Upload ExaoneForCausalLM

config.json

@@ -0,0 +1,33 @@
+{
+  "_name_or_path": "LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct",
+  "activation_function": "silu",
+  "architectures": [
+    "ExaoneForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "auto_map": {
+    "AutoConfig": "LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct--configuration_exaone.ExaoneConfig",
+    "AutoModelForCausalLM": "LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct--modeling_exaone.ExaoneForCausalLM",
+    "AutoModelForSequenceClassification": "LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct--modeling_exaone.ExaoneForSequenceClassification"
+  },
+  "bos_token_id": 1,
+  "embed_dropout": 0.0,
+  "eos_token_id": 361,
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "layer_norm_epsilon": 1e-05,
+  "max_position_embeddings": 4096,
+  "model_type": "exaone",
+  "num_attention_heads": 32,
+  "num_key_value_heads": 8,
+  "num_layers": 32,
+  "pad_token_id": 0,
+  "rope_scaling": null,
+  "rope_theta": 500000.0,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.46.2",
+  "use_cache": true,
+  "vocab_size": 102400
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "eos_token_id": 361,
+  "pad_token_id": 0,
+  "transformers_version": "4.46.2"
+}

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f54165e07c670c590830dbcd9f7736c7d1966b7606040206311451b56e5e05eb
+size 4966229888

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0eb752078cf24a313e54192cd966080e3390c526fa06306bb1fd426a20f5365d
+size 4915915984

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a115931d89038b854f93e6c4876ee9a881932a1d8c7247a52dbd4376080a959
+size 4915924296

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:806f3e67b07fa035afb1f5f68dac3e4e150016596029be74c263e52d7a48ef82
+size 838860928

model.safetensors.index.json

@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 15636897792
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00004-of-00004.safetensors",
+    "transformer.h.0.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.0.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.0.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.0.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.0.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.0.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.0.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.0.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.0.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.1.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.10.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.10.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.10.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.10.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.10.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.10.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.10.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.10.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.10.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.11.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.12.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.13.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.14.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.15.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.16.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.17.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.18.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.19.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.2.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.2.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.2.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.2.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.2.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.2.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.2.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.2.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.2.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.20.attn.attention.k_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.20.attn.attention.out_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.20.attn.attention.q_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.20.attn.attention.v_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.20.ln_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.20.ln_2.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.20.mlp.c_fc_0.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.20.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.20.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.21.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.22.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.23.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.24.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.25.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.26.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.27.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.28.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.29.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.3.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.3.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.3.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.3.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.3.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.3.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.3.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.3.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.3.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.30.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.30.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.30.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.30.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.30.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.30.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.30.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.30.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.30.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.attn.attention.k_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.attn.attention.out_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.attn.attention.q_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.attn.attention.v_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.ln_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.ln_2.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.mlp.c_fc_0.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.mlp.c_fc_1.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.31.mlp.c_proj.weight": "model-00003-of-00004.safetensors",
+    "transformer.h.4.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.4.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.4.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.4.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.4.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.4.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.4.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.4.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.4.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.5.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.6.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.7.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.mlp.c_fc_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.8.mlp.c_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.9.attn.attention.k_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.9.attn.attention.out_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.9.attn.attention.q_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.9.attn.attention.v_proj.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.9.ln_1.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.9.ln_2.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.9.mlp.c_fc_0.weight": "model-00001-of-00004.safetensors",
+    "transformer.h.9.mlp.c_fc_1.weight": "model-00002-of-00004.safetensors",
+    "transformer.h.9.mlp.c_proj.weight": "model-00002-of-00004.safetensors",
+    "transformer.ln_f.weight": "model-00003-of-00004.safetensors",
+    "transformer.wte.weight": "model-00001-of-00004.safetensors"
+  }
+}