Upload folder using huggingface_hub (#1)

- 761fa4952c8976d586e35d753b0fbebeaf4fd9781618ad04a51a90681faae75c (3ed47cd934260dac3318772e8b714f44eda79670)
- 4119ba73f3e8a0e7356050e75bdc0f63db8d9dbc970ec889cff2655a7b02ef7d (ea6915ec041921ec74dacdb4b9be07f58edb5ec4)
- fe1754b29eecd95f9b42bbc31c11e82a08d6bf063a5677575165bf1de46c870e (0429fd9ab78a08cb3147a9bfdce5e32c701456e6)
- 5a36974303df4b41c907eadd49fd006357205eb0d89689947dbff629e494f12e (a7bf65b367d55272e91c67a43157eeb599f89625)
- cd7024a965569bdbbc5d29fd54238e8db7edce8e1b7558ae5c39a5e0ce59d465 (c766d6d238c150ca056ef078da88872aabc78725)
- 2c65ea79c9abbcfc1cb94edbf118e2e83c3f7350e1d7ca87147713e70d0c25a4 (041635989dc5fde8d02744672a730a40625747cb)
- 5563c9d27edb29ef3f0d88b1ba55cdbbe350136e5e39be4830028029c8e3b165 (8a11fe36c839fce5dc3e476ec140d38bc575ede3)
- 05d87e3bce8a837fde5c58789a3a0fc4bcd7155d9a783c40515b823ea5d7123b (275038e9fdd58e21ca8245c650f6f063303fabb6)
- a58dac418dd99bd450e3c9714e6284ef3744f5326e04cb949b9b0e34f2b67384 (89c2b872fba5903dbb09d32fa8e9388bf085b5a7)

README.md

@@ -0,0 +1,40 @@
+---
+base_model: []
+library_name: transformers
+tags:
+- mergekit
+- merge
+
+---
+# lumikabra_behemoth_195B_v2
+
+This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
+
+## Merge Details
+### Merge Method
+
+This model was merged using the della_linear merge method using /workspace/models/schnapper79_lumikabra-195B_v0.3 as a base.
+
+### Models Merged
+
+The following models were included in the merge:
+* /workspace/merges/lumikabra_behemoth_195b
+
+### Configuration
+
+The following YAML configuration was used to produce this model:
+
+```yaml
+models:
+  - model: /workspace/merges/lumikabra_behemoth_195b
+    parameters:
+      weight: 0.5
+      density: 0.8    
+merge_method: della_linear
+base_model: /workspace/models/schnapper79_lumikabra-195B_v0.3
+parameters:
+  epsilon: 0.05
+  lambda: 1
+  int8_mask: true
+dtype: bfloat16
+```

config.json

@@ -0,0 +1,38 @@
+{
+    "_name_or_path": "/workspace/models/schnapper79_lumikabra-195B_v0.3",
+    "architectures": [
+        "MistralForCausalLM"
+    ],
+    "attention_dropout": 0.0,
+    "bos_token_id": 1,
+    "eos_token_id": 2,
+    "head_dim": 128,
+    "hidden_act": "silu",
+    "hidden_size": 12288,
+    "initializer_range": 0.02,
+    "intermediate_size": 28672,
+    "max_position_embeddings": 131072,
+    "model_type": "mistral",
+    "num_attention_heads": 96,
+    "num_hidden_layers": 140,
+    "num_key_value_heads": 8,
+    "rms_norm_eps": 1e-05,
+    "rope_theta": 1000000.0,
+    "sliding_window": null,
+    "tie_word_embeddings": false,
+    "torch_dtype": "bfloat16",
+    "transformers_version": "4.45.2",
+    "use_cache": true,
+    "vocab_size": 32768,
+    "quantization_config": {
+        "quant_method": "exl2",
+        "version": "0.2.3",
+        "bits": 2.7,
+        "head_bits": 6,
+        "calibration": {
+            "rows": 115,
+            "length": 2048,
+            "dataset": "(default)"
+        }
+    }
+}

mergekit_config.yml

@@ -0,0 +1,12 @@
+models:
+  - model: /workspace/merges/lumikabra_behemoth_195b
+    parameters:
+      weight: 0.5
+      density: 0.8    
+merge_method: della_linear
+base_model: /workspace/models/schnapper79_lumikabra-195B_v0.3
+parameters:
+  epsilon: 0.05
+  lambda: 1
+  int8_mask: true
+dtype: bfloat16

model.safetensors.index.json

@@ -0,0 +1 @@
+{"metadata": {"mergekit_version": "0.0.4.4", "total_size": 389171208192}, "weight_map": {"lm_head.weight": "model-00001-of-00081.safetensors", "model.embed_tokens.weight": "model-00001-of-00081.safetensors", "model.layers.0.input_layernorm.weight": "model-00001-of-00081.safetensors", "model.layers.0.mlp.down_proj.weight": "model-00001-of-00081.safetensors", "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00081.safetensors", "model.layers.0.mlp.up_proj.weight": "model-00001-of-00081.safetensors", "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00081.safetensors", "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00081.safetensors", "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00081.safetensors", "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00081.safetensors", "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00081.safetensors", "model.layers.1.input_layernorm.weight": "model-00001-of-00081.safetensors", "model.layers.1.mlp.down_proj.weight": "model-00002-of-00081.safetensors", "model.layers.1.mlp.gate_proj.weight": "model-00002-of-00081.safetensors", "model.layers.1.mlp.up_proj.weight": "model-00002-of-00081.safetensors", "model.layers.1.post_attention_layernorm.weight": "model-00002-of-00081.safetensors", "model.layers.1.self_attn.k_proj.weight": "model-00002-of-00081.safetensors", "model.layers.1.self_attn.o_proj.weight": "model-00002-of-00081.safetensors", "model.layers.1.self_attn.q_proj.weight": "model-00002-of-00081.safetensors", "model.layers.1.self_attn.v_proj.weight": "model-00002-of-00081.safetensors", "model.layers.10.input_layernorm.weight": "model-00002-of-00081.safetensors", "model.layers.10.mlp.down_proj.weight": "model-00002-of-00081.safetensors", "model.layers.10.mlp.gate_proj.weight": "model-00002-of-00081.safetensors", "model.layers.10.mlp.up_proj.weight": "model-00002-of-00081.safetensors", "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00081.safetensors", "model.layers.10.self_attn.k_proj.weight": "model-00002-of-00081.safetensors", "model.layers.10.self_attn.o_proj.weight": "model-00003-of-00081.safetensors", "model.layers.10.self_attn.q_proj.weight": "model-00003-of-00081.safetensors", "model.layers.10.self_attn.v_proj.weight": "model-00003-of-00081.safetensors", "model.layers.100.input_layernorm.weight": "model-00003-of-00081.safetensors", "model.layers.100.mlp.down_proj.weight": "model-00003-of-00081.safetensors", "model.layers.100.mlp.gate_proj.weight": "model-00003-of-00081.safetensors", "model.layers.100.mlp.up_proj.weight": "model-00003-of-00081.safetensors", "model.layers.100.post_attention_layernorm.weight": "model-00003-of-00081.safetensors", "model.layers.100.self_attn.k_proj.weight": "model-00003-of-00081.safetensors", "model.layers.100.self_attn.o_proj.weight": "model-00003-of-00081.safetensors", "model.layers.100.self_attn.q_proj.weight": "model-00003-of-00081.safetensors", "model.layers.100.self_attn.v_proj.weight": "model-00003-of-00081.safetensors", "model.layers.101.input_layernorm.weight": "model-00003-of-00081.safetensors", "model.layers.101.mlp.down_proj.weight": "model-00003-of-00081.safetensors", "model.layers.101.mlp.gate_proj.weight": "model-00003-of-00081.safetensors", "model.layers.101.mlp.up_proj.weight": "model-00004-of-00081.safetensors", "model.layers.101.post_attention_layernorm.weight": "model-00004-of-00081.safetensors", "model.layers.101.self_attn.k_proj.weight": "model-00004-of-00081.safetensors", "model.layers.101.self_attn.o_proj.weight": "model-00004-of-00081.safetensors", "model.layers.101.self_attn.q_proj.weight": "model-00004-of-00081.safetensors", "model.layers.101.self_attn.v_proj.weight": "model-00004-of-00081.safetensors", "model.layers.102.input_layernorm.weight": "model-00004-of-00081.safetensors", "model.layers.102.mlp.down_proj.weight": "model-00004-of-00081.safetensors", "model.layers.102.mlp.gate_proj.weight": "model-00004-of-00081.safetensors", "model.layers.102.mlp.up_proj.weight": "model-00004-of-00081.safetensors", "model.layers.102.post_attention_layernorm.weight": "model-00004-of-00081.safetensors", "model.layers.102.self_attn.k_proj.weight": "model-00004-of-00081.safetensors", "model.layers.102.self_attn.o_proj.weight": "model-00004-of-00081.safetensors", "model.layers.102.self_attn.q_proj.weight": "model-00004-of-00081.safetensors", "model.layers.102.self_attn.v_proj.weight": "model-00004-of-00081.safetensors", "model.layers.103.input_layernorm.weight": "model-00004-of-00081.safetensors", "model.layers.103.mlp.down_proj.weight": "model-00004-of-00081.safetensors", "model.layers.103.mlp.gate_proj.weight": "model-00005-of-00081.safetensors", "model.layers.103.mlp.up_proj.weight": "model-00005-of-00081.safetensors", "model.layers.103.post_attention_layernorm.weight": "model-00005-of-00081.safetensors", "model.layers.103.self_attn.k_proj.weight": "model-00005-of-00081.safetensors", "model.layers.103.self_attn.o_proj.weight": "model-00005-of-00081.safetensors", "model.layers.103.self_attn.q_proj.weight": "model-00005-of-00081.safetensors", "model.layers.103.self_attn.v_proj.weight": "model-00005-of-00081.safetensors", "model.layers.104.input_layernorm.weight": "model-00005-of-00081.safetensors", "model.layers.104.mlp.down_proj.weight": "model-00005-of-00081.safetensors", "model.layers.104.mlp.gate_proj.weight": "model-00005-of-00081.safetensors", "model.layers.104.mlp.up_proj.weight": "model-00005-of-00081.safetensors", "model.layers.104.post_attention_layernorm.weight": "model-00005-of-00081.safetensors", "model.layers.104.self_attn.k_proj.weight": "model-00005-of-00081.safetensors", "model.layers.104.self_attn.o_proj.weight": "model-00005-of-00081.safetensors", "model.layers.104.self_attn.q_proj.weight": "model-00005-of-00081.safetensors", "model.layers.104.self_attn.v_proj.weight": "model-00005-of-00081.safetensors", "model.layers.105.input_layernorm.weight": "model-00005-of-00081.safetensors", "model.layers.105.mlp.down_proj.weight": "model-00006-of-00081.safetensors", "model.layers.105.mlp.gate_proj.weight": "model-00006-of-00081.safetensors", "model.layers.105.mlp.up_proj.weight": "model-00006-of-00081.safetensors", "model.layers.105.post_attention_layernorm.weight": "model-00006-of-00081.safetensors", "model.layers.105.self_attn.k_proj.weight": "model-00006-of-00081.safetensors", "model.layers.105.self_attn.o_proj.weight": "model-00006-of-00081.safetensors", "model.layers.105.self_attn.q_proj.weight": "model-00006-of-00081.safetensors", "model.layers.105.self_attn.v_proj.weight": "model-00006-of-00081.safetensors", "model.layers.106.input_layernorm.weight": "model-00006-of-00081.safetensors", "model.layers.106.mlp.down_proj.weight": "model-00006-of-00081.safetensors", "model.layers.106.mlp.gate_proj.weight": "model-00006-of-00081.safetensors", "model.layers.106.mlp.up_proj.weight": "model-00006-of-00081.safetensors", "model.layers.106.post_attention_layernorm.weight": "model-00006-of-00081.safetensors", "model.layers.106.self_attn.k_proj.weight": "model-00006-of-00081.safetensors", "model.layers.106.self_attn.o_proj.weight": "model-00007-of-00081.safetensors", "model.layers.106.self_attn.q_proj.weight": "model-00007-of-00081.safetensors", "model.layers.106.self_attn.v_proj.weight": "model-00007-of-00081.safetensors", "model.layers.107.input_layernorm.weight": "model-00007-of-00081.safetensors", "model.layers.107.mlp.down_proj.weight": "model-00007-of-00081.safetensors", "model.layers.107.mlp.gate_proj.weight": "model-00007-of-00081.safetensors", "model.layers.107.mlp.up_proj.weight": "model-00007-of-00081.safetensors", "model.layers.107.post_attention_layernorm.weight": "model-00007-of-00081.safetensors", "model.layers.107.self_attn.k_proj.weight": "model-00007-of-00081.safetensors", "model.layers.107.self_attn.o_proj.weight": "model-00007-of-00081.safetensors", "model.layers.107.self_attn.q_proj.weight": "model-00007-of-00081.safetensors", "model.layers.107.self_attn.v_proj.weight": "model-00007-of-00081.safetensors", "model.layers.108.input_layernorm.weight": "model-00007-of-00081.safetensors", "model.layers.108.mlp.down_proj.weight": "model-00007-of-00081.safetensors", "model.layers.108.mlp.gate_proj.weight": "model-00007-of-00081.safetensors", "model.layers.108.mlp.up_proj.weight": "model-00008-of-00081.safetensors", "model.layers.108.post_attention_layernorm.weight": "model-00008-of-00081.safetensors", "model.layers.108.self_attn.k_proj.weight": "model-00008-of-00081.safetensors", "model.layers.108.self_attn.o_proj.weight": "model-00008-of-00081.safetensors", "model.layers.108.self_attn.q_proj.weight": "model-00008-of-00081.safetensors", "model.layers.108.self_attn.v_proj.weight": "model-00008-of-00081.safetensors", "model.layers.109.input_layernorm.weight": "model-00008-of-00081.safetensors", "model.layers.109.mlp.down_proj.weight": "model-00008-of-00081.safetensors", "model.layers.109.mlp.gate_proj.weight": "model-00008-of-00081.safetensors", "model.layers.109.mlp.up_proj.weight": "model-00008-of-00081.safetensors", "model.layers.109.post_attention_layernorm.weight": "model-00008-of-00081.safetensors", "model.layers.109.self_attn.k_proj.weight": "model-00008-of-00081.safetensors", "model.layers.109.self_attn.o_proj.weight": "model-00008-of-00081.safetensors", "model.layers.109.self_attn.q_proj.weight": "model-00008-of-00081.safetensors", "model.layers.109.self_attn.v_proj.weight": "model-00008-of-00081.safetensors", "model.layers.11.input_layernorm.weight": "model-00008-of-00081.safetensors", "model.layers.11.mlp.down_proj.weight": "model-00008-of-00081.safetensors", "model.layers.11.mlp.gate_proj.weight": "model-00009-of-00081.safetensors", "model.layers.11.mlp.up_proj.weight": "model-00009-of-00081.safetensors", "model.layers.11.post_attention_layernorm.weight": "model-00009-of-00081.safetensors", "model.layers.11.self_attn.k_proj.weight": "model-00009-of-00081.safetensors", "model.layers.11.self_attn.o_proj.weight": "model-00009-of-00081.safetensors", "model.layers.11.self_attn.q_proj.weight": "model-00009-of-00081.safetensors", "model.layers.11.self_attn.v_proj.weight": "model-00009-of-00081.safetensors", "model.layers.110.input_layernorm.weight": "model-00009-of-00081.safetensors", "model.layers.110.mlp.down_proj.weight": "model-00009-of-00081.safetensors", "model.layers.110.mlp.gate_proj.weight": "model-00009-of-00081.safetensors", "model.layers.110.mlp.up_proj.weight": "model-00009-of-00081.safetensors", "model.layers.110.post_attention_layernorm.weight": "model-00009-of-00081.safetensors", "model.layers.110.self_attn.k_proj.weight": "model-00009-of-00081.safetensors", "model.layers.110.self_attn.o_proj.weight": "model-00009-of-00081.safetensors", "model.layers.110.self_attn.q_proj.weight": "model-00009-of-00081.safetensors", "model.layers.110.self_attn.v_proj.weight": "model-00009-of-00081.safetensors", "model.layers.111.input_layernorm.weight": "model-00009-of-00081.safetensors", "model.layers.111.mlp.down_proj.weight": "model-00010-of-00081.safetensors", "model.layers.111.mlp.gate_proj.weight": "model-00010-of-00081.safetensors", "model.layers.111.mlp.up_proj.weight": "model-00010-of-00081.safetensors", "model.layers.111.post_attention_layernorm.weight": "model-00010-of-00081.safetensors", "model.layers.111.self_attn.k_proj.weight": "model-00010-of-00081.safetensors", "model.layers.111.self_attn.o_proj.weight": "model-00010-of-00081.safetensors", "model.layers.111.self_attn.q_proj.weight": "model-00010-of-00081.safetensors", "model.layers.111.self_attn.v_proj.weight": "model-00010-of-00081.safetensors", "model.layers.112.input_layernorm.weight": "model-00010-of-00081.safetensors", "model.layers.112.mlp.down_proj.weight": "model-00010-of-00081.safetensors", "model.layers.112.mlp.gate_proj.weight": "model-00010-of-00081.safetensors", "model.layers.112.mlp.up_proj.weight": "model-00010-of-00081.safetensors", "model.layers.112.post_attention_layernorm.weight": "model-00010-of-00081.safetensors", "model.layers.112.self_attn.k_proj.weight": "model-00010-of-00081.safetensors", "model.layers.112.self_attn.o_proj.weight": "model-00011-of-00081.safetensors", "model.layers.112.self_attn.q_proj.weight": "model-00011-of-00081.safetensors", "model.layers.112.self_attn.v_proj.weight": "model-00011-of-00081.safetensors", "model.layers.113.input_layernorm.weight": "model-00011-of-00081.safetensors", "model.layers.113.mlp.down_proj.weight": "model-00011-of-00081.safetensors", "model.layers.113.mlp.gate_proj.weight": "model-00011-of-00081.safetensors", "model.layers.113.mlp.up_proj.weight": "model-00011-of-00081.safetensors", "model.layers.113.post_attention_layernorm.weight": "model-00011-of-00081.safetensors", "model.layers.113.self_attn.k_proj.weight": "model-00011-of-00081.safetensors", "model.layers.113.self_attn.o_proj.weight": "model-00011-of-00081.safetensors", "model.layers.113.self_attn.q_proj.weight": "model-00011-of-00081.safetensors", "model.layers.113.self_attn.v_proj.weight": "model-00011-of-00081.safetensors", "model.layers.114.input_layernorm.weight": "model-00011-of-00081.safetensors", "model.layers.114.mlp.down_proj.weight": "model-00011-of-00081.safetensors", "model.layers.114.mlp.gate_proj.weight": "model-00011-of-00081.safetensors", "model.layers.114.mlp.up_proj.weight": "model-00012-of-00081.safetensors", "model.layers.114.post_attention_layernorm.weight": "model-00012-of-00081.safetensors", "model.layers.114.self_attn.k_proj.weight": "model-00012-of-00081.safetensors", "model.layers.114.self_attn.o_proj.weight": "model-00012-of-00081.safetensors", "model.layers.114.self_attn.q_proj.weight": "model-00012-of-00081.safetensors", "model.layers.114.self_attn.v_proj.weight": "model-00012-of-00081.safetensors", "model.layers.115.input_layernorm.weight": "model-00012-of-00081.safetensors", "model.layers.115.mlp.down_proj.weight": "model-00012-of-00081.safetensors", "model.layers.115.mlp.gate_proj.weight": "model-00012-of-00081.safetensors", "model.layers.115.mlp.up_proj.weight": "model-00012-of-00081.safetensors", "model.layers.115.post_attention_layernorm.weight": "model-00012-of-00081.safetensors", "model.layers.115.self_attn.k_proj.weight": "model-00012-of-00081.safetensors", "model.layers.115.self_attn.o_proj.weight": "model-00012-of-00081.safetensors", "model.layers.115.self_attn.q_proj.weight": "model-00012-of-00081.safetensors", "model.layers.115.self_attn.v_proj.weight": "model-00012-of-00081.safetensors", "model.layers.116.input_layernorm.weight": "model-00012-of-00081.safetensors", "model.layers.116.mlp.down_proj.weight": "model-00012-of-00081.safetensors", "model.layers.116.mlp.gate_proj.weight": "model-00013-of-00081.safetensors", "model.layers.116.mlp.up_proj.weight": "model-00013-of-00081.safetensors", "model.layers.116.post_attention_layernorm.weight": "model-00013-of-00081.safetensors", "model.layers.116.self_attn.k_proj.weight": "model-00013-of-00081.safetensors", "model.layers.116.self_attn.o_proj.weight": "model-00013-of-00081.safetensors", "model.layers.116.self_attn.q_proj.weight": "model-00013-of-00081.safetensors", "model.layers.116.self_attn.v_proj.weight": "model-00013-of-00081.safetensors", "model.layers.117.input_layernorm.weight": "model-00013-of-00081.safetensors", "model.layers.117.mlp.down_proj.weight": "model-00013-of-00081.safetensors", "model.layers.117.mlp.gate_proj.weight": "model-00013-of-00081.safetensors", "model.layers.117.mlp.up_proj.weight": "model-00013-of-00081.safetensors", "model.layers.117.post_attention_layernorm.weight": "model-00013-of-00081.safetensors", "model.layers.117.self_attn.k_proj.weight": "model-00013-of-00081.safetensors", "model.layers.117.self_attn.o_proj.weight": "model-00013-of-00081.safetensors", "model.layers.117.self_attn.q_proj.weight": "model-00013-of-00081.safetensors", "model.layers.117.self_attn.v_proj.weight": "model-00013-of-00081.safetensors", "model.layers.118.input_layernorm.weight": "model-00013-of-00081.safetensors", "model.layers.118.mlp.down_proj.weight": "model-00014-of-00081.safetensors", "model.layers.118.mlp.gate_proj.weight": "model-00014-of-00081.safetensors", "model.layers.118.mlp.up_proj.weight": "model-00014-of-00081.safetensors", "model.layers.118.post_attention_layernorm.weight": "model-00014-of-00081.safetensors", "model.layers.118.self_attn.k_proj.weight": "model-00014-of-00081.safetensors", "model.layers.118.self_attn.o_proj.weight": "model-00014-of-00081.safetensors", "model.layers.118.self_attn.q_proj.weight": "model-00014-of-00081.safetensors", "model.layers.118.self_attn.v_proj.weight": "model-00014-of-00081.safetensors", "model.layers.119.input_layernorm.weight": "model-00014-of-00081.safetensors", "model.layers.119.mlp.down_proj.weight": "model-00014-of-00081.safetensors", "model.layers.119.mlp.gate_proj.weight": "model-00014-of-00081.safetensors", "model.layers.119.mlp.up_proj.weight": "model-00014-of-00081.safetensors", "model.layers.119.post_attention_layernorm.weight": "model-00014-of-00081.safetensors", "model.layers.119.self_attn.k_proj.weight": "model-00014-of-00081.safetensors", "model.layers.119.self_attn.o_proj.weight": "model-00015-of-00081.safetensors", "model.layers.119.self_attn.q_proj.weight": "model-00015-of-00081.safetensors", "model.layers.119.self_attn.v_proj.weight": "model-00015-of-00081.safetensors", "model.layers.12.input_layernorm.weight": "model-00015-of-00081.safetensors", "model.layers.12.mlp.down_proj.weight": "model-00015-of-00081.safetensors", "model.layers.12.mlp.gate_proj.weight": "model-00015-of-00081.safetensors", "model.layers.12.mlp.up_proj.weight": "model-00015-of-00081.safetensors", "model.layers.12.post_attention_layernorm.weight": "model-00015-of-00081.safetensors", "model.layers.12.self_attn.k_proj.weight": "model-00015-of-00081.safetensors", "model.layers.12.self_attn.o_proj.weight": "model-00015-of-00081.safetensors", "model.layers.12.self_attn.q_proj.weight": "model-00015-of-00081.safetensors", "model.layers.12.self_attn.v_proj.weight": "model-00015-of-00081.safetensors", "model.layers.120.input_layernorm.weight": "model-00015-of-00081.safetensors", "model.layers.120.mlp.down_proj.weight": "model-00015-of-00081.safetensors", "model.layers.120.mlp.gate_proj.weight": "model-00015-of-00081.safetensors", "model.layers.120.mlp.up_proj.weight": "model-00016-of-00081.safetensors", "model.layers.120.post_attention_layernorm.weight": "model-00016-of-00081.safetensors", "model.layers.120.self_attn.k_proj.weight": "model-00016-of-00081.safetensors", "model.layers.120.self_attn.o_proj.weight": "model-00016-of-00081.safetensors", "model.layers.120.self_attn.q_proj.weight": "model-00016-of-00081.safetensors", "model.layers.120.self_attn.v_proj.weight": "model-00016-of-00081.safetensors", "model.layers.121.input_layernorm.weight": "model-00016-of-00081.safetensors", "model.layers.121.mlp.down_proj.weight": "model-00016-of-00081.safetensors", "model.layers.121.mlp.gate_proj.weight": "model-00016-of-00081.safetensors", "model.layers.121.mlp.up_proj.weight": "model-00016-of-00081.safetensors", "model.layers.121.post_attention_layernorm.weight": "model-00016-of-00081.safetensors", "model.layers.121.self_attn.k_proj.weight": "model-00016-of-00081.safetensors", "model.layers.121.self_attn.o_proj.weight": "model-00016-of-00081.safetensors", "model.layers.121.self_attn.q_proj.weight": "model-00016-of-00081.safetensors", "model.layers.121.self_attn.v_proj.weight": "model-00016-of-00081.safetensors", "model.layers.122.input_layernorm.weight": "model-00016-of-00081.safetensors", "model.layers.122.mlp.down_proj.weight": "model-00016-of-00081.safetensors", "model.layers.122.mlp.gate_proj.weight": "model-00017-of-00081.safetensors", "model.layers.122.mlp.up_proj.weight": "model-00017-of-00081.safetensors", "model.layers.122.post_attention_layernorm.weight": "model-00017-of-00081.safetensors", "model.layers.122.self_attn.k_proj.weight": "model-00017-of-00081.safetensors", "model.layers.122.self_attn.o_proj.weight": "model-00017-of-00081.safetensors", "model.layers.122.self_attn.q_proj.weight": "model-00017-of-00081.safetensors", "model.layers.122.self_attn.v_proj.weight": "model-00017-of-00081.safetensors", "model.layers.123.input_layernorm.weight": "model-00017-of-00081.safetensors", "model.layers.123.mlp.down_proj.weight": "model-00017-of-00081.safetensors", "model.layers.123.mlp.gate_proj.weight": "model-00017-of-00081.safetensors", "model.layers.123.mlp.up_proj.weight": "model-00017-of-00081.safetensors", "model.layers.123.post_attention_layernorm.weight": "model-00017-of-00081.safetensors", "model.layers.123.self_attn.k_proj.weight": "model-00017-of-00081.safetensors", "model.layers.123.self_attn.o_proj.weight": "model-00017-of-00081.safetensors", "model.layers.123.self_attn.q_proj.weight": "model-00017-of-00081.safetensors", "model.layers.123.self_attn.v_proj.weight": "model-00017-of-00081.safetensors", "model.layers.124.input_layernorm.weight": "model-00017-of-00081.safetensors", "model.layers.124.mlp.down_proj.weight": "model-00018-of-00081.safetensors", "model.layers.124.mlp.gate_proj.weight": "model-00018-of-00081.safetensors", "model.layers.124.mlp.up_proj.weight": "model-00018-of-00081.safetensors", "model.layers.124.post_attention_layernorm.weight": "model-00018-of-00081.safetensors", "model.layers.124.self_attn.k_proj.weight": "model-00018-of-00081.safetensors", "model.layers.124.self_attn.o_proj.weight": "model-00018-of-00081.safetensors", "model.layers.124.self_attn.q_proj.weight": "model-00018-of-00081.safetensors", "model.layers.124.self_attn.v_proj.weight": "model-00018-of-00081.safetensors", "model.layers.125.input_layernorm.weight": "model-00018-of-00081.safetensors", "model.layers.125.mlp.down_proj.weight": "model-00018-of-00081.safetensors", "model.layers.125.mlp.gate_proj.weight": "model-00018-of-00081.safetensors", "model.layers.125.mlp.up_proj.weight": "model-00018-of-00081.safetensors", "model.layers.125.post_attention_layernorm.weight": "model-00018-of-00081.safetensors", "model.layers.125.self_attn.k_proj.weight": "model-00018-of-00081.safetensors", "model.layers.125.self_attn.o_proj.weight": "model-00019-of-00081.safetensors", "model.layers.125.self_attn.q_proj.weight": "model-00019-of-00081.safetensors", "model.layers.125.self_attn.v_proj.weight": "model-00019-of-00081.safetensors", "model.layers.126.input_layernorm.weight": "model-00019-of-00081.safetensors", "model.layers.126.mlp.down_proj.weight": "model-00019-of-00081.safetensors", "model.layers.126.mlp.gate_proj.weight": "model-00019-of-00081.safetensors", "model.layers.126.mlp.up_proj.weight": "model-00019-of-00081.safetensors", "model.layers.126.post_attention_layernorm.weight": "model-00019-of-00081.safetensors", "model.layers.126.self_attn.k_proj.weight": "model-00019-of-00081.safetensors", "model.layers.126.self_attn.o_proj.weight": "model-00019-of-00081.safetensors", "model.layers.126.self_attn.q_proj.weight": "model-00019-of-00081.safetensors", "model.layers.126.self_attn.v_proj.weight": "model-00019-of-00081.safetensors", "model.layers.127.input_layernorm.weight": "model-00019-of-00081.safetensors", "model.layers.127.mlp.down_proj.weight": "model-00019-of-00081.safetensors", "model.layers.127.mlp.gate_proj.weight": "model-00019-of-00081.safetensors", "model.layers.127.mlp.up_proj.weight": "model-00020-of-00081.safetensors", "model.layers.127.post_attention_layernorm.weight": "model-00020-of-00081.safetensors", "model.layers.127.self_attn.k_proj.weight": "model-00020-of-00081.safetensors", "model.layers.127.self_attn.o_proj.weight": "model-00020-of-00081.safetensors", "model.layers.127.self_attn.q_proj.weight": "model-00020-of-00081.safetensors", "model.layers.127.self_attn.v_proj.weight": "model-00020-of-00081.safetensors", "model.layers.128.input_layernorm.weight": "model-00020-of-00081.safetensors", "model.layers.128.mlp.down_proj.weight": "model-00020-of-00081.safetensors", "model.layers.128.mlp.gate_proj.weight": "model-00020-of-00081.safetensors", "model.layers.128.mlp.up_proj.weight": "model-00020-of-00081.safetensors", "model.layers.128.post_attention_layernorm.weight": "model-00020-of-00081.safetensors", "model.layers.128.self_attn.k_proj.weight": "model-00020-of-00081.safetensors", "model.layers.128.self_attn.o_proj.weight": "model-00020-of-00081.safetensors", "model.layers.128.self_attn.q_proj.weight": "model-00020-of-00081.safetensors", "model.layers.128.self_attn.v_proj.weight": "model-00020-of-00081.safetensors", "model.layers.129.input_layernorm.weight": "model-00020-of-00081.safetensors", "model.layers.129.mlp.down_proj.weight": "model-00020-of-00081.safetensors", "model.layers.129.mlp.gate_proj.weight": "model-00021-of-00081.safetensors", "model.layers.129.mlp.up_proj.weight": "model-00021-of-00081.safetensors", "model.layers.129.post_attention_layernorm.weight": "model-00021-of-00081.safetensors", "model.layers.129.self_attn.k_proj.weight": "model-00021-of-00081.safetensors", "model.layers.129.self_attn.o_proj.weight": "model-00021-of-00081.safetensors", "model.layers.129.self_attn.q_proj.weight": "model-00021-of-00081.safetensors", "model.layers.129.self_attn.v_proj.weight": "model-00021-of-00081.safetensors", "model.layers.13.input_layernorm.weight": "model-00021-of-00081.safetensors", "model.layers.13.mlp.down_proj.weight": "model-00021-of-00081.safetensors", "model.layers.13.mlp.gate_proj.weight": "model-00021-of-00081.safetensors", "model.layers.13.mlp.up_proj.weight": "model-00021-of-00081.safetensors", "model.layers.13.post_attention_layernorm.weight": "model-00021-of-00081.safetensors", "model.layers.13.self_attn.k_proj.weight": "model-00021-of-00081.safetensors", "model.layers.13.self_attn.o_proj.weight": "model-00021-of-00081.safetensors", "model.layers.13.self_attn.q_proj.weight": "model-00021-of-00081.safetensors", "model.layers.13.self_attn.v_proj.weight": "model-00021-of-00081.safetensors", "model.layers.130.input_layernorm.weight": "model-00021-of-00081.safetensors", "model.layers.130.mlp.down_proj.weight": "model-00022-of-00081.safetensors", "model.layers.130.mlp.gate_proj.weight": "model-00022-of-00081.safetensors", "model.layers.130.mlp.up_proj.weight": "model-00022-of-00081.safetensors", "model.layers.130.post_attention_layernorm.weight": "model-00022-of-00081.safetensors", "model.layers.130.self_attn.k_proj.weight": "model-00022-of-00081.safetensors", "model.layers.130.self_attn.o_proj.weight": "model-00022-of-00081.safetensors", "model.layers.130.self_attn.q_proj.weight": "model-00022-of-00081.safetensors", "model.layers.130.self_attn.v_proj.weight": "model-00022-of-00081.safetensors", "model.layers.131.input_layernorm.weight": "model-00022-of-00081.safetensors", "model.layers.131.mlp.down_proj.weight": "model-00022-of-00081.safetensors", "model.layers.131.mlp.gate_proj.weight": "model-00022-of-00081.safetensors", "model.layers.131.mlp.up_proj.weight": "model-00022-of-00081.safetensors", "model.layers.131.post_attention_layernorm.weight": "model-00022-of-00081.safetensors", "model.layers.131.self_attn.k_proj.weight": "model-00022-of-00081.safetensors", "model.layers.131.self_attn.o_proj.weight": "model-00023-of-00081.safetensors", "model.layers.131.self_attn.q_proj.weight": "model-00023-of-00081.safetensors", "model.layers.131.self_attn.v_proj.weight": "model-00023-of-00081.safetensors", "model.layers.132.input_layernorm.weight": "model-00023-of-00081.safetensors", "model.layers.132.mlp.down_proj.weight": "model-00023-of-00081.safetensors", "model.layers.132.mlp.gate_proj.weight": "model-00023-of-00081.safetensors", "model.layers.132.mlp.up_proj.weight": "model-00023-of-00081.safetensors", "model.layers.132.post_attention_layernorm.weight": "model-00023-of-00081.safetensors", "model.layers.132.self_attn.k_proj.weight": "model-00023-of-00081.safetensors", "model.layers.132.self_attn.o_proj.weight": "model-00023-of-00081.safetensors", "model.layers.132.self_attn.q_proj.weight": "model-00023-of-00081.safetensors", "model.layers.132.self_attn.v_proj.weight": "model-00023-of-00081.safetensors", "model.layers.133.input_layernorm.weight": "model-00023-of-00081.safetensors", "model.layers.133.mlp.down_proj.weight": "model-00023-of-00081.safetensors", "model.layers.133.mlp.gate_proj.weight": "model-00023-of-00081.safetensors", "model.layers.133.mlp.up_proj.weight": "model-00024-of-00081.safetensors", "model.layers.133.post_attention_layernorm.weight": "model-00024-of-00081.safetensors", "model.layers.133.self_attn.k_proj.weight": "model-00024-of-00081.safetensors", "model.layers.133.self_attn.o_proj.weight": "model-00024-of-00081.safetensors", "model.layers.133.self_attn.q_proj.weight": "model-00024-of-00081.safetensors", "model.layers.133.self_attn.v_proj.weight": "model-00024-of-00081.safetensors", "model.layers.134.input_layernorm.weight": "model-00024-of-00081.safetensors", "model.layers.134.mlp.down_proj.weight": "model-00024-of-00081.safetensors", "model.layers.134.mlp.gate_proj.weight": "model-00024-of-00081.safetensors", "model.layers.134.mlp.up_proj.weight": "model-00024-of-00081.safetensors", "model.layers.134.post_attention_layernorm.weight": "model-00024-of-00081.safetensors", "model.layers.134.self_attn.k_proj.weight": "model-00024-of-00081.safetensors", "model.layers.134.self_attn.o_proj.weight": "model-00024-of-00081.safetensors", "model.layers.134.self_attn.q_proj.weight": "model-00024-of-00081.safetensors", "model.layers.134.self_attn.v_proj.weight": "model-00024-of-00081.safetensors", "model.layers.135.input_layernorm.weight": "model-00024-of-00081.safetensors", "model.layers.135.mlp.down_proj.weight": "model-00024-of-00081.safetensors", "model.layers.135.mlp.gate_proj.weight": "model-00025-of-00081.safetensors", "model.layers.135.mlp.up_proj.weight": "model-00025-of-00081.safetensors", "model.layers.135.post_attention_layernorm.weight": "model-00025-of-00081.safetensors", "model.layers.135.self_attn.k_proj.weight": "model-00025-of-00081.safetensors", "model.layers.135.self_attn.o_proj.weight": "model-00025-of-00081.safetensors", "model.layers.135.self_attn.q_proj.weight": "model-00025-of-00081.safetensors", "model.layers.135.self_attn.v_proj.weight": "model-00025-of-00081.safetensors", "model.layers.136.input_layernorm.weight": "model-00025-of-00081.safetensors", "model.layers.136.mlp.down_proj.weight": "model-00025-of-00081.safetensors", "model.layers.136.mlp.gate_proj.weight": "model-00025-of-00081.safetensors", "model.layers.136.mlp.up_proj.weight": "model-00025-of-00081.safetensors", "model.layers.136.post_attention_layernorm.weight": "model-00025-of-00081.safetensors", "model.layers.136.self_attn.k_proj.weight": "model-00025-of-00081.safetensors", "model.layers.136.self_attn.o_proj.weight": "model-00025-of-00081.safetensors", "model.layers.136.self_attn.q_proj.weight": "model-00025-of-00081.safetensors", "model.layers.136.self_attn.v_proj.weight": "model-00025-of-00081.safetensors", "model.layers.137.input_layernorm.weight": "model-00025-of-00081.safetensors", "model.layers.137.mlp.down_proj.weight": "model-00026-of-00081.safetensors", "model.layers.137.mlp.gate_proj.weight": "model-00026-of-00081.safetensors", "model.layers.137.mlp.up_proj.weight": "model-00026-of-00081.safetensors", "model.layers.137.post_attention_layernorm.weight": "model-00026-of-00081.safetensors", "model.layers.137.self_attn.k_proj.weight": "model-00026-of-00081.safetensors", "model.layers.137.self_attn.o_proj.weight": "model-00026-of-00081.safetensors", "model.layers.137.self_attn.q_proj.weight": "model-00026-of-00081.safetensors", "model.layers.137.self_attn.v_proj.weight": "model-00026-of-00081.safetensors", "model.layers.138.input_layernorm.weight": "model-00026-of-00081.safetensors", "model.layers.138.mlp.down_proj.weight": "model-00026-of-00081.safetensors", "model.layers.138.mlp.gate_proj.weight": "model-00026-of-00081.safetensors", "model.layers.138.mlp.up_proj.weight": "model-00026-of-00081.safetensors", "model.layers.138.post_attention_layernorm.weight": "model-00026-of-00081.safetensors", "model.layers.138.self_attn.k_proj.weight": "model-00026-of-00081.safetensors", "model.layers.138.self_attn.o_proj.weight": "model-00027-of-00081.safetensors", "model.layers.138.self_attn.q_proj.weight": "model-00027-of-00081.safetensors", "model.layers.138.self_attn.v_proj.weight": "model-00027-of-00081.safetensors", "model.layers.139.input_layernorm.weight": "model-00027-of-00081.safetensors", "model.layers.139.mlp.down_proj.weight": "model-00027-of-00081.safetensors", "model.layers.139.mlp.gate_proj.weight": "model-00027-of-00081.safetensors", "model.layers.139.mlp.up_proj.weight": "model-00027-of-00081.safetensors", "model.layers.139.post_attention_layernorm.weight": "model-00027-of-00081.safetensors", "model.layers.139.self_attn.k_proj.weight": "model-00027-of-00081.safetensors", "model.layers.139.self_attn.o_proj.weight": "model-00027-of-00081.safetensors", "model.layers.139.self_attn.q_proj.weight": "model-00027-of-00081.safetensors", "model.layers.139.self_attn.v_proj.weight": "model-00027-of-00081.safetensors", "model.layers.14.input_layernorm.weight": "model-00027-of-00081.safetensors", "model.layers.14.mlp.down_proj.weight": "model-00027-of-00081.safetensors", "model.layers.14.mlp.gate_proj.weight": "model-00027-of-00081.safetensors", "model.layers.14.mlp.up_proj.weight": "model-00028-of-00081.safetensors", "model.layers.14.post_attention_layernorm.weight": "model-00028-of-00081.safetensors", "model.layers.14.self_attn.k_proj.weight": "model-00028-of-00081.safetensors", "model.layers.14.self_attn.o_proj.weight": "model-00028-of-00081.safetensors", "model.layers.14.self_attn.q_proj.weight": "model-00028-of-00081.safetensors", "model.layers.14.self_attn.v_proj.weight": "model-00028-of-00081.safetensors", "model.layers.15.input_layernorm.weight": "model-00028-of-00081.safetensors", "model.layers.15.mlp.down_proj.weight": "model-00028-of-00081.safetensors", "model.layers.15.mlp.gate_proj.weight": "model-00028-of-00081.safetensors", "model.layers.15.mlp.up_proj.weight": "model-00028-of-00081.safetensors", "model.layers.15.post_attention_layernorm.weight": "model-00028-of-00081.safetensors", "model.layers.15.self_attn.k_proj.weight": "model-00028-of-00081.safetensors", "model.layers.15.self_attn.o_proj.weight": "model-00028-of-00081.safetensors", "model.layers.15.self_attn.q_proj.weight": "model-00028-of-00081.safetensors", "model.layers.15.self_attn.v_proj.weight": "model-00028-of-00081.safetensors", "model.layers.16.input_layernorm.weight": "model-00028-of-00081.safetensors", "model.layers.16.mlp.down_proj.weight": "model-00028-of-00081.safetensors", "model.layers.16.mlp.gate_proj.weight": "model-00029-of-00081.safetensors", "model.layers.16.mlp.up_proj.weight": "model-00029-of-00081.safetensors", "model.layers.16.post_attention_layernorm.weight": "model-00029-of-00081.safetensors", "model.layers.16.self_attn.k_proj.weight": "model-00029-of-00081.safetensors", "model.layers.16.self_attn.o_proj.weight": "model-00029-of-00081.safetensors", "model.layers.16.self_attn.q_proj.weight": "model-00029-of-00081.safetensors", "model.layers.16.self_attn.v_proj.weight": "model-00029-of-00081.safetensors", "model.layers.17.input_layernorm.weight": "model-00029-of-00081.safetensors", "model.layers.17.mlp.down_proj.weight": "model-00029-of-00081.safetensors", "model.layers.17.mlp.gate_proj.weight": "model-00029-of-00081.safetensors", "model.layers.17.mlp.up_proj.weight": "model-00029-of-00081.safetensors", "model.layers.17.post_attention_layernorm.weight": "model-00029-of-00081.safetensors", "model.layers.17.self_attn.k_proj.weight": "model-00029-of-00081.safetensors", "model.layers.17.self_attn.o_proj.weight": "model-00029-of-00081.safetensors", "model.layers.17.self_attn.q_proj.weight": "model-00029-of-00081.safetensors", "model.layers.17.self_attn.v_proj.weight": "model-00029-of-00081.safetensors", "model.layers.18.input_layernorm.weight": "model-00029-of-00081.safetensors", "model.layers.18.mlp.down_proj.weight": "model-00030-of-00081.safetensors", "model.layers.18.mlp.gate_proj.weight": "model-00030-of-00081.safetensors", "model.layers.18.mlp.up_proj.weight": "model-00030-of-00081.safetensors", "model.layers.18.post_attention_layernorm.weight": "model-00030-of-00081.safetensors", "model.layers.18.self_attn.k_proj.weight": "model-00030-of-00081.safetensors", "model.layers.18.self_attn.o_proj.weight": "model-00030-of-00081.safetensors", "model.layers.18.self_attn.q_proj.weight": "model-00030-of-00081.safetensors", "model.layers.18.self_attn.v_proj.weight": "model-00030-of-00081.safetensors", "model.layers.19.input_layernorm.weight": "model-00030-of-00081.safetensors", "model.layers.19.mlp.down_proj.weight": "model-00030-of-00081.safetensors", "model.layers.19.mlp.gate_proj.weight": "model-00030-of-00081.safetensors", "model.layers.19.mlp.up_proj.weight": "model-00030-of-00081.safetensors", "model.layers.19.post_attention_layernorm.weight": "model-00030-of-00081.safetensors", "model.layers.19.self_attn.k_proj.weight": "model-00030-of-00081.safetensors", "model.layers.19.self_attn.o_proj.weight": "model-00031-of-00081.safetensors", "model.layers.19.self_attn.q_proj.weight": "model-00031-of-00081.safetensors", "model.layers.19.self_attn.v_proj.weight": "model-00031-of-00081.safetensors", "model.layers.2.input_layernorm.weight": "model-00031-of-00081.safetensors", "model.layers.2.mlp.down_proj.weight": "model-00031-of-00081.safetensors", "model.layers.2.mlp.gate_proj.weight": "model-00031-of-00081.safetensors", "model.layers.2.mlp.up_proj.weight": "model-00031-of-00081.safetensors", "model.layers.2.post_attention_layernorm.weight": "model-00031-of-00081.safetensors", "model.layers.2.self_attn.k_proj.weight": "model-00031-of-00081.safetensors", "model.layers.2.self_attn.o_proj.weight": "model-00031-of-00081.safetensors", "model.layers.2.self_attn.q_proj.weight": "model-00031-of-00081.safetensors", "model.layers.2.self_attn.v_proj.weight": "model-00031-of-00081.safetensors", "model.layers.20.input_layernorm.weight": "model-00031-of-00081.safetensors", "model.layers.20.mlp.down_proj.weight": "model-00031-of-00081.safetensors", "model.layers.20.mlp.gate_proj.weight": "model-00031-of-00081.safetensors", "model.layers.20.mlp.up_proj.weight": "model-00032-of-00081.safetensors", "model.layers.20.post_attention_layernorm.weight": "model-00032-of-00081.safetensors", "model.layers.20.self_attn.k_proj.weight": "model-00032-of-00081.safetensors", "model.layers.20.self_attn.o_proj.weight": "model-00032-of-00081.safetensors", "model.layers.20.self_attn.q_proj.weight": "model-00032-of-00081.safetensors", "model.layers.20.self_attn.v_proj.weight": "model-00032-of-00081.safetensors", "model.layers.21.input_layernorm.weight": "model-00032-of-00081.safetensors", "model.layers.21.mlp.down_proj.weight": "model-00032-of-00081.safetensors", "model.layers.21.mlp.gate_proj.weight": "model-00032-of-00081.safetensors", "model.layers.21.mlp.up_proj.weight": "model-00032-of-00081.safetensors", "model.layers.21.post_attention_layernorm.weight": "model-00032-of-00081.safetensors", "model.layers.21.self_attn.k_proj.weight": "model-00032-of-00081.safetensors", "model.layers.21.self_attn.o_proj.weight": "model-00032-of-00081.safetensors", "model.layers.21.self_attn.q_proj.weight": "model-00032-of-00081.safetensors", "model.layers.21.self_attn.v_proj.weight": "model-00032-of-00081.safetensors", "model.layers.22.input_layernorm.weight": "model-00032-of-00081.safetensors", "model.layers.22.mlp.down_proj.weight": "model-00032-of-00081.safetensors", "model.layers.22.mlp.gate_proj.weight": "model-00033-of-00081.safetensors", "model.layers.22.mlp.up_proj.weight": "model-00033-of-00081.safetensors", "model.layers.22.post_attention_layernorm.weight": "model-00033-of-00081.safetensors", "model.layers.22.self_attn.k_proj.weight": "model-00033-of-00081.safetensors", "model.layers.22.self_attn.o_proj.weight": "model-00033-of-00081.safetensors", "model.layers.22.self_attn.q_proj.weight": "model-00033-of-00081.safetensors", "model.layers.22.self_attn.v_proj.weight": "model-00033-of-00081.safetensors", "model.layers.23.input_layernorm.weight": "model-00033-of-00081.safetensors", "model.layers.23.mlp.down_proj.weight": "model-00033-of-00081.safetensors", "model.layers.23.mlp.gate_proj.weight": "model-00033-of-00081.safetensors", "model.layers.23.mlp.up_proj.weight": "model-00033-of-00081.safetensors", "model.layers.23.post_attention_layernorm.weight": "model-00033-of-00081.safetensors", "model.layers.23.self_attn.k_proj.weight": "model-00033-of-00081.safetensors", "model.layers.23.self_attn.o_proj.weight": "model-00033-of-00081.safetensors", "model.layers.23.self_attn.q_proj.weight": "model-00033-of-00081.safetensors", "model.layers.23.self_attn.v_proj.weight": "model-00033-of-00081.safetensors", "model.layers.24.input_layernorm.weight": "model-00033-of-00081.safetensors", "model.layers.24.mlp.down_proj.weight": "model-00034-of-00081.safetensors", "model.layers.24.mlp.gate_proj.weight": "model-00034-of-00081.safetensors", "model.layers.24.mlp.up_proj.weight": "model-00034-of-00081.safetensors", "model.layers.24.post_attention_layernorm.weight": "model-00034-of-00081.safetensors", "model.layers.24.self_attn.k_proj.weight": "model-00034-of-00081.safetensors", "model.layers.24.self_attn.o_proj.weight": "model-00034-of-00081.safetensors", "model.layers.24.self_attn.q_proj.weight": "model-00034-of-00081.safetensors", "model.layers.24.self_attn.v_proj.weight": "model-00034-of-00081.safetensors", "model.layers.25.input_layernorm.weight": "model-00034-of-00081.safetensors", "model.layers.25.mlp.down_proj.weight": "model-00034-of-00081.safetensors", "model.layers.25.mlp.gate_proj.weight": "model-00034-of-00081.safetensors", "model.layers.25.mlp.up_proj.weight": "model-00034-of-00081.safetensors", "model.layers.25.post_attention_layernorm.weight": "model-00034-of-00081.safetensors", "model.layers.25.self_attn.k_proj.weight": "model-00034-of-00081.safetensors", "model.layers.25.self_attn.o_proj.weight": "model-00035-of-00081.safetensors", "model.layers.25.self_attn.q_proj.weight": "model-00035-of-00081.safetensors", "model.layers.25.self_attn.v_proj.weight": "model-00035-of-00081.safetensors", "model.layers.26.input_layernorm.weight": "model-00035-of-00081.safetensors", "model.layers.26.mlp.down_proj.weight": "model-00035-of-00081.safetensors", "model.layers.26.mlp.gate_proj.weight": "model-00035-of-00081.safetensors", "model.layers.26.mlp.up_proj.weight": "model-00035-of-00081.safetensors", "model.layers.26.post_attention_layernorm.weight": "model-00035-of-00081.safetensors", "model.layers.26.self_attn.k_proj.weight": "model-00035-of-00081.safetensors", "model.layers.26.self_attn.o_proj.weight": "model-00035-of-00081.safetensors", "model.layers.26.self_attn.q_proj.weight": "model-00035-of-00081.safetensors", "model.layers.26.self_attn.v_proj.weight": "model-00035-of-00081.safetensors", "model.layers.27.input_layernorm.weight": "model-00035-of-00081.safetensors", "model.layers.27.mlp.down_proj.weight": "model-00035-of-00081.safetensors", "model.layers.27.mlp.gate_proj.weight": "model-00035-of-00081.safetensors", "model.layers.27.mlp.up_proj.weight": "model-00036-of-00081.safetensors", "model.layers.27.post_attention_layernorm.weight": "model-00036-of-00081.safetensors", "model.layers.27.self_attn.k_proj.weight": "model-00036-of-00081.safetensors", "model.layers.27.self_attn.o_proj.weight": "model-00036-of-00081.safetensors", "model.layers.27.self_attn.q_proj.weight": "model-00036-of-00081.safetensors", "model.layers.27.self_attn.v_proj.weight": "model-00036-of-00081.safetensors", "model.layers.28.input_layernorm.weight": "model-00036-of-00081.safetensors", "model.layers.28.mlp.down_proj.weight": "model-00036-of-00081.safetensors", "model.layers.28.mlp.gate_proj.weight": "model-00036-of-00081.safetensors", "model.layers.28.mlp.up_proj.weight": "model-00036-of-00081.safetensors", "model.layers.28.post_attention_layernorm.weight": "model-00036-of-00081.safetensors", "model.layers.28.self_attn.k_proj.weight": "model-00036-of-00081.safetensors", "model.layers.28.self_attn.o_proj.weight": "model-00036-of-00081.safetensors", "model.layers.28.self_attn.q_proj.weight": "model-00036-of-00081.safetensors", "model.layers.28.self_attn.v_proj.weight": "model-00036-of-00081.safetensors", "model.layers.29.input_layernorm.weight": "model-00036-of-00081.safetensors", "model.layers.29.mlp.down_proj.weight": "model-00036-of-00081.safetensors", "model.layers.29.mlp.gate_proj.weight": "model-00037-of-00081.safetensors", "model.layers.29.mlp.up_proj.weight": "model-00037-of-00081.safetensors", "model.layers.29.post_attention_layernorm.weight": "model-00037-of-00081.safetensors", "model.layers.29.self_attn.k_proj.weight": "model-00037-of-00081.safetensors", "model.layers.29.self_attn.o_proj.weight": "model-00037-of-00081.safetensors", "model.layers.29.self_attn.q_proj.weight": "model-00037-of-00081.safetensors", "model.layers.29.self_attn.v_proj.weight": "model-00037-of-00081.safetensors", "model.layers.3.input_layernorm.weight": "model-00037-of-00081.safetensors", "model.layers.3.mlp.down_proj.weight": "model-00037-of-00081.safetensors", "model.layers.3.mlp.gate_proj.weight": "model-00037-of-00081.safetensors", "model.layers.3.mlp.up_proj.weight": "model-00037-of-00081.safetensors", "model.layers.3.post_attention_layernorm.weight": "model-00037-of-00081.safetensors", "model.layers.3.self_attn.k_proj.weight": "model-00037-of-00081.safetensors", "model.layers.3.self_attn.o_proj.weight": "model-00037-of-00081.safetensors", "model.layers.3.self_attn.q_proj.weight": "model-00037-of-00081.safetensors", "model.layers.3.self_attn.v_proj.weight": "model-00037-of-00081.safetensors", "model.layers.30.input_layernorm.weight": "model-00037-of-00081.safetensors", "model.layers.30.mlp.down_proj.weight": "model-00038-of-00081.safetensors", "model.layers.30.mlp.gate_proj.weight": "model-00038-of-00081.safetensors", "model.layers.30.mlp.up_proj.weight": "model-00038-of-00081.safetensors", "model.layers.30.post_attention_layernorm.weight": "model-00038-of-00081.safetensors", "model.layers.30.self_attn.k_proj.weight": "model-00038-of-00081.safetensors", "model.layers.30.self_attn.o_proj.weight": "model-00038-of-00081.safetensors", "model.layers.30.self_attn.q_proj.weight": "model-00038-of-00081.safetensors", "model.layers.30.self_attn.v_proj.weight": "model-00038-of-00081.safetensors", "model.layers.31.input_layernorm.weight": "model-00038-of-00081.safetensors", "model.layers.31.mlp.down_proj.weight": "model-00038-of-00081.safetensors", "model.layers.31.mlp.gate_proj.weight": "model-00038-of-00081.safetensors", "model.layers.31.mlp.up_proj.weight": "model-00038-of-00081.safetensors", "model.layers.31.post_attention_layernorm.weight": "model-00038-of-00081.safetensors", "model.layers.31.self_attn.k_proj.weight": "model-00038-of-00081.safetensors", "model.layers.31.self_attn.o_proj.weight": "model-00039-of-00081.safetensors", "model.layers.31.self_attn.q_proj.weight": "model-00039-of-00081.safetensors", "model.layers.31.self_attn.v_proj.weight": "model-00039-of-00081.safetensors", "model.layers.32.input_layernorm.weight": "model-00039-of-00081.safetensors", "model.layers.32.mlp.down_proj.weight": "model-00039-of-00081.safetensors", "model.layers.32.mlp.gate_proj.weight": "model-00039-of-00081.safetensors", "model.layers.32.mlp.up_proj.weight": "model-00039-of-00081.safetensors", "model.layers.32.post_attention_layernorm.weight": "model-00039-of-00081.safetensors", "model.layers.32.self_attn.k_proj.weight": "model-00039-of-00081.safetensors", "model.layers.32.self_attn.o_proj.weight": "model-00039-of-00081.safetensors", "model.layers.32.self_attn.q_proj.weight": "model-00039-of-00081.safetensors", "model.layers.32.self_attn.v_proj.weight": "model-00039-of-00081.safetensors", "model.layers.33.input_layernorm.weight": "model-00039-of-00081.safetensors", "model.layers.33.mlp.down_proj.weight": "model-00039-of-00081.safetensors", "model.layers.33.mlp.gate_proj.weight": "model-00039-of-00081.safetensors", "model.layers.33.mlp.up_proj.weight": "model-00040-of-00081.safetensors", "model.layers.33.post_attention_layernorm.weight": "model-00040-of-00081.safetensors", "model.layers.33.self_attn.k_proj.weight": "model-00040-of-00081.safetensors", "model.layers.33.self_attn.o_proj.weight": "model-00040-of-00081.safetensors", "model.layers.33.self_attn.q_proj.weight": "model-00040-of-00081.safetensors", "model.layers.33.self_attn.v_proj.weight": "model-00040-of-00081.safetensors", "model.layers.34.input_layernorm.weight": "model-00040-of-00081.safetensors", "model.layers.34.mlp.down_proj.weight": "model-00040-of-00081.safetensors", "model.layers.34.mlp.gate_proj.weight": "model-00040-of-00081.safetensors", "model.layers.34.mlp.up_proj.weight": "model-00040-of-00081.safetensors", "model.layers.34.post_attention_layernorm.weight": "model-00040-of-00081.safetensors", "model.layers.34.self_attn.k_proj.weight": "model-00040-of-00081.safetensors", "model.layers.34.self_attn.o_proj.weight": "model-00040-of-00081.safetensors", "model.layers.34.self_attn.q_proj.weight": "model-00040-of-00081.safetensors", "model.layers.34.self_attn.v_proj.weight": "model-00040-of-00081.safetensors", "model.layers.35.input_layernorm.weight": "model-00040-of-00081.safetensors", "model.layers.35.mlp.down_proj.weight": "model-00040-of-00081.safetensors", "model.layers.35.mlp.gate_proj.weight": "model-00041-of-00081.safetensors", "model.layers.35.mlp.up_proj.weight": "model-00041-of-00081.safetensors", "model.layers.35.post_attention_layernorm.weight": "model-00041-of-00081.safetensors", "model.layers.35.self_attn.k_proj.weight": "model-00041-of-00081.safetensors", "model.layers.35.self_attn.o_proj.weight": "model-00041-of-00081.safetensors", "model.layers.35.self_attn.q_proj.weight": "model-00041-of-00081.safetensors", "model.layers.35.self_attn.v_proj.weight": "model-00041-of-00081.safetensors", "model.layers.36.input_layernorm.weight": "model-00041-of-00081.safetensors", "model.layers.36.mlp.down_proj.weight": "model-00041-of-00081.safetensors", "model.layers.36.mlp.gate_proj.weight": "model-00041-of-00081.safetensors", "model.layers.36.mlp.up_proj.weight": "model-00041-of-00081.safetensors", "model.layers.36.post_attention_layernorm.weight": "model-00041-of-00081.safetensors", "model.layers.36.self_attn.k_proj.weight": "model-00041-of-00081.safetensors", "model.layers.36.self_attn.o_proj.weight": "model-00041-of-00081.safetensors", "model.layers.36.self_attn.q_proj.weight": "model-00041-of-00081.safetensors", "model.layers.36.self_attn.v_proj.weight": "model-00041-of-00081.safetensors", "model.layers.37.input_layernorm.weight": "model-00041-of-00081.safetensors", "model.layers.37.mlp.down_proj.weight": "model-00042-of-00081.safetensors", "model.layers.37.mlp.gate_proj.weight": "model-00042-of-00081.safetensors", "model.layers.37.mlp.up_proj.weight": "model-00042-of-00081.safetensors", "model.layers.37.post_attention_layernorm.weight": "model-00042-of-00081.safetensors", "model.layers.37.self_attn.k_proj.weight": "model-00042-of-00081.safetensors", "model.layers.37.self_attn.o_proj.weight": "model-00042-of-00081.safetensors", "model.layers.37.self_attn.q_proj.weight": "model-00042-of-00081.safetensors", "model.layers.37.self_attn.v_proj.weight": "model-00042-of-00081.safetensors", "model.layers.38.input_layernorm.weight": "model-00042-of-00081.safetensors", "model.layers.38.mlp.down_proj.weight": "model-00042-of-00081.safetensors", "model.layers.38.mlp.gate_proj.weight": "model-00042-of-00081.safetensors", "model.layers.38.mlp.up_proj.weight": "model-00042-of-00081.safetensors", "model.layers.38.post_attention_layernorm.weight": "model-00042-of-00081.safetensors", "model.layers.38.self_attn.k_proj.weight": "model-00042-of-00081.safetensors", "model.layers.38.self_attn.o_proj.weight": "model-00043-of-00081.safetensors", "model.layers.38.self_attn.q_proj.weight": "model-00043-of-00081.safetensors", "model.layers.38.self_attn.v_proj.weight": "model-00043-of-00081.safetensors", "model.layers.39.input_layernorm.weight": "model-00043-of-00081.safetensors", "model.layers.39.mlp.down_proj.weight": "model-00043-of-00081.safetensors", "model.layers.39.mlp.gate_proj.weight": "model-00043-of-00081.safetensors", "model.layers.39.mlp.up_proj.weight": "model-00043-of-00081.safetensors", "model.layers.39.post_attention_layernorm.weight": "model-00043-of-00081.safetensors", "model.layers.39.self_attn.k_proj.weight": "model-00043-of-00081.safetensors", "model.layers.39.self_attn.o_proj.weight": "model-00043-of-00081.safetensors", "model.layers.39.self_attn.q_proj.weight": "model-00043-of-00081.safetensors", "model.layers.39.self_attn.v_proj.weight": "model-00043-of-00081.safetensors", "model.layers.4.input_layernorm.weight": "model-00043-of-00081.safetensors", "model.layers.4.mlp.down_proj.weight": "model-00043-of-00081.safetensors", "model.layers.4.mlp.gate_proj.weight": "model-00043-of-00081.safetensors", "model.layers.4.mlp.up_proj.weight": "model-00044-of-00081.safetensors", "model.layers.4.post_attention_layernorm.weight": "model-00044-of-00081.safetensors", "model.layers.4.self_attn.k_proj.weight": "model-00044-of-00081.safetensors", "model.layers.4.self_attn.o_proj.weight": "model-00044-of-00081.safetensors", "model.layers.4.self_attn.q_proj.weight": "model-00044-of-00081.safetensors", "model.layers.4.self_attn.v_proj.weight": "model-00044-of-00081.safetensors", "model.layers.40.input_layernorm.weight": "model-00044-of-00081.safetensors", "model.layers.40.mlp.down_proj.weight": "model-00044-of-00081.safetensors", "model.layers.40.mlp.gate_proj.weight": "model-00044-of-00081.safetensors", "model.layers.40.mlp.up_proj.weight": "model-00044-of-00081.safetensors", "model.layers.40.post_attention_layernorm.weight": "model-00044-of-00081.safetensors", "model.layers.40.self_attn.k_proj.weight": "model-00044-of-00081.safetensors", "model.layers.40.self_attn.o_proj.weight": "model-00044-of-00081.safetensors", "model.layers.40.self_attn.q_proj.weight": "model-00044-of-00081.safetensors", "model.layers.40.self_attn.v_proj.weight": "model-00044-of-00081.safetensors", "model.layers.41.input_layernorm.weight": "model-00044-of-00081.safetensors", "model.layers.41.mlp.down_proj.weight": "model-00044-of-00081.safetensors", "model.layers.41.mlp.gate_proj.weight": "model-00045-of-00081.safetensors", "model.layers.41.mlp.up_proj.weight": "model-00045-of-00081.safetensors", "model.layers.41.post_attention_layernorm.weight": "model-00045-of-00081.safetensors", "model.layers.41.self_attn.k_proj.weight": "model-00045-of-00081.safetensors", "model.layers.41.self_attn.o_proj.weight": "model-00045-of-00081.safetensors", "model.layers.41.self_attn.q_proj.weight": "model-00045-of-00081.safetensors", "model.layers.41.self_attn.v_proj.weight": "model-00045-of-00081.safetensors", "model.layers.42.input_layernorm.weight": "model-00045-of-00081.safetensors", "model.layers.42.mlp.down_proj.weight": "model-00045-of-00081.safetensors", "model.layers.42.mlp.gate_proj.weight": "model-00045-of-00081.safetensors", "model.layers.42.mlp.up_proj.weight": "model-00045-of-00081.safetensors", "model.layers.42.post_attention_layernorm.weight": "model-00045-of-00081.safetensors", "model.layers.42.self_attn.k_proj.weight": "model-00045-of-00081.safetensors", "model.layers.42.self_attn.o_proj.weight": "model-00045-of-00081.safetensors", "model.layers.42.self_attn.q_proj.weight": "model-00045-of-00081.safetensors", "model.layers.42.self_attn.v_proj.weight": "model-00045-of-00081.safetensors", "model.layers.43.input_layernorm.weight": "model-00045-of-00081.safetensors", "model.layers.43.mlp.down_proj.weight": "model-00046-of-00081.safetensors", "model.layers.43.mlp.gate_proj.weight": "model-00046-of-00081.safetensors", "model.layers.43.mlp.up_proj.weight": "model-00046-of-00081.safetensors", "model.layers.43.post_attention_layernorm.weight": "model-00046-of-00081.safetensors", "model.layers.43.self_attn.k_proj.weight": "model-00046-of-00081.safetensors", "model.layers.43.self_attn.o_proj.weight": "model-00046-of-00081.safetensors", "model.layers.43.self_attn.q_proj.weight": "model-00046-of-00081.safetensors", "model.layers.43.self_attn.v_proj.weight": "model-00046-of-00081.safetensors", "model.layers.44.input_layernorm.weight": "model-00046-of-00081.safetensors", "model.layers.44.mlp.down_proj.weight": "model-00046-of-00081.safetensors", "model.layers.44.mlp.gate_proj.weight": "model-00046-of-00081.safetensors", "model.layers.44.mlp.up_proj.weight": "model-00046-of-00081.safetensors", "model.layers.44.post_attention_layernorm.weight": "model-00046-of-00081.safetensors", "model.layers.44.self_attn.k_proj.weight": "model-00046-of-00081.safetensors", "model.layers.44.self_attn.o_proj.weight": "model-00047-of-00081.safetensors", "model.layers.44.self_attn.q_proj.weight": "model-00047-of-00081.safetensors", "model.layers.44.self_attn.v_proj.weight": "model-00047-of-00081.safetensors", "model.layers.45.input_layernorm.weight": "model-00047-of-00081.safetensors", "model.layers.45.mlp.down_proj.weight": "model-00047-of-00081.safetensors", "model.layers.45.mlp.gate_proj.weight": "model-00047-of-00081.safetensors", "model.layers.45.mlp.up_proj.weight": "model-00047-of-00081.safetensors", "model.layers.45.post_attention_layernorm.weight": "model-00047-of-00081.safetensors", "model.layers.45.self_attn.k_proj.weight": "model-00047-of-00081.safetensors", "model.layers.45.self_attn.o_proj.weight": "model-00047-of-00081.safetensors", "model.layers.45.self_attn.q_proj.weight": "model-00047-of-00081.safetensors", "model.layers.45.self_attn.v_proj.weight": "model-00047-of-00081.safetensors", "model.layers.46.input_layernorm.weight": "model-00047-of-00081.safetensors", "model.layers.46.mlp.down_proj.weight": "model-00047-of-00081.safetensors", "model.layers.46.mlp.gate_proj.weight": "model-00047-of-00081.safetensors", "model.layers.46.mlp.up_proj.weight": "model-00048-of-00081.safetensors", "model.layers.46.post_attention_layernorm.weight": "model-00048-of-00081.safetensors", "model.layers.46.self_attn.k_proj.weight": "model-00048-of-00081.safetensors", "model.layers.46.self_attn.o_proj.weight": "model-00048-of-00081.safetensors", "model.layers.46.self_attn.q_proj.weight": "model-00048-of-00081.safetensors", "model.layers.46.self_attn.v_proj.weight": "model-00048-of-00081.safetensors", "model.layers.47.input_layernorm.weight": "model-00048-of-00081.safetensors", "model.layers.47.mlp.down_proj.weight": "model-00048-of-00081.safetensors", "model.layers.47.mlp.gate_proj.weight": "model-00048-of-00081.safetensors", "model.layers.47.mlp.up_proj.weight": "model-00048-of-00081.safetensors", "model.layers.47.post_attention_layernorm.weight": "model-00048-of-00081.safetensors", "model.layers.47.self_attn.k_proj.weight": "model-00048-of-00081.safetensors", "model.layers.47.self_attn.o_proj.weight": "model-00048-of-00081.safetensors", "model.layers.47.self_attn.q_proj.weight": "model-00048-of-00081.safetensors", "model.layers.47.self_attn.v_proj.weight": "model-00048-of-00081.safetensors", "model.layers.48.input_layernorm.weight": "model-00048-of-00081.safetensors", "model.layers.48.mlp.down_proj.weight": "model-00048-of-00081.safetensors", "model.layers.48.mlp.gate_proj.weight": "model-00049-of-00081.safetensors", "model.layers.48.mlp.up_proj.weight": "model-00049-of-00081.safetensors", "model.layers.48.post_attention_layernorm.weight": "model-00049-of-00081.safetensors", "model.layers.48.self_attn.k_proj.weight": "model-00049-of-00081.safetensors", "model.layers.48.self_attn.o_proj.weight": "model-00049-of-00081.safetensors", "model.layers.48.self_attn.q_proj.weight": "model-00049-of-00081.safetensors", "model.layers.48.self_attn.v_proj.weight": "model-00049-of-00081.safetensors", "model.layers.49.input_layernorm.weight": "model-00049-of-00081.safetensors", "model.layers.49.mlp.down_proj.weight": "model-00049-of-00081.safetensors", "model.layers.49.mlp.gate_proj.weight": "model-00049-of-00081.safetensors", "model.layers.49.mlp.up_proj.weight": "model-00049-of-00081.safetensors", "model.layers.49.post_attention_layernorm.weight": "model-00049-of-00081.safetensors", "model.layers.49.self_attn.k_proj.weight": "model-00049-of-00081.safetensors", "model.layers.49.self_attn.o_proj.weight": "model-00049-of-00081.safetensors", "model.layers.49.self_attn.q_proj.weight": "model-00049-of-00081.safetensors", "model.layers.49.self_attn.v_proj.weight": "model-00049-of-00081.safetensors", "model.layers.5.input_layernorm.weight": "model-00049-of-00081.safetensors", "model.layers.5.mlp.down_proj.weight": "model-00050-of-00081.safetensors", "model.layers.5.mlp.gate_proj.weight": "model-00050-of-00081.safetensors", "model.layers.5.mlp.up_proj.weight": "model-00050-of-00081.safetensors", "model.layers.5.post_attention_layernorm.weight": "model-00050-of-00081.safetensors", "model.layers.5.self_attn.k_proj.weight": "model-00050-of-00081.safetensors", "model.layers.5.self_attn.o_proj.weight": "model-00050-of-00081.safetensors", "model.layers.5.self_attn.q_proj.weight": "model-00050-of-00081.safetensors", "model.layers.5.self_attn.v_proj.weight": "model-00050-of-00081.safetensors", "model.layers.50.input_layernorm.weight": "model-00050-of-00081.safetensors", "model.layers.50.mlp.down_proj.weight": "model-00050-of-00081.safetensors", "model.layers.50.mlp.gate_proj.weight": "model-00050-of-00081.safetensors", "model.layers.50.mlp.up_proj.weight": "model-00050-of-00081.safetensors", "model.layers.50.post_attention_layernorm.weight": "model-00050-of-00081.safetensors", "model.layers.50.self_attn.k_proj.weight": "model-00050-of-00081.safetensors", "model.layers.50.self_attn.o_proj.weight": "model-00051-of-00081.safetensors", "model.layers.50.self_attn.q_proj.weight": "model-00051-of-00081.safetensors", "model.layers.50.self_attn.v_proj.weight": "model-00051-of-00081.safetensors", "model.layers.51.input_layernorm.weight": "model-00051-of-00081.safetensors", "model.layers.51.mlp.down_proj.weight": "model-00051-of-00081.safetensors", "model.layers.51.mlp.gate_proj.weight": "model-00051-of-00081.safetensors", "model.layers.51.mlp.up_proj.weight": "model-00051-of-00081.safetensors", "model.layers.51.post_attention_layernorm.weight": "model-00051-of-00081.safetensors", "model.layers.51.self_attn.k_proj.weight": "model-00051-of-00081.safetensors", "model.layers.51.self_attn.o_proj.weight": "model-00051-of-00081.safetensors", "model.layers.51.self_attn.q_proj.weight": "model-00051-of-00081.safetensors", "model.layers.51.self_attn.v_proj.weight": "model-00051-of-00081.safetensors", "model.layers.52.input_layernorm.weight": "model-00051-of-00081.safetensors", "model.layers.52.mlp.down_proj.weight": "model-00051-of-00081.safetensors", "model.layers.52.mlp.gate_proj.weight": "model-00051-of-00081.safetensors", "model.layers.52.mlp.up_proj.weight": "model-00052-of-00081.safetensors", "model.layers.52.post_attention_layernorm.weight": "model-00052-of-00081.safetensors", "model.layers.52.self_attn.k_proj.weight": "model-00052-of-00081.safetensors", "model.layers.52.self_attn.o_proj.weight": "model-00052-of-00081.safetensors", "model.layers.52.self_attn.q_proj.weight": "model-00052-of-00081.safetensors", "model.layers.52.self_attn.v_proj.weight": "model-00052-of-00081.safetensors", "model.layers.53.input_layernorm.weight": "model-00052-of-00081.safetensors", "model.layers.53.mlp.down_proj.weight": "model-00052-of-00081.safetensors", "model.layers.53.mlp.gate_proj.weight": "model-00052-of-00081.safetensors", "model.layers.53.mlp.up_proj.weight": "model-00052-of-00081.safetensors", "model.layers.53.post_attention_layernorm.weight": "model-00052-of-00081.safetensors", "model.layers.53.self_attn.k_proj.weight": "model-00052-of-00081.safetensors", "model.layers.53.self_attn.o_proj.weight": "model-00052-of-00081.safetensors", "model.layers.53.self_attn.q_proj.weight": "model-00052-of-00081.safetensors", "model.layers.53.self_attn.v_proj.weight": "model-00052-of-00081.safetensors", "model.layers.54.input_layernorm.weight": "model-00052-of-00081.safetensors", "model.layers.54.mlp.down_proj.weight": "model-00052-of-00081.safetensors", "model.layers.54.mlp.gate_proj.weight": "model-00053-of-00081.safetensors", "model.layers.54.mlp.up_proj.weight": "model-00053-of-00081.safetensors", "model.layers.54.post_attention_layernorm.weight": "model-00053-of-00081.safetensors", "model.layers.54.self_attn.k_proj.weight": "model-00053-of-00081.safetensors", "model.layers.54.self_attn.o_proj.weight": "model-00053-of-00081.safetensors", "model.layers.54.self_attn.q_proj.weight": "model-00053-of-00081.safetensors", "model.layers.54.self_attn.v_proj.weight": "model-00053-of-00081.safetensors", "model.layers.55.input_layernorm.weight": "model-00053-of-00081.safetensors", "model.layers.55.mlp.down_proj.weight": "model-00053-of-00081.safetensors", "model.layers.55.mlp.gate_proj.weight": "model-00053-of-00081.safetensors", "model.layers.55.mlp.up_proj.weight": "model-00053-of-00081.safetensors", "model.layers.55.post_attention_layernorm.weight": "model-00053-of-00081.safetensors", "model.layers.55.self_attn.k_proj.weight": "model-00053-of-00081.safetensors", "model.layers.55.self_attn.o_proj.weight": "model-00053-of-00081.safetensors", "model.layers.55.self_attn.q_proj.weight": "model-00053-of-00081.safetensors", "model.layers.55.self_attn.v_proj.weight": "model-00053-of-00081.safetensors", "model.layers.56.input_layernorm.weight": "model-00053-of-00081.safetensors", "model.layers.56.mlp.down_proj.weight": "model-00054-of-00081.safetensors", "model.layers.56.mlp.gate_proj.weight": "model-00054-of-00081.safetensors", "model.layers.56.mlp.up_proj.weight": "model-00054-of-00081.safetensors", "model.layers.56.post_attention_layernorm.weight": "model-00054-of-00081.safetensors", "model.layers.56.self_attn.k_proj.weight": "model-00054-of-00081.safetensors", "model.layers.56.self_attn.o_proj.weight": "model-00054-of-00081.safetensors", "model.layers.56.self_attn.q_proj.weight": "model-00054-of-00081.safetensors", "model.layers.56.self_attn.v_proj.weight": "model-00054-of-00081.safetensors", "model.layers.57.input_layernorm.weight": "model-00054-of-00081.safetensors", "model.layers.57.mlp.down_proj.weight": "model-00054-of-00081.safetensors", "model.layers.57.mlp.gate_proj.weight": "model-00054-of-00081.safetensors", "model.layers.57.mlp.up_proj.weight": "model-00054-of-00081.safetensors", "model.layers.57.post_attention_layernorm.weight": "model-00054-of-00081.safetensors", "model.layers.57.self_attn.k_proj.weight": "model-00054-of-00081.safetensors", "model.layers.57.self_attn.o_proj.weight": "model-00055-of-00081.safetensors", "model.layers.57.self_attn.q_proj.weight": "model-00055-of-00081.safetensors", "model.layers.57.self_attn.v_proj.weight": "model-00055-of-00081.safetensors", "model.layers.58.input_layernorm.weight": "model-00055-of-00081.safetensors", "model.layers.58.mlp.down_proj.weight": "model-00055-of-00081.safetensors", "model.layers.58.mlp.gate_proj.weight": "model-00055-of-00081.safetensors", "model.layers.58.mlp.up_proj.weight": "model-00055-of-00081.safetensors", "model.layers.58.post_attention_layernorm.weight": "model-00055-of-00081.safetensors", "model.layers.58.self_attn.k_proj.weight": "model-00055-of-00081.safetensors", "model.layers.58.self_attn.o_proj.weight": "model-00055-of-00081.safetensors", "model.layers.58.self_attn.q_proj.weight": "model-00055-of-00081.safetensors", "model.layers.58.self_attn.v_proj.weight": "model-00055-of-00081.safetensors", "model.layers.59.input_layernorm.weight": "model-00055-of-00081.safetensors", "model.layers.59.mlp.down_proj.weight": "model-00055-of-00081.safetensors", "model.layers.59.mlp.gate_proj.weight": "model-00055-of-00081.safetensors", "model.layers.59.mlp.up_proj.weight": "model-00056-of-00081.safetensors", "model.layers.59.post_attention_layernorm.weight": "model-00056-of-00081.safetensors", "model.layers.59.self_attn.k_proj.weight": "model-00056-of-00081.safetensors", "model.layers.59.self_attn.o_proj.weight": "model-00056-of-00081.safetensors", "model.layers.59.self_attn.q_proj.weight": "model-00056-of-00081.safetensors", "model.layers.59.self_attn.v_proj.weight": "model-00056-of-00081.safetensors", "model.layers.6.input_layernorm.weight": "model-00056-of-00081.safetensors", "model.layers.6.mlp.down_proj.weight": "model-00056-of-00081.safetensors", "model.layers.6.mlp.gate_proj.weight": "model-00056-of-00081.safetensors", "model.layers.6.mlp.up_proj.weight": "model-00056-of-00081.safetensors", "model.layers.6.post_attention_layernorm.weight": "model-00056-of-00081.safetensors", "model.layers.6.self_attn.k_proj.weight": "model-00056-of-00081.safetensors", "model.layers.6.self_attn.o_proj.weight": "model-00056-of-00081.safetensors", "model.layers.6.self_attn.q_proj.weight": "model-00056-of-00081.safetensors", "model.layers.6.self_attn.v_proj.weight": "model-00056-of-00081.safetensors", "model.layers.60.input_layernorm.weight": "model-00056-of-00081.safetensors", "model.layers.60.mlp.down_proj.weight": "model-00056-of-00081.safetensors", "model.layers.60.mlp.gate_proj.weight": "model-00057-of-00081.safetensors", "model.layers.60.mlp.up_proj.weight": "model-00057-of-00081.safetensors", "model.layers.60.post_attention_layernorm.weight": "model-00057-of-00081.safetensors", "model.layers.60.self_attn.k_proj.weight": "model-00057-of-00081.safetensors", "model.layers.60.self_attn.o_proj.weight": "model-00057-of-00081.safetensors", "model.layers.60.self_attn.q_proj.weight": "model-00057-of-00081.safetensors", "model.layers.60.self_attn.v_proj.weight": "model-00057-of-00081.safetensors", "model.layers.61.input_layernorm.weight": "model-00057-of-00081.safetensors", "model.layers.61.mlp.down_proj.weight": "model-00057-of-00081.safetensors", "model.layers.61.mlp.gate_proj.weight": "model-00057-of-00081.safetensors", "model.layers.61.mlp.up_proj.weight": "model-00057-of-00081.safetensors", "model.layers.61.post_attention_layernorm.weight": "model-00057-of-00081.safetensors", "model.layers.61.self_attn.k_proj.weight": "model-00057-of-00081.safetensors", "model.layers.61.self_attn.o_proj.weight": "model-00057-of-00081.safetensors", "model.layers.61.self_attn.q_proj.weight": "model-00057-of-00081.safetensors", "model.layers.61.self_attn.v_proj.weight": "model-00057-of-00081.safetensors", "model.layers.62.input_layernorm.weight": "model-00057-of-00081.safetensors", "model.layers.62.mlp.down_proj.weight": "model-00058-of-00081.safetensors", "model.layers.62.mlp.gate_proj.weight": "model-00058-of-00081.safetensors", "model.layers.62.mlp.up_proj.weight": "model-00058-of-00081.safetensors", "model.layers.62.post_attention_layernorm.weight": "model-00058-of-00081.safetensors", "model.layers.62.self_attn.k_proj.weight": "model-00058-of-00081.safetensors", "model.layers.62.self_attn.o_proj.weight": "model-00058-of-00081.safetensors", "model.layers.62.self_attn.q_proj.weight": "model-00058-of-00081.safetensors", "model.layers.62.self_attn.v_proj.weight": "model-00058-of-00081.safetensors", "model.layers.63.input_layernorm.weight": "model-00058-of-00081.safetensors", "model.layers.63.mlp.down_proj.weight": "model-00058-of-00081.safetensors", "model.layers.63.mlp.gate_proj.weight": "model-00058-of-00081.safetensors", "model.layers.63.mlp.up_proj.weight": "model-00058-of-00081.safetensors", "model.layers.63.post_attention_layernorm.weight": "model-00058-of-00081.safetensors", "model.layers.63.self_attn.k_proj.weight": "model-00058-of-00081.safetensors", "model.layers.63.self_attn.o_proj.weight": "model-00059-of-00081.safetensors", "model.layers.63.self_attn.q_proj.weight": "model-00059-of-00081.safetensors", "model.layers.63.self_attn.v_proj.weight": "model-00059-of-00081.safetensors", "model.layers.64.input_layernorm.weight": "model-00059-of-00081.safetensors", "model.layers.64.mlp.down_proj.weight": "model-00059-of-00081.safetensors", "model.layers.64.mlp.gate_proj.weight": "model-00059-of-00081.safetensors", "model.layers.64.mlp.up_proj.weight": "model-00059-of-00081.safetensors", "model.layers.64.post_attention_layernorm.weight": "model-00059-of-00081.safetensors", "model.layers.64.self_attn.k_proj.weight": "model-00059-of-00081.safetensors", "model.layers.64.self_attn.o_proj.weight": "model-00059-of-00081.safetensors", "model.layers.64.self_attn.q_proj.weight": "model-00059-of-00081.safetensors", "model.layers.64.self_attn.v_proj.weight": "model-00059-of-00081.safetensors", "model.layers.65.input_layernorm.weight": "model-00059-of-00081.safetensors", "model.layers.65.mlp.down_proj.weight": "model-00059-of-00081.safetensors", "model.layers.65.mlp.gate_proj.weight": "model-00059-of-00081.safetensors", "model.layers.65.mlp.up_proj.weight": "model-00060-of-00081.safetensors", "model.layers.65.post_attention_layernorm.weight": "model-00060-of-00081.safetensors", "model.layers.65.self_attn.k_proj.weight": "model-00060-of-00081.safetensors", "model.layers.65.self_attn.o_proj.weight": "model-00060-of-00081.safetensors", "model.layers.65.self_attn.q_proj.weight": "model-00060-of-00081.safetensors", "model.layers.65.self_attn.v_proj.weight": "model-00060-of-00081.safetensors", "model.layers.66.input_layernorm.weight": "model-00060-of-00081.safetensors", "model.layers.66.mlp.down_proj.weight": "model-00060-of-00081.safetensors", "model.layers.66.mlp.gate_proj.weight": "model-00060-of-00081.safetensors", "model.layers.66.mlp.up_proj.weight": "model-00060-of-00081.safetensors", "model.layers.66.post_attention_layernorm.weight": "model-00060-of-00081.safetensors", "model.layers.66.self_attn.k_proj.weight": "model-00060-of-00081.safetensors", "model.layers.66.self_attn.o_proj.weight": "model-00060-of-00081.safetensors", "model.layers.66.self_attn.q_proj.weight": "model-00060-of-00081.safetensors", "model.layers.66.self_attn.v_proj.weight": "model-00060-of-00081.safetensors", "model.layers.67.input_layernorm.weight": "model-00060-of-00081.safetensors", "model.layers.67.mlp.down_proj.weight": "model-00060-of-00081.safetensors", "model.layers.67.mlp.gate_proj.weight": "model-00061-of-00081.safetensors", "model.layers.67.mlp.up_proj.weight": "model-00061-of-00081.safetensors", "model.layers.67.post_attention_layernorm.weight": "model-00061-of-00081.safetensors", "model.layers.67.self_attn.k_proj.weight": "model-00061-of-00081.safetensors", "model.layers.67.self_attn.o_proj.weight": "model-00061-of-00081.safetensors", "model.layers.67.self_attn.q_proj.weight": "model-00061-of-00081.safetensors", "model.layers.67.self_attn.v_proj.weight": "model-00061-of-00081.safetensors", "model.layers.68.input_layernorm.weight": "model-00061-of-00081.safetensors", "model.layers.68.mlp.down_proj.weight": "model-00061-of-00081.safetensors", "model.layers.68.mlp.gate_proj.weight": "model-00061-of-00081.safetensors", "model.layers.68.mlp.up_proj.weight": "model-00061-of-00081.safetensors", "model.layers.68.post_attention_layernorm.weight": "model-00061-of-00081.safetensors", "model.layers.68.self_attn.k_proj.weight": "model-00061-of-00081.safetensors", "model.layers.68.self_attn.o_proj.weight": "model-00061-of-00081.safetensors", "model.layers.68.self_attn.q_proj.weight": "model-00061-of-00081.safetensors", "model.layers.68.self_attn.v_proj.weight": "model-00061-of-00081.safetensors", "model.layers.69.input_layernorm.weight": "model-00061-of-00081.safetensors", "model.layers.69.mlp.down_proj.weight": "model-00062-of-00081.safetensors", "model.layers.69.mlp.gate_proj.weight": "model-00062-of-00081.safetensors", "model.layers.69.mlp.up_proj.weight": "model-00062-of-00081.safetensors", "model.layers.69.post_attention_layernorm.weight": "model-00062-of-00081.safetensors", "model.layers.69.self_attn.k_proj.weight": "model-00062-of-00081.safetensors", "model.layers.69.self_attn.o_proj.weight": "model-00062-of-00081.safetensors", "model.layers.69.self_attn.q_proj.weight": "model-00062-of-00081.safetensors", "model.layers.69.self_attn.v_proj.weight": "model-00062-of-00081.safetensors", "model.layers.7.input_layernorm.weight": "model-00062-of-00081.safetensors", "model.layers.7.mlp.down_proj.weight": "model-00062-of-00081.safetensors", "model.layers.7.mlp.gate_proj.weight": "model-00062-of-00081.safetensors", "model.layers.7.mlp.up_proj.weight": "model-00062-of-00081.safetensors", "model.layers.7.post_attention_layernorm.weight": "model-00062-of-00081.safetensors", "model.layers.7.self_attn.k_proj.weight": "model-00062-of-00081.safetensors", "model.layers.7.self_attn.o_proj.weight": "model-00063-of-00081.safetensors", "model.layers.7.self_attn.q_proj.weight": "model-00063-of-00081.safetensors", "model.layers.7.self_attn.v_proj.weight": "model-00063-of-00081.safetensors", "model.layers.70.input_layernorm.weight": "model-00063-of-00081.safetensors", "model.layers.70.mlp.down_proj.weight": "model-00063-of-00081.safetensors", "model.layers.70.mlp.gate_proj.weight": "model-00063-of-00081.safetensors", "model.layers.70.mlp.up_proj.weight": "model-00063-of-00081.safetensors", "model.layers.70.post_attention_layernorm.weight": "model-00063-of-00081.safetensors", "model.layers.70.self_attn.k_proj.weight": "model-00063-of-00081.safetensors", "model.layers.70.self_attn.o_proj.weight": "model-00063-of-00081.safetensors", "model.layers.70.self_attn.q_proj.weight": "model-00063-of-00081.safetensors", "model.layers.70.self_attn.v_proj.weight": "model-00063-of-00081.safetensors", "model.layers.71.input_layernorm.weight": "model-00063-of-00081.safetensors", "model.layers.71.mlp.down_proj.weight": "model-00063-of-00081.safetensors", "model.layers.71.mlp.gate_proj.weight": "model-00063-of-00081.safetensors", "model.layers.71.mlp.up_proj.weight": "model-00064-of-00081.safetensors", "model.layers.71.post_attention_layernorm.weight": "model-00064-of-00081.safetensors", "model.layers.71.self_attn.k_proj.weight": "model-00064-of-00081.safetensors", "model.layers.71.self_attn.o_proj.weight": "model-00064-of-00081.safetensors", "model.layers.71.self_attn.q_proj.weight": "model-00064-of-00081.safetensors", "model.layers.71.self_attn.v_proj.weight": "model-00064-of-00081.safetensors", "model.layers.72.input_layernorm.weight": "model-00064-of-00081.safetensors", "model.layers.72.mlp.down_proj.weight": "model-00064-of-00081.safetensors", "model.layers.72.mlp.gate_proj.weight": "model-00064-of-00081.safetensors", "model.layers.72.mlp.up_proj.weight": "model-00064-of-00081.safetensors", "model.layers.72.post_attention_layernorm.weight": "model-00064-of-00081.safetensors", "model.layers.72.self_attn.k_proj.weight": "model-00064-of-00081.safetensors", "model.layers.72.self_attn.o_proj.weight": "model-00064-of-00081.safetensors", "model.layers.72.self_attn.q_proj.weight": "model-00064-of-00081.safetensors", "model.layers.72.self_attn.v_proj.weight": "model-00064-of-00081.safetensors", "model.layers.73.input_layernorm.weight": "model-00064-of-00081.safetensors", "model.layers.73.mlp.down_proj.weight": "model-00064-of-00081.safetensors", "model.layers.73.mlp.gate_proj.weight": "model-00065-of-00081.safetensors", "model.layers.73.mlp.up_proj.weight": "model-00065-of-00081.safetensors", "model.layers.73.post_attention_layernorm.weight": "model-00065-of-00081.safetensors", "model.layers.73.self_attn.k_proj.weight": "model-00065-of-00081.safetensors", "model.layers.73.self_attn.o_proj.weight": "model-00065-of-00081.safetensors", "model.layers.73.self_attn.q_proj.weight": "model-00065-of-00081.safetensors", "model.layers.73.self_attn.v_proj.weight": "model-00065-of-00081.safetensors", "model.layers.74.input_layernorm.weight": "model-00065-of-00081.safetensors", "model.layers.74.mlp.down_proj.weight": "model-00065-of-00081.safetensors", "model.layers.74.mlp.gate_proj.weight": "model-00065-of-00081.safetensors", "model.layers.74.mlp.up_proj.weight": "model-00065-of-00081.safetensors", "model.layers.74.post_attention_layernorm.weight": "model-00065-of-00081.safetensors", "model.layers.74.self_attn.k_proj.weight": "model-00065-of-00081.safetensors", "model.layers.74.self_attn.o_proj.weight": "model-00065-of-00081.safetensors", "model.layers.74.self_attn.q_proj.weight": "model-00065-of-00081.safetensors", "model.layers.74.self_attn.v_proj.weight": "model-00065-of-00081.safetensors", "model.layers.75.input_layernorm.weight": "model-00065-of-00081.safetensors", "model.layers.75.mlp.down_proj.weight": "model-00066-of-00081.safetensors", "model.layers.75.mlp.gate_proj.weight": "model-00066-of-00081.safetensors", "model.layers.75.mlp.up_proj.weight": "model-00066-of-00081.safetensors", "model.layers.75.post_attention_layernorm.weight": "model-00066-of-00081.safetensors", "model.layers.75.self_attn.k_proj.weight": "model-00066-of-00081.safetensors", "model.layers.75.self_attn.o_proj.weight": "model-00066-of-00081.safetensors", "model.layers.75.self_attn.q_proj.weight": "model-00066-of-00081.safetensors", "model.layers.75.self_attn.v_proj.weight": "model-00066-of-00081.safetensors", "model.layers.76.input_layernorm.weight": "model-00066-of-00081.safetensors", "model.layers.76.mlp.down_proj.weight": "model-00066-of-00081.safetensors", "model.layers.76.mlp.gate_proj.weight": "model-00066-of-00081.safetensors", "model.layers.76.mlp.up_proj.weight": "model-00066-of-00081.safetensors", "model.layers.76.post_attention_layernorm.weight": "model-00066-of-00081.safetensors", "model.layers.76.self_attn.k_proj.weight": "model-00066-of-00081.safetensors", "model.layers.76.self_attn.o_proj.weight": "model-00067-of-00081.safetensors", "model.layers.76.self_attn.q_proj.weight": "model-00067-of-00081.safetensors", "model.layers.76.self_attn.v_proj.weight": "model-00067-of-00081.safetensors", "model.layers.77.input_layernorm.weight": "model-00067-of-00081.safetensors", "model.layers.77.mlp.down_proj.weight": "model-00067-of-00081.safetensors", "model.layers.77.mlp.gate_proj.weight": "model-00067-of-00081.safetensors", "model.layers.77.mlp.up_proj.weight": "model-00067-of-00081.safetensors", "model.layers.77.post_attention_layernorm.weight": "model-00067-of-00081.safetensors", "model.layers.77.self_attn.k_proj.weight": "model-00067-of-00081.safetensors", "model.layers.77.self_attn.o_proj.weight": "model-00067-of-00081.safetensors", "model.layers.77.self_attn.q_proj.weight": "model-00067-of-00081.safetensors", "model.layers.77.self_attn.v_proj.weight": "model-00067-of-00081.safetensors", "model.layers.78.input_layernorm.weight": "model-00067-of-00081.safetensors", "model.layers.78.mlp.down_proj.weight": "model-00067-of-00081.safetensors", "model.layers.78.mlp.gate_proj.weight": "model-00067-of-00081.safetensors", "model.layers.78.mlp.up_proj.weight": "model-00068-of-00081.safetensors", "model.layers.78.post_attention_layernorm.weight": "model-00068-of-00081.safetensors", "model.layers.78.self_attn.k_proj.weight": "model-00068-of-00081.safetensors", "model.layers.78.self_attn.o_proj.weight": "model-00068-of-00081.safetensors", "model.layers.78.self_attn.q_proj.weight": "model-00068-of-00081.safetensors", "model.layers.78.self_attn.v_proj.weight": "model-00068-of-00081.safetensors", "model.layers.79.input_layernorm.weight": "model-00068-of-00081.safetensors", "model.layers.79.mlp.down_proj.weight": "model-00068-of-00081.safetensors", "model.layers.79.mlp.gate_proj.weight": "model-00068-of-00081.safetensors", "model.layers.79.mlp.up_proj.weight": "model-00068-of-00081.safetensors", "model.layers.79.post_attention_layernorm.weight": "model-00068-of-00081.safetensors", "model.layers.79.self_attn.k_proj.weight": "model-00068-of-00081.safetensors", "model.layers.79.self_attn.o_proj.weight": "model-00068-of-00081.safetensors", "model.layers.79.self_attn.q_proj.weight": "model-00068-of-00081.safetensors", "model.layers.79.self_attn.v_proj.weight": "model-00068-of-00081.safetensors", "model.layers.8.input_layernorm.weight": "model-00068-of-00081.safetensors", "model.layers.8.mlp.down_proj.weight": "model-00068-of-00081.safetensors", "model.layers.8.mlp.gate_proj.weight": "model-00069-of-00081.safetensors", "model.layers.8.mlp.up_proj.weight": "model-00069-of-00081.safetensors", "model.layers.8.post_attention_layernorm.weight": "model-00069-of-00081.safetensors", "model.layers.8.self_attn.k_proj.weight": "model-00069-of-00081.safetensors", "model.layers.8.self_attn.o_proj.weight": "model-00069-of-00081.safetensors", "model.layers.8.self_attn.q_proj.weight": "model-00069-of-00081.safetensors", "model.layers.8.self_attn.v_proj.weight": "model-00069-of-00081.safetensors", "model.layers.80.input_layernorm.weight": "model-00069-of-00081.safetensors", "model.layers.80.mlp.down_proj.weight": "model-00069-of-00081.safetensors", "model.layers.80.mlp.gate_proj.weight": "model-00069-of-00081.safetensors", "model.layers.80.mlp.up_proj.weight": "model-00069-of-00081.safetensors", "model.layers.80.post_attention_layernorm.weight": "model-00069-of-00081.safetensors", "model.layers.80.self_attn.k_proj.weight": "model-00069-of-00081.safetensors", "model.layers.80.self_attn.o_proj.weight": "model-00069-of-00081.safetensors", "model.layers.80.self_attn.q_proj.weight": "model-00069-of-00081.safetensors", "model.layers.80.self_attn.v_proj.weight": "model-00069-of-00081.safetensors", "model.layers.81.input_layernorm.weight": "model-00069-of-00081.safetensors", "model.layers.81.mlp.down_proj.weight": "model-00070-of-00081.safetensors", "model.layers.81.mlp.gate_proj.weight": "model-00070-of-00081.safetensors", "model.layers.81.mlp.up_proj.weight": "model-00070-of-00081.safetensors", "model.layers.81.post_attention_layernorm.weight": "model-00070-of-00081.safetensors", "model.layers.81.self_attn.k_proj.weight": "model-00070-of-00081.safetensors", "model.layers.81.self_attn.o_proj.weight": "model-00070-of-00081.safetensors", "model.layers.81.self_attn.q_proj.weight": "model-00070-of-00081.safetensors", "model.layers.81.self_attn.v_proj.weight": "model-00070-of-00081.safetensors", "model.layers.82.input_layernorm.weight": "model-00070-of-00081.safetensors", "model.layers.82.mlp.down_proj.weight": "model-00070-of-00081.safetensors", "model.layers.82.mlp.gate_proj.weight": "model-00070-of-00081.safetensors", "model.layers.82.mlp.up_proj.weight": "model-00070-of-00081.safetensors", "model.layers.82.post_attention_layernorm.weight": "model-00070-of-00081.safetensors", "model.layers.82.self_attn.k_proj.weight": "model-00070-of-00081.safetensors", "model.layers.82.self_attn.o_proj.weight": "model-00071-of-00081.safetensors", "model.layers.82.self_attn.q_proj.weight": "model-00071-of-00081.safetensors", "model.layers.82.self_attn.v_proj.weight": "model-00071-of-00081.safetensors", "model.layers.83.input_layernorm.weight": "model-00071-of-00081.safetensors", "model.layers.83.mlp.down_proj.weight": "model-00071-of-00081.safetensors", "model.layers.83.mlp.gate_proj.weight": "model-00071-of-00081.safetensors", "model.layers.83.mlp.up_proj.weight": "model-00071-of-00081.safetensors", "model.layers.83.post_attention_layernorm.weight": "model-00071-of-00081.safetensors", "model.layers.83.self_attn.k_proj.weight": "model-00071-of-00081.safetensors", "model.layers.83.self_attn.o_proj.weight": "model-00071-of-00081.safetensors", "model.layers.83.self_attn.q_proj.weight": "model-00071-of-00081.safetensors", "model.layers.83.self_attn.v_proj.weight": "model-00071-of-00081.safetensors", "model.layers.84.input_layernorm.weight": "model-00071-of-00081.safetensors", "model.layers.84.mlp.down_proj.weight": "model-00071-of-00081.safetensors", "model.layers.84.mlp.gate_proj.weight": "model-00071-of-00081.safetensors", "model.layers.84.mlp.up_proj.weight": "model-00072-of-00081.safetensors", "model.layers.84.post_attention_layernorm.weight": "model-00072-of-00081.safetensors", "model.layers.84.self_attn.k_proj.weight": "model-00072-of-00081.safetensors", "model.layers.84.self_attn.o_proj.weight": "model-00072-of-00081.safetensors", "model.layers.84.self_attn.q_proj.weight": "model-00072-of-00081.safetensors", "model.layers.84.self_attn.v_proj.weight": "model-00072-of-00081.safetensors", "model.layers.85.input_layernorm.weight": "model-00072-of-00081.safetensors", "model.layers.85.mlp.down_proj.weight": "model-00072-of-00081.safetensors", "model.layers.85.mlp.gate_proj.weight": "model-00072-of-00081.safetensors", "model.layers.85.mlp.up_proj.weight": "model-00072-of-00081.safetensors", "model.layers.85.post_attention_layernorm.weight": "model-00072-of-00081.safetensors", "model.layers.85.self_attn.k_proj.weight": "model-00072-of-00081.safetensors", "model.layers.85.self_attn.o_proj.weight": "model-00072-of-00081.safetensors", "model.layers.85.self_attn.q_proj.weight": "model-00072-of-00081.safetensors", "model.layers.85.self_attn.v_proj.weight": "model-00072-of-00081.safetensors", "model.layers.86.input_layernorm.weight": "model-00072-of-00081.safetensors", "model.layers.86.mlp.down_proj.weight": "model-00072-of-00081.safetensors", "model.layers.86.mlp.gate_proj.weight": "model-00073-of-00081.safetensors", "model.layers.86.mlp.up_proj.weight": "model-00073-of-00081.safetensors", "model.layers.86.post_attention_layernorm.weight": "model-00073-of-00081.safetensors", "model.layers.86.self_attn.k_proj.weight": "model-00073-of-00081.safetensors", "model.layers.86.self_attn.o_proj.weight": "model-00073-of-00081.safetensors", "model.layers.86.self_attn.q_proj.weight": "model-00073-of-00081.safetensors", "model.layers.86.self_attn.v_proj.weight": "model-00073-of-00081.safetensors", "model.layers.87.input_layernorm.weight": "model-00073-of-00081.safetensors", "model.layers.87.mlp.down_proj.weight": "model-00073-of-00081.safetensors", "model.layers.87.mlp.gate_proj.weight": "model-00073-of-00081.safetensors", "model.layers.87.mlp.up_proj.weight": "model-00073-of-00081.safetensors", "model.layers.87.post_attention_layernorm.weight": "model-00073-of-00081.safetensors", "model.layers.87.self_attn.k_proj.weight": "model-00073-of-00081.safetensors", "model.layers.87.self_attn.o_proj.weight": "model-00073-of-00081.safetensors", "model.layers.87.self_attn.q_proj.weight": "model-00073-of-00081.safetensors", "model.layers.87.self_attn.v_proj.weight": "model-00073-of-00081.safetensors", "model.layers.88.input_layernorm.weight": "model-00073-of-00081.safetensors", "model.layers.88.mlp.down_proj.weight": "model-00074-of-00081.safetensors", "model.layers.88.mlp.gate_proj.weight": "model-00074-of-00081.safetensors", "model.layers.88.mlp.up_proj.weight": "model-00074-of-00081.safetensors", "model.layers.88.post_attention_layernorm.weight": "model-00074-of-00081.safetensors", "model.layers.88.self_attn.k_proj.weight": "model-00074-of-00081.safetensors", "model.layers.88.self_attn.o_proj.weight": "model-00074-of-00081.safetensors", "model.layers.88.self_attn.q_proj.weight": "model-00074-of-00081.safetensors", "model.layers.88.self_attn.v_proj.weight": "model-00074-of-00081.safetensors", "model.layers.89.input_layernorm.weight": "model-00074-of-00081.safetensors", "model.layers.89.mlp.down_proj.weight": "model-00074-of-00081.safetensors", "model.layers.89.mlp.gate_proj.weight": "model-00074-of-00081.safetensors", "model.layers.89.mlp.up_proj.weight": "model-00074-of-00081.safetensors", "model.layers.89.post_attention_layernorm.weight": "model-00074-of-00081.safetensors", "model.layers.89.self_attn.k_proj.weight": "model-00074-of-00081.safetensors", "model.layers.89.self_attn.o_proj.weight": "model-00075-of-00081.safetensors", "model.layers.89.self_attn.q_proj.weight": "model-00075-of-00081.safetensors", "model.layers.89.self_attn.v_proj.weight": "model-00075-of-00081.safetensors", "model.layers.9.input_layernorm.weight": "model-00075-of-00081.safetensors", "model.layers.9.mlp.down_proj.weight": "model-00075-of-00081.safetensors", "model.layers.9.mlp.gate_proj.weight": "model-00075-of-00081.safetensors", "model.layers.9.mlp.up_proj.weight": "model-00075-of-00081.safetensors", "model.layers.9.post_attention_layernorm.weight": "model-00075-of-00081.safetensors", "model.layers.9.self_attn.k_proj.weight": "model-00075-of-00081.safetensors", "model.layers.9.self_attn.o_proj.weight": "model-00075-of-00081.safetensors", "model.layers.9.self_attn.q_proj.weight": "model-00075-of-00081.safetensors", "model.layers.9.self_attn.v_proj.weight": "model-00075-of-00081.safetensors", "model.layers.90.input_layernorm.weight": "model-00075-of-00081.safetensors", "model.layers.90.mlp.down_proj.weight": "model-00075-of-00081.safetensors", "model.layers.90.mlp.gate_proj.weight": "model-00075-of-00081.safetensors", "model.layers.90.mlp.up_proj.weight": "model-00076-of-00081.safetensors", "model.layers.90.post_attention_layernorm.weight": "model-00076-of-00081.safetensors", "model.layers.90.self_attn.k_proj.weight": "model-00076-of-00081.safetensors", "model.layers.90.self_attn.o_proj.weight": "model-00076-of-00081.safetensors", "model.layers.90.self_attn.q_proj.weight": "model-00076-of-00081.safetensors", "model.layers.90.self_attn.v_proj.weight": "model-00076-of-00081.safetensors", "model.layers.91.input_layernorm.weight": "model-00076-of-00081.safetensors", "model.layers.91.mlp.down_proj.weight": "model-00076-of-00081.safetensors", "model.layers.91.mlp.gate_proj.weight": "model-00076-of-00081.safetensors", "model.layers.91.mlp.up_proj.weight": "model-00076-of-00081.safetensors", "model.layers.91.post_attention_layernorm.weight": "model-00076-of-00081.safetensors", "model.layers.91.self_attn.k_proj.weight": "model-00076-of-00081.safetensors", "model.layers.91.self_attn.o_proj.weight": "model-00076-of-00081.safetensors", "model.layers.91.self_attn.q_proj.weight": "model-00076-of-00081.safetensors", "model.layers.91.self_attn.v_proj.weight": "model-00076-of-00081.safetensors", "model.layers.92.input_layernorm.weight": "model-00076-of-00081.safetensors", "model.layers.92.mlp.down_proj.weight": "model-00076-of-00081.safetensors", "model.layers.92.mlp.gate_proj.weight": "model-00077-of-00081.safetensors", "model.layers.92.mlp.up_proj.weight": "model-00077-of-00081.safetensors", "model.layers.92.post_attention_layernorm.weight": "model-00077-of-00081.safetensors", "model.layers.92.self_attn.k_proj.weight": "model-00077-of-00081.safetensors", "model.layers.92.self_attn.o_proj.weight": "model-00077-of-00081.safetensors", "model.layers.92.self_attn.q_proj.weight": "model-00077-of-00081.safetensors", "model.layers.92.self_attn.v_proj.weight": "model-00077-of-00081.safetensors", "model.layers.93.input_layernorm.weight": "model-00077-of-00081.safetensors", "model.layers.93.mlp.down_proj.weight": "model-00077-of-00081.safetensors", "model.layers.93.mlp.gate_proj.weight": "model-00077-of-00081.safetensors", "model.layers.93.mlp.up_proj.weight": "model-00077-of-00081.safetensors", "model.layers.93.post_attention_layernorm.weight": "model-00077-of-00081.safetensors", "model.layers.93.self_attn.k_proj.weight": "model-00077-of-00081.safetensors", "model.layers.93.self_attn.o_proj.weight": "model-00077-of-00081.safetensors", "model.layers.93.self_attn.q_proj.weight": "model-00077-of-00081.safetensors", "model.layers.93.self_attn.v_proj.weight": "model-00077-of-00081.safetensors", "model.layers.94.input_layernorm.weight": "model-00077-of-00081.safetensors", "model.layers.94.mlp.down_proj.weight": "model-00078-of-00081.safetensors", "model.layers.94.mlp.gate_proj.weight": "model-00078-of-00081.safetensors", "model.layers.94.mlp.up_proj.weight": "model-00078-of-00081.safetensors", "model.layers.94.post_attention_layernorm.weight": "model-00078-of-00081.safetensors", "model.layers.94.self_attn.k_proj.weight": "model-00078-of-00081.safetensors", "model.layers.94.self_attn.o_proj.weight": "model-00078-of-00081.safetensors", "model.layers.94.self_attn.q_proj.weight": "model-00078-of-00081.safetensors", "model.layers.94.self_attn.v_proj.weight": "model-00078-of-00081.safetensors", "model.layers.95.input_layernorm.weight": "model-00078-of-00081.safetensors", "model.layers.95.mlp.down_proj.weight": "model-00078-of-00081.safetensors", "model.layers.95.mlp.gate_proj.weight": "model-00078-of-00081.safetensors", "model.layers.95.mlp.up_proj.weight": "model-00078-of-00081.safetensors", "model.layers.95.post_attention_layernorm.weight": "model-00078-of-00081.safetensors", "model.layers.95.self_attn.k_proj.weight": "model-00078-of-00081.safetensors", "model.layers.95.self_attn.o_proj.weight": "model-00079-of-00081.safetensors", "model.layers.95.self_attn.q_proj.weight": "model-00079-of-00081.safetensors", "model.layers.95.self_attn.v_proj.weight": "model-00079-of-00081.safetensors", "model.layers.96.input_layernorm.weight": "model-00079-of-00081.safetensors", "model.layers.96.mlp.down_proj.weight": "model-00079-of-00081.safetensors", "model.layers.96.mlp.gate_proj.weight": "model-00079-of-00081.safetensors", "model.layers.96.mlp.up_proj.weight": "model-00079-of-00081.safetensors", "model.layers.96.post_attention_layernorm.weight": "model-00079-of-00081.safetensors", "model.layers.96.self_attn.k_proj.weight": "model-00079-of-00081.safetensors", "model.layers.96.self_attn.o_proj.weight": "model-00079-of-00081.safetensors", "model.layers.96.self_attn.q_proj.weight": "model-00079-of-00081.safetensors", "model.layers.96.self_attn.v_proj.weight": "model-00079-of-00081.safetensors", "model.layers.97.input_layernorm.weight": "model-00079-of-00081.safetensors", "model.layers.97.mlp.down_proj.weight": "model-00079-of-00081.safetensors", "model.layers.97.mlp.gate_proj.weight": "model-00079-of-00081.safetensors", "model.layers.97.mlp.up_proj.weight": "model-00080-of-00081.safetensors", "model.layers.97.post_attention_layernorm.weight": "model-00080-of-00081.safetensors", "model.layers.97.self_attn.k_proj.weight": "model-00080-of-00081.safetensors", "model.layers.97.self_attn.o_proj.weight": "model-00080-of-00081.safetensors", "model.layers.97.self_attn.q_proj.weight": "model-00080-of-00081.safetensors", "model.layers.97.self_attn.v_proj.weight": "model-00080-of-00081.safetensors", "model.layers.98.input_layernorm.weight": "model-00080-of-00081.safetensors", "model.layers.98.mlp.down_proj.weight": "model-00080-of-00081.safetensors", "model.layers.98.mlp.gate_proj.weight": "model-00080-of-00081.safetensors", "model.layers.98.mlp.up_proj.weight": "model-00080-of-00081.safetensors", "model.layers.98.post_attention_layernorm.weight": "model-00080-of-00081.safetensors", "model.layers.98.self_attn.k_proj.weight": "model-00080-of-00081.safetensors", "model.layers.98.self_attn.o_proj.weight": "model-00080-of-00081.safetensors", "model.layers.98.self_attn.q_proj.weight": "model-00080-of-00081.safetensors", "model.layers.98.self_attn.v_proj.weight": "model-00080-of-00081.safetensors", "model.layers.99.input_layernorm.weight": "model-00080-of-00081.safetensors", "model.layers.99.mlp.down_proj.weight": "model-00080-of-00081.safetensors", "model.layers.99.mlp.gate_proj.weight": "model-00081-of-00081.safetensors", "model.layers.99.mlp.up_proj.weight": "model-00081-of-00081.safetensors", "model.layers.99.post_attention_layernorm.weight": "model-00081-of-00081.safetensors", "model.layers.99.self_attn.k_proj.weight": "model-00081-of-00081.safetensors", "model.layers.99.self_attn.o_proj.weight": "model-00081-of-00081.safetensors", "model.layers.99.self_attn.q_proj.weight": "model-00081-of-00081.safetensors", "model.layers.99.self_attn.v_proj.weight": "model-00081-of-00081.safetensors", "model.norm.weight": "model-00081-of-00081.safetensors"}}

output-00001-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb6b6eebd4b5712b92abb09ff233ceb9d6aaa25d42f34d5098dbe58dd3ef503d
+size 8511614020

output-00002-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a4f5982b3826221c794c851456db92d6f82381603039187c1a180fb68b155f5
+size 8586988174

output-00003-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3bf728d7df6189b7df478a21c97245170d37354ae42d12cc6d9504c98d8eeb14
+size 8566588124

output-00004-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35914e6129657e406805ab5aee177eaa4f51174ec0cbc5421d82898219f07770
+size 8503536258

output-00005-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1186dd608f8bbdf33d1b79caf6d5ce2d7783315a85d3c6ecee228cc060f3932c
+size 8582640140

output-00006-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:af4076f8458a239ae937f6d671a9da7a18b8857b5c2be6f5e32873aabfd560bd
+size 8559695636

output-00007-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f206dcb8bdae98e8a03e7303d97f78213416b9b30db5781c208479771e95e23
+size 8587700226

output-00008-of-00008.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b29a0c20123bc9a989c44249718e4b50226ed4e7a71ff7d61d914d22b278c17
+size 6629678740

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59f95e28944c062244741268596badc900df86c7f5ded05088d2da22a7379e06
+size 587583