dccfbb49897244fe9" + }, + { + "dataPath": "params_shard_12.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.13.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.13.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.13.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.13.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.13.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.13.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.13.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.13.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.13.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.14.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.14.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.14.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "4ae0f16883658cc23ef7ae87f4d8fc4d" + }, + { + "dataPath": "params_shard_13.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.14.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.14.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.14.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.14.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.14.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.14.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.14.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.14.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.14.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.15.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.15.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.15.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "ab50f345a4ac97461bf07c8408d87a6f" + }, + { + "dataPath": "params_shard_14.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.15.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.15.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.15.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.15.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.15.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.15.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.15.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.15.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.15.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.16.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.16.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.16.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "9b3bdcef0674bc4656ecc1fda38f39ed" + }, + { + "dataPath": "params_shard_15.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.16.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.16.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.16.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.16.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.16.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.16.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.16.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.16.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.16.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.17.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.17.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.17.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "0f260f95345406805c449ec965de283c" + }, + { + "dataPath": "params_shard_16.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.17.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.17.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.17.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.17.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.17.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.17.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.17.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.17.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.17.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.18.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.18.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.18.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "e1c9e9a7039ad2982e15f9402dd1a320" + }, + { + "dataPath": "params_shard_17.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.18.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.18.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.18.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.18.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.18.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.18.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.18.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.18.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.18.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.19.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.19.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.19.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "1c27e2bdfec0defdbb610b2c85551131" + }, + { + "dataPath": "params_shard_18.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.19.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.19.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.19.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.19.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.19.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.19.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.19.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.19.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.19.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.2.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.2.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.2.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "c27aa9cf5c3d850bd5b5f8373a4d11be" + }, + { + "dataPath": "params_shard_19.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.2.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.2.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.2.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.2.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.2.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.2.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.2.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.2.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.2.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.20.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.20.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.20.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "8dbcd4e41a6e8853e3907b32c2735c8f" + }, + { + "dataPath": "params_shard_20.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.20.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.20.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.20.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.20.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.20.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.20.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.20.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.20.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.20.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.21.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.21.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.21.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "2afb946a41f7e6a3e320814ed8f2f840" + }, + { + "dataPath": "params_shard_21.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.21.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.21.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.21.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.21.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.21.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.21.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.21.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.21.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.21.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.22.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.22.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.22.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "af821f2a3624447a51e713b02c51ccf9" + }, + { + "dataPath": "params_shard_22.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.22.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.22.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.22.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.22.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.22.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.22.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.22.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.22.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.22.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.23.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.23.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.23.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "dda805e13ee7fcad796f1e45d5e6e42c" + }, + { + "dataPath": "params_shard_23.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.23.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.23.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.23.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.23.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.23.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.23.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.23.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.23.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.23.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.24.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.24.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.24.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "045f93b29c0d66e29f7e42ff9783ac3c" + }, + { + "dataPath": "params_shard_24.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.24.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.24.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.24.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.24.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.24.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.24.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.24.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.24.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.24.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.25.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.25.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.25.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "eedb6dda6685e17f7a6b2c2eafd0f407" + }, + { + "dataPath": "params_shard_25.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.25.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.25.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.25.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.25.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.25.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.25.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.25.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.25.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.25.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.3.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.3.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.3.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "19dd1f2ac66cca0491cd50c845dd94eb" + }, + { + "dataPath": "params_shard_26.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.3.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.3.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.3.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.3.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.3.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.3.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.3.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.3.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.3.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.4.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.4.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.4.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "6788f4d1f7fa056b27ad996689ddf628" + }, + { + "dataPath": "params_shard_27.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.4.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.4.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.4.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.4.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.4.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.4.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.4.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.4.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.4.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.5.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.5.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.5.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "da49fda13457516aa8570c3443f3f872" + }, + { + "dataPath": "params_shard_28.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.5.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.5.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.5.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.5.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.5.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.5.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.5.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.5.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.5.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.6.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.6.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.6.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "6af5d8de40f68c73db16e9610289f318" + }, + { + "dataPath": "params_shard_29.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.6.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.6.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.6.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.6.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.6.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.6.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.6.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.6.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.6.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.7.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.7.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.7.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "0d4ceca3b257163c00dc1068bea5224f" + }, + { + "dataPath": "params_shard_30.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.7.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.7.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.7.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.7.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.7.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.7.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.7.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.7.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.7.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.8.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.8.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.8.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "287638c1f051e3824256437107a74204" + }, + { + "dataPath": "params_shard_31.bin", + "format": "raw-shard", + "nbytes": 30479008, + "records": [ + { + "name": "vpm.blocks.8.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.8.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.8.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.8.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.8.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.8.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.8.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.8.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.8.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.blocks.9.attn.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.blocks.9.attn.proj.weight", + "shape": [ + 1152, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 27817888 + }, + { + "name": "vpm.blocks.9.attn.qkv.bias", + "shape": [ + 3456 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 6912, + "byteOffset": 30472096 + } + ], + "md5sum": "b114610d1b36bf2553f674a0b8fab3fb" + }, + { + "dataPath": "params_shard_32.bin", + "format": "raw-shard", + "nbytes": 565645824, + "records": [ + { + "name": "llm.model.embed_tokens.weight", + "shape": [ + 122753, + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 565645824, + "byteOffset": 0 + } + ], + "md5sum": "02f59850bf732470b7c83d87210f4b79" + }, + { + "dataPath": "params_shard_33.bin", + "format": "raw-shard", + "nbytes": 31541152, + "records": [ + { + "name": "vpm.blocks.9.attn.qkv.weight", + "shape": [ + 3456, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "vpm.blocks.9.mlp.fc1.bias", + "shape": [ + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 8608, + "byteOffset": 7962624 + }, + { + "name": "vpm.blocks.9.mlp.fc1.weight", + "shape": [ + 4304, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 7971232 + }, + { + "name": "vpm.blocks.9.mlp.fc2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 17887648 + }, + { + "name": "vpm.blocks.9.mlp.fc2.weight", + "shape": [ + 1152, + 4304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 9916416, + "byteOffset": 17889952 + }, + { + "name": "vpm.blocks.9.norm1.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27806368 + }, + { + "name": "vpm.blocks.9.norm1.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27808672 + }, + { + "name": "vpm.blocks.9.norm2.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27810976 + }, + { + "name": "vpm.blocks.9.norm2.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27813280 + }, + { + "name": "vpm.norm.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27815584 + }, + { + "name": "vpm.norm.weight", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27817888 + }, + { + "name": "vpm.patch_embed.proj.bias", + "shape": [ + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2304, + "byteOffset": 27820192 + }, + { + "name": "vpm.patch_embed.proj.weight", + "shape": [ + 1152, + 3, + 14, + 14 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1354752, + "byteOffset": 27822496 + }, + { + "name": "vpm.pos_embed", + "shape": [ + 1, + 1024, + 1152 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 2359296, + "byteOffset": 29177248 + }, + { + "name": "llm.model.layers.0.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 31536544 + } + ], + "md5sum": "3b0465eb5dacf342c63d8433766ccfb3" + }, + { + "dataPath": "params_shard_34.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.0.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.0.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.0.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.0.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.0.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.0.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.0.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.0.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "0dbc5229b535940f8f1abb4a8ba99c7f" + }, + { + "dataPath": "params_shard_35.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.0.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.0.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.0.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.0.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.0.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.1.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.1.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.1.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.1.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.1.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.1.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.1.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.1.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "5b9312985ea77d7a9bdead27bf54ed44" + }, + { + "dataPath": "params_shard_36.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.1.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.1.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.1.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.1.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.1.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.1.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.10.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.10.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.10.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.10.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "3bc78c7b03657ecc6e3db691e11c8dd0" + }, + { + "dataPath": "params_shard_37.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.10.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.10.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.10.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.10.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.10.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.10.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.10.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.10.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.10.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.10.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.11.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "2eb9ae047f1d08451ca7b4371e959abb" + }, + { + "dataPath": "params_shard_38.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.11.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.11.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.11.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.11.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.11.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.11.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.11.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.11.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "bcde85dcfe5102b8a3b976c22d702528" + }, + { + "dataPath": "params_shard_39.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.11.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.11.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.11.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.11.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.11.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.12.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.12.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.12.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.12.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.12.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.12.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.12.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.12.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "7a7512c0da8b04112754123dd03c0a54" + }, + { + "dataPath": "params_shard_40.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.12.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.12.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.12.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.12.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.12.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.12.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.13.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.13.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.13.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.13.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "be9a5d488f8b31939678dc976702b803" + }, + { + "dataPath": "params_shard_41.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.13.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.13.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.13.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.13.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.13.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.13.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.13.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.13.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.13.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.13.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.14.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "ca4e619827037459ea9f8fe2154fd310" + }, + { + "dataPath": "params_shard_42.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.14.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.14.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.14.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.14.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.14.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.14.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.14.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.14.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "67d703e1bbf6501fb8213036ee9bd843" + }, + { + "dataPath": "params_shard_43.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.14.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.14.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.14.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.14.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.14.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.15.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.15.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.15.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.15.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.15.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.15.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.15.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.15.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "2ddf36644f25b8997edaee3186a463eb" + }, + { + "dataPath": "params_shard_44.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.15.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.15.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.15.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.15.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.15.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.15.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.16.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.16.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.16.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.16.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "e970d3419c57d33cae9353a42fceab1a" + }, + { + "dataPath": "params_shard_45.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.16.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.16.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.16.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.16.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.16.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.16.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.16.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.16.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.16.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.16.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.17.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "166b74ba4507e712827c3bcc015ec3c0" + }, + { + "dataPath": "params_shard_46.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.17.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.17.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.17.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.17.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.17.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.17.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.17.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.17.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "a69d268ece2807f675f4ba28cca72486" + }, + { + "dataPath": "params_shard_47.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.17.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.17.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.17.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.17.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.17.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.18.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.18.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.18.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.18.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.18.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.18.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.18.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.18.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "0d68b513081b20d4725b7e757df40d34" + }, + { + "dataPath": "params_shard_48.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.18.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.18.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.18.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.18.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.18.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.18.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.19.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.19.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.19.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.19.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "ee1c975ff22c3d9a92ac7dc743837e4f" + }, + { + "dataPath": "params_shard_49.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.19.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.19.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.19.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.19.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.19.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.19.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.19.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.19.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.19.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.19.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.2.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "e39f521a2d80808243ccbb45ffda7e59" + }, + { + "dataPath": "params_shard_50.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.2.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.2.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.2.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.2.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.2.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.2.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.2.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.2.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "05adac5518d4632c487f2b3636443685" + }, + { + "dataPath": "params_shard_51.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.2.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.2.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.2.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.2.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.2.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.20.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.20.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.20.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.20.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.20.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.20.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.20.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.20.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "b93523dc0644319d75b53e6e664621bf" + }, + { + "dataPath": "params_shard_52.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.20.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.20.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.20.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.20.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.20.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.20.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.21.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.21.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.21.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.21.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "86580fb013cda9b02b79b7284ece3f8e" + }, + { + "dataPath": "params_shard_53.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.21.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.21.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.21.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.21.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.21.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.21.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.21.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.21.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.21.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.21.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.22.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "39bed558a4fd2f42ef398c14bd5eddb5" + }, + { + "dataPath": "params_shard_54.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.22.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.22.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.22.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.22.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.22.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.22.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.22.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.22.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "50ea7c70ac913d7f9cac703a1ff80140" + }, + { + "dataPath": "params_shard_55.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.22.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.22.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.22.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.22.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.22.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.23.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.23.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.23.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.23.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.23.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.23.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.23.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.23.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "32c56da14c68670bff9e3bcec790fb7d" + }, + { + "dataPath": "params_shard_56.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.23.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.23.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.23.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.23.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.23.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.23.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.24.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.24.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.24.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.24.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "cc12333f48379a33000e4c51f1c0f850" + }, + { + "dataPath": "params_shard_57.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.24.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.24.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.24.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.24.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.24.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.24.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.24.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.24.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.24.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.24.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.25.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "1c475d0582c4ccfb1a31bf70370fc289" + }, + { + "dataPath": "params_shard_58.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.25.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.25.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.25.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.25.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.25.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.25.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.25.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.25.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "254dcb7dc16168dae1bcb49c47d612d4" + }, + { + "dataPath": "params_shard_59.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.25.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.25.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.25.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.25.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.25.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.26.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.26.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.26.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.26.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.26.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.26.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.26.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.26.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "4ccb623c56961d4a14127cf310d8a0f4" + }, + { + "dataPath": "params_shard_60.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.26.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.26.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.26.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.26.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.26.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.26.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.27.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.27.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.27.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.27.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "74a771618a793fce90c7b5db6f2fea5a" + }, + { + "dataPath": "params_shard_61.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.27.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.27.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.27.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.27.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.27.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.27.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.27.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.27.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.27.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.27.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.28.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "64ef568084ecbff1c611d3e41aba463c" + }, + { + "dataPath": "params_shard_62.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.28.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.28.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.28.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.28.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.28.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.28.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.28.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.28.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "be4a2b25cfe5b4c41eebd0f36efc81fd" + }, + { + "dataPath": "params_shard_63.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.28.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.28.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.28.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.28.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.28.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.29.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.29.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.29.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.29.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.29.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.29.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.29.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.29.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "6d00a96471ec6b2da46529dd222ac6c4" + }, + { + "dataPath": "params_shard_64.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.29.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.29.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.29.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.29.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.29.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.29.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.3.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.3.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.3.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.3.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "52bda64d796b202d4b5d524ee59a6542" + }, + { + "dataPath": "params_shard_65.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.3.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.3.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.3.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.3.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.3.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.3.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.3.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.3.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.3.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.3.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.30.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "970375b93a1bc2e9896545afb1efe41a" + }, + { + "dataPath": "params_shard_66.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.30.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.30.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.30.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.30.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.30.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.30.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.30.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.30.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "5f08202dfa4888bb19fc6d9fcd4b07cd" + }, + { + "dataPath": "params_shard_67.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.30.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.30.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.30.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.30.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.30.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.31.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.31.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.31.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.31.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.31.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.31.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.31.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.31.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "402d9cdb76cee45b19ee96b115848a5f" + }, + { + "dataPath": "params_shard_68.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.31.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.31.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.31.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.31.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.31.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.31.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.32.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.32.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.32.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.32.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "af333716a27060a77bd1309dfea682e4" + }, + { + "dataPath": "params_shard_69.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.32.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.32.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.32.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.32.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.32.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.32.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.32.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.32.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.32.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.32.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.33.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "7b74ba34b51889dc34948df6c504efcb" + }, + { + "dataPath": "params_shard_70.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.33.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.33.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.33.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.33.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.33.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.33.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.33.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.33.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "1884f7d17246c6c65d71f76b512c42cd" + }, + { + "dataPath": "params_shard_71.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.33.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.33.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.33.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.33.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.33.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.34.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.34.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.34.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.34.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.34.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.34.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.34.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.34.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "67bdd7015a394715f4b935865f07d5f0" + }, + { + "dataPath": "params_shard_72.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.34.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.34.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.34.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.34.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.34.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.34.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.35.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.35.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.35.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.35.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "ac02b462700a37127c6b7008c0ff9a86" + }, + { + "dataPath": "params_shard_73.bin", + "format": "raw-shard", + "nbytes": 29864448, + "records": [ + { + "name": "llm.model.layers.35.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.35.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.35.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.35.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.35.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.35.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.35.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.35.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.35.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.35.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + } + ], + "md5sum": "e62639dffa8480de096da424ded83d15" + }, + { + "dataPath": "params_shard_74.bin", + "format": "raw-shard", + "nbytes": 33182208, + "records": [ + { + "name": "llm.model.layers.36.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.36.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.36.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.4.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.4.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 9957888 + }, + { + "name": "llm.model.layers.4.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.4.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 17422848 + }, + { + "name": "llm.model.layers.4.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 18252288 + }, + { + "name": "llm.model.layers.4.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 31523328 + } + ], + "md5sum": "03502bcbe2e396c002bdcc55c55e1f64" + }, + { + "dataPath": "params_shard_75.bin", + "format": "raw-shard", + "nbytes": 23233536, + "records": [ + { + "name": "llm.model.layers.4.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.4.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 1658880 + }, + { + "name": "llm.model.layers.4.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 1663488 + }, + { + "name": "llm.model.layers.4.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 9626112 + }, + { + "name": "llm.model.layers.4.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 10621440 + }, + { + "name": "llm.model.layers.4.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 11616768 + }, + { + "name": "llm.model.layers.4.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 14270976 + }, + { + "name": "llm.model.layers.4.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 14602752 + }, + { + "name": "llm.model.layers.5.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 14934528 + }, + { + "name": "llm.model.layers.5.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 14939136 + }, + { + "name": "llm.model.layers.5.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 21574656 + }, + { + "name": "llm.model.layers.5.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 22404096 + } + ], + "md5sum": "a18b564c182246e9494b4f35776a17d9" + }, + { + "dataPath": "params_shard_76.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.5.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.5.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.5.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.5.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.5.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.5.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.5.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.5.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.5.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.5.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.6.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "fe555e656ca09a653dad6cd536eb2213" + }, + { + "dataPath": "params_shard_77.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.6.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.6.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.6.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.6.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.6.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.6.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.6.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.6.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "a3b3a6951a8987588d4d01f5e9bba1f6" + }, + { + "dataPath": "params_shard_78.bin", + "format": "raw-shard", + "nbytes": 30200832, + "records": [ + { + "name": "llm.model.layers.6.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.6.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.6.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.6.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.6.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + }, + { + "name": "llm.model.layers.7.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 5308416 + }, + { + "name": "llm.model.layers.7.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 5313024 + }, + { + "name": "llm.model.layers.7.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 11948544 + }, + { + "name": "llm.model.layers.7.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 12777984 + }, + { + "name": "llm.model.layers.7.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 13607424 + }, + { + "name": "llm.model.layers.7.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 26878464 + }, + { + "name": "llm.model.layers.7.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 28537344 + }, + { + "name": "llm.model.layers.7.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 30196224 + } + ], + "md5sum": "5fd870ae774eb8be84a3bd5a4a12b724" + }, + { + "dataPath": "params_shard_79.bin", + "format": "raw-shard", + "nbytes": 21570048, + "records": [ + { + "name": "llm.model.layers.7.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.7.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 7962624 + }, + { + "name": "llm.model.layers.7.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 8957952 + }, + { + "name": "llm.model.layers.7.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 9953280 + }, + { + "name": "llm.model.layers.7.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12607488 + }, + { + "name": "llm.model.layers.7.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 12939264 + }, + { + "name": "llm.model.layers.8.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.8.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 13275648 + }, + { + "name": "llm.model.layers.8.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 19911168 + }, + { + "name": "llm.model.layers.8.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 20740608 + } + ], + "md5sum": "1d0cd0d59fb651ddafe82ee35b787b0d" + }, + { + "dataPath": "params_shard_80.bin", + "format": "raw-shard", + "nbytes": 29869056, + "records": [ + { + "name": "llm.model.layers.8.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.8.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 13271040 + }, + { + "name": "llm.model.layers.8.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 14929920 + }, + { + "name": "llm.model.layers.8.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 16588800 + }, + { + "name": "llm.model.layers.8.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 16593408 + }, + { + "name": "llm.model.layers.8.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 24556032 + }, + { + "name": "llm.model.layers.8.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 25551360 + }, + { + "name": "llm.model.layers.8.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 26546688 + }, + { + "name": "llm.model.layers.8.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29200896 + }, + { + "name": "llm.model.layers.8.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 29532672 + }, + { + "name": "llm.model.layers.9.input_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 29864448 + } + ], + "md5sum": "d9559d4c1684bf9254309705b0e3240a" + }, + { + "dataPath": "params_shard_81.bin", + "format": "raw-shard", + "nbytes": 32850432, + "records": [ + { + "name": "llm.model.layers.9.mlp.down_proj.q_weight", + "shape": [ + 2304, + 720 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 6635520, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.9.mlp.down_proj.q_scale", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 6635520 + }, + { + "name": "llm.model.layers.9.mlp.down_proj.q_zero", + "shape": [ + 2304, + 180 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 829440, + "byteOffset": 7464960 + }, + { + "name": "llm.model.layers.9.mlp.gate_up_proj.q_weight", + "shape": [ + 11520, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 13271040, + "byteOffset": 8294400 + }, + { + "name": "llm.model.layers.9.mlp.gate_up_proj.q_scale", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 21565440 + }, + { + "name": "llm.model.layers.9.mlp.gate_up_proj.q_zero", + "shape": [ + 11520, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 1658880, + "byteOffset": 23224320 + }, + { + "name": "llm.model.layers.9.post_attention_layernorm.weight", + "shape": [ + 2304 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 4608, + "byteOffset": 24883200 + }, + { + "name": "llm.model.layers.9.self_attn.qkv_proj.q_weight", + "shape": [ + 6912, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 7962624, + "byteOffset": 24887808 + } + ], + "md5sum": "d2800904eaf9cbc7b0aaacfcd24ad75c" + }, + { + "dataPath": "params_shard_82.bin", + "format": "raw-shard", + "nbytes": 5308416, + "records": [ + { + "name": "llm.model.layers.9.self_attn.qkv_proj.q_scale", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 0 + }, + { + "name": "llm.model.layers.9.self_attn.qkv_proj.q_zero", + "shape": [ + 6912, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 995328, + "byteOffset": 995328 + }, + { + "name": "llm.model.layers.9.self_attn.o_proj.q_weight", + "shape": [ + 2304, + 288 + ], + "dtype": "uint32", + "format": "f32-to-bf16", + "nbytes": 2654208, + "byteOffset": 1990656 + }, + { + "name": "llm.model.layers.9.self_attn.o_proj.q_scale", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4644864 + }, + { + "name": "llm.model.layers.9.self_attn.o_proj.q_zero", + "shape": [ + 2304, + 72 + ], + "dtype": "float16", + "format": "f32-to-bf16", + "nbytes": 331776, + "byteOffset": 4976640 + } + ], + "md5sum": "38ee4d8b0f6b09524bae331f6c4f8c11" + } + ] +} \ No newline at end of file