upload model

config.json

@@ -0,0 +1,28 @@
+{
+  "activation_function": "gelu_pytorch_tanh",
+  "architectures": [
+    "GPTBigCodeForCausalLM"
+  ],
+  "attention_softmax_in_fp32": true,
+  "attn_pdrop": 0.1,
+  "bos_token_id": 0,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 0,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "gpt_bigcode",
+  "multi_query": true,
+  "n_embd": 6144,
+  "n_head": 48,
+  "n_inner": 24576,
+  "n_layer": 88,
+  "n_positions": 8192,
+  "pad_token_id": 0,
+  "resid_pdrop": 0.1,
+  "scale_attention_softmax_in_fp32": true,
+  "scale_attn_weights": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.41.0.dev0",
+  "use_cache": false,
+  "vocab_size": 49152
+}

generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "pad_token_id": 0,
+  "transformers_version": "4.41.0.dev0"
+}

model-00001-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36414f8f77469990d7ba1c82df700bd39d803754fe309e6f1a8d65f1e9063f9e
+size 4952185256

model-00002-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:286215ef6c205ad32e84c3cf9111565d1353fad0071e57b352f7117934bca4a4
+size 4930227792

model-00003-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e61edf7c3a56023e1dbc1dbf20299ed5362a4894ce6f83cbd564b03c9f8b4e38
+size 4927118464

model-00004-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ae8241e0a4d865c94a175e503741b9cf036c8b92a156e557ee51f74b910cfcf
+size 4930227864

model-00005-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c54d554317c073acb5578f1f83c5942d35619bc928b83d0979f7b76f993b284e
+size 4927118464

model-00006-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cee7d8513bfc83f1fdf565c5beab489af624bdf857d559e287111ada10cced4
+size 4930227864

model-00007-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:330d7e9fd9dad0dde381daa0a673ffa2deaeed698622401e77d98ab9c42c0900
+size 4927118464

model-00008-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb66fbd2cbff2ae7c3f1f5479c8df12813ce792d2bbe8ae582b40a122f7d3b2f
+size 4930227864

model-00009-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5606598d155dc829f12c4fd6871d0e5d65fdb23fb8b8450a1907722e3f1da905
+size 4927118464

model-00010-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec26ffba465b25a25b6272a264ee44cc1b3fad9e152ece959535d3d8de7038a3
+size 4930227864

model-00011-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ed405fca3d01901b0b6d8db4a3e868dcdc96b84b45e756e87f83f8079315193
+size 4927118464

model-00012-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6a3584aec66ab0d402126f8d2310952843d514a2d636618d2f1732a62836f67
+size 4930227864

model-00013-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afc6745525ba75facf9adc67e5aeae4d501785384b6b5e153824f6a8a0f8fd1b
+size 4927118464

model-00014-of-00014.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25a548342740e93e5933da2879c274ec3136caaf8738d569774fb291b43b96b4
+size 3335056808

model.safetensors.index.json

@@ -0,0 +1,1067 @@
+{
+  "metadata": {
+    "total_size": 67431206912
+  },
+  "weight_map": {
+    "transformer.h.0.attn.c_attn.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.0.attn.c_attn.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.0.attn.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.0.attn.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.0.ln_1.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.0.ln_1.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.0.ln_2.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.0.ln_2.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.0.mlp.c_fc.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.0.mlp.c_fc.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.0.mlp.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.0.mlp.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.1.attn.c_attn.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.1.attn.c_attn.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.1.attn.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.1.attn.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.1.ln_1.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.1.ln_1.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.1.ln_2.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.1.ln_2.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.1.mlp.c_fc.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.1.mlp.c_fc.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.1.mlp.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.1.mlp.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.10.attn.c_attn.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.10.attn.c_attn.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.10.attn.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.10.attn.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.10.ln_1.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.10.ln_1.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.10.ln_2.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.10.ln_2.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.10.mlp.c_fc.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.10.mlp.c_fc.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.10.mlp.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.10.mlp.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.11.attn.c_attn.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.11.attn.c_attn.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.11.attn.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.11.attn.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.11.ln_1.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.11.ln_1.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.11.ln_2.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.11.ln_2.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.11.mlp.c_fc.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.11.mlp.c_fc.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.11.mlp.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.11.mlp.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.12.attn.c_attn.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.12.attn.c_attn.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.12.attn.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.12.attn.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.12.ln_1.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.12.ln_1.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.12.ln_2.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.12.ln_2.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.12.mlp.c_fc.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.12.mlp.c_fc.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.12.mlp.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.12.mlp.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.13.attn.c_attn.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.13.attn.c_attn.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.13.attn.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.13.attn.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.13.ln_1.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.13.ln_1.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.13.ln_2.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.13.ln_2.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.13.mlp.c_fc.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.13.mlp.c_fc.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.13.mlp.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.13.mlp.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.14.attn.c_attn.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.14.attn.c_attn.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.14.attn.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.14.attn.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.14.ln_1.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.14.ln_1.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.14.ln_2.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.14.ln_2.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.14.mlp.c_fc.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.14.mlp.c_fc.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.14.mlp.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.14.mlp.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.15.attn.c_attn.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.15.attn.c_attn.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.15.attn.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.15.attn.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.15.ln_1.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.15.ln_1.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.15.ln_2.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.15.ln_2.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.15.mlp.c_fc.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.15.mlp.c_fc.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.15.mlp.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.15.mlp.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.16.attn.c_attn.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.16.attn.c_attn.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.16.attn.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.16.attn.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.16.ln_1.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.16.ln_1.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.16.ln_2.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.16.ln_2.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.16.mlp.c_fc.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.16.mlp.c_fc.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.16.mlp.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.16.mlp.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.17.attn.c_attn.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.17.attn.c_attn.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.17.attn.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.17.attn.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.17.ln_1.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.17.ln_1.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.17.ln_2.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.17.ln_2.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.17.mlp.c_fc.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.17.mlp.c_fc.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.17.mlp.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.17.mlp.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.18.attn.c_attn.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.18.attn.c_attn.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.18.attn.c_proj.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.18.attn.c_proj.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.18.ln_1.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.18.ln_1.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.18.ln_2.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.18.ln_2.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.18.mlp.c_fc.bias": "model-00003-of-00014.safetensors",
+    "transformer.h.18.mlp.c_fc.weight": "model-00003-of-00014.safetensors",
+    "transformer.h.18.mlp.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.18.mlp.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.19.attn.c_attn.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.19.attn.c_attn.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.19.attn.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.19.attn.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.19.ln_1.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.19.ln_1.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.19.ln_2.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.19.ln_2.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.19.mlp.c_fc.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.19.mlp.c_fc.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.19.mlp.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.19.mlp.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.2.attn.c_attn.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.2.attn.c_attn.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.2.attn.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.2.attn.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.2.ln_1.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.2.ln_1.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.2.ln_2.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.2.ln_2.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.2.mlp.c_fc.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.2.mlp.c_fc.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.2.mlp.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.2.mlp.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.20.attn.c_attn.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.20.attn.c_attn.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.20.attn.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.20.attn.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.20.ln_1.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.20.ln_1.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.20.ln_2.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.20.ln_2.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.20.mlp.c_fc.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.20.mlp.c_fc.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.20.mlp.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.20.mlp.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.21.attn.c_attn.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.21.attn.c_attn.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.21.attn.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.21.attn.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.21.ln_1.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.21.ln_1.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.21.ln_2.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.21.ln_2.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.21.mlp.c_fc.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.21.mlp.c_fc.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.21.mlp.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.21.mlp.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.22.attn.c_attn.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.22.attn.c_attn.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.22.attn.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.22.attn.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.22.ln_1.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.22.ln_1.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.22.ln_2.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.22.ln_2.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.22.mlp.c_fc.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.22.mlp.c_fc.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.22.mlp.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.22.mlp.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.23.attn.c_attn.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.23.attn.c_attn.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.23.attn.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.23.attn.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.23.ln_1.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.23.ln_1.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.23.ln_2.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.23.ln_2.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.23.mlp.c_fc.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.23.mlp.c_fc.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.23.mlp.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.23.mlp.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.24.attn.c_attn.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.24.attn.c_attn.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.24.attn.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.24.attn.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.24.ln_1.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.24.ln_1.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.24.ln_2.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.24.ln_2.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.24.mlp.c_fc.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.24.mlp.c_fc.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.24.mlp.c_proj.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.24.mlp.c_proj.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.25.attn.c_attn.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.25.attn.c_attn.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.25.attn.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.25.attn.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.25.ln_1.bias": "model-00004-of-00014.safetensors",
+    "transformer.h.25.ln_1.weight": "model-00004-of-00014.safetensors",
+    "transformer.h.25.ln_2.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.25.ln_2.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.25.mlp.c_fc.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.25.mlp.c_fc.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.25.mlp.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.25.mlp.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.26.attn.c_attn.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.26.attn.c_attn.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.26.attn.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.26.attn.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.26.ln_1.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.26.ln_1.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.26.ln_2.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.26.ln_2.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.26.mlp.c_fc.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.26.mlp.c_fc.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.26.mlp.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.26.mlp.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.27.attn.c_attn.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.27.attn.c_attn.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.27.attn.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.27.attn.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.27.ln_1.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.27.ln_1.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.27.ln_2.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.27.ln_2.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.27.mlp.c_fc.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.27.mlp.c_fc.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.27.mlp.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.27.mlp.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.28.attn.c_attn.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.28.attn.c_attn.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.28.attn.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.28.attn.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.28.ln_1.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.28.ln_1.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.28.ln_2.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.28.ln_2.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.28.mlp.c_fc.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.28.mlp.c_fc.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.28.mlp.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.28.mlp.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.29.attn.c_attn.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.29.attn.c_attn.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.29.attn.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.29.attn.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.29.ln_1.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.29.ln_1.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.29.ln_2.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.29.ln_2.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.29.mlp.c_fc.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.29.mlp.c_fc.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.29.mlp.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.29.mlp.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.3.attn.c_attn.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.3.attn.c_attn.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.3.attn.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.3.attn.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.3.ln_1.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.3.ln_1.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.3.ln_2.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.3.ln_2.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.3.mlp.c_fc.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.3.mlp.c_fc.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.3.mlp.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.3.mlp.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.30.attn.c_attn.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.30.attn.c_attn.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.30.attn.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.30.attn.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.30.ln_1.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.30.ln_1.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.30.ln_2.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.30.ln_2.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.30.mlp.c_fc.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.30.mlp.c_fc.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.30.mlp.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.30.mlp.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.31.attn.c_attn.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.31.attn.c_attn.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.31.attn.c_proj.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.31.attn.c_proj.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.31.ln_1.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.31.ln_1.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.31.ln_2.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.31.ln_2.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.31.mlp.c_fc.bias": "model-00005-of-00014.safetensors",
+    "transformer.h.31.mlp.c_fc.weight": "model-00005-of-00014.safetensors",
+    "transformer.h.31.mlp.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.31.mlp.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.32.attn.c_attn.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.32.attn.c_attn.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.32.attn.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.32.attn.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.32.ln_1.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.32.ln_1.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.32.ln_2.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.32.ln_2.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.32.mlp.c_fc.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.32.mlp.c_fc.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.32.mlp.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.32.mlp.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.33.attn.c_attn.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.33.attn.c_attn.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.33.attn.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.33.attn.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.33.ln_1.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.33.ln_1.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.33.ln_2.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.33.ln_2.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.33.mlp.c_fc.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.33.mlp.c_fc.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.33.mlp.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.33.mlp.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.34.attn.c_attn.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.34.attn.c_attn.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.34.attn.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.34.attn.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.34.ln_1.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.34.ln_1.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.34.ln_2.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.34.ln_2.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.34.mlp.c_fc.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.34.mlp.c_fc.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.34.mlp.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.34.mlp.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.35.attn.c_attn.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.35.attn.c_attn.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.35.attn.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.35.attn.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.35.ln_1.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.35.ln_1.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.35.ln_2.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.35.ln_2.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.35.mlp.c_fc.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.35.mlp.c_fc.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.35.mlp.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.35.mlp.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.36.attn.c_attn.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.36.attn.c_attn.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.36.attn.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.36.attn.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.36.ln_1.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.36.ln_1.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.36.ln_2.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.36.ln_2.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.36.mlp.c_fc.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.36.mlp.c_fc.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.36.mlp.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.36.mlp.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.37.attn.c_attn.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.37.attn.c_attn.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.37.attn.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.37.attn.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.37.ln_1.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.37.ln_1.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.37.ln_2.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.37.ln_2.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.37.mlp.c_fc.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.37.mlp.c_fc.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.37.mlp.c_proj.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.37.mlp.c_proj.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.38.attn.c_attn.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.38.attn.c_attn.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.38.attn.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.38.attn.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.38.ln_1.bias": "model-00006-of-00014.safetensors",
+    "transformer.h.38.ln_1.weight": "model-00006-of-00014.safetensors",
+    "transformer.h.38.ln_2.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.38.ln_2.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.38.mlp.c_fc.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.38.mlp.c_fc.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.38.mlp.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.38.mlp.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.39.attn.c_attn.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.39.attn.c_attn.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.39.attn.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.39.attn.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.39.ln_1.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.39.ln_1.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.39.ln_2.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.39.ln_2.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.39.mlp.c_fc.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.39.mlp.c_fc.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.39.mlp.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.39.mlp.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.4.attn.c_attn.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.4.attn.c_attn.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.4.attn.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.4.attn.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.4.ln_1.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.4.ln_1.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.4.ln_2.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.4.ln_2.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.4.mlp.c_fc.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.4.mlp.c_fc.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.4.mlp.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.4.mlp.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.40.attn.c_attn.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.40.attn.c_attn.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.40.attn.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.40.attn.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.40.ln_1.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.40.ln_1.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.40.ln_2.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.40.ln_2.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.40.mlp.c_fc.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.40.mlp.c_fc.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.40.mlp.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.40.mlp.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.41.attn.c_attn.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.41.attn.c_attn.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.41.attn.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.41.attn.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.41.ln_1.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.41.ln_1.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.41.ln_2.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.41.ln_2.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.41.mlp.c_fc.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.41.mlp.c_fc.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.41.mlp.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.41.mlp.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.42.attn.c_attn.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.42.attn.c_attn.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.42.attn.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.42.attn.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.42.ln_1.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.42.ln_1.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.42.ln_2.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.42.ln_2.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.42.mlp.c_fc.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.42.mlp.c_fc.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.42.mlp.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.42.mlp.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.43.attn.c_attn.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.43.attn.c_attn.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.43.attn.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.43.attn.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.43.ln_1.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.43.ln_1.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.43.ln_2.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.43.ln_2.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.43.mlp.c_fc.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.43.mlp.c_fc.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.43.mlp.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.43.mlp.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.44.attn.c_attn.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.44.attn.c_attn.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.44.attn.c_proj.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.44.attn.c_proj.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.44.ln_1.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.44.ln_1.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.44.ln_2.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.44.ln_2.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.44.mlp.c_fc.bias": "model-00007-of-00014.safetensors",
+    "transformer.h.44.mlp.c_fc.weight": "model-00007-of-00014.safetensors",
+    "transformer.h.44.mlp.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.44.mlp.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.45.attn.c_attn.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.45.attn.c_attn.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.45.attn.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.45.attn.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.45.ln_1.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.45.ln_1.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.45.ln_2.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.45.ln_2.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.45.mlp.c_fc.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.45.mlp.c_fc.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.45.mlp.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.45.mlp.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.46.attn.c_attn.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.46.attn.c_attn.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.46.attn.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.46.attn.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.46.ln_1.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.46.ln_1.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.46.ln_2.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.46.ln_2.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.46.mlp.c_fc.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.46.mlp.c_fc.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.46.mlp.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.46.mlp.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.47.attn.c_attn.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.47.attn.c_attn.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.47.attn.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.47.attn.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.47.ln_1.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.47.ln_1.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.47.ln_2.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.47.ln_2.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.47.mlp.c_fc.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.47.mlp.c_fc.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.47.mlp.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.47.mlp.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.48.attn.c_attn.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.48.attn.c_attn.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.48.attn.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.48.attn.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.48.ln_1.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.48.ln_1.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.48.ln_2.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.48.ln_2.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.48.mlp.c_fc.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.48.mlp.c_fc.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.48.mlp.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.48.mlp.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.49.attn.c_attn.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.49.attn.c_attn.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.49.attn.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.49.attn.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.49.ln_1.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.49.ln_1.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.49.ln_2.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.49.ln_2.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.49.mlp.c_fc.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.49.mlp.c_fc.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.49.mlp.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.49.mlp.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.5.attn.c_attn.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.5.attn.c_attn.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.5.attn.c_proj.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.5.attn.c_proj.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.5.ln_1.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.5.ln_1.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.5.ln_2.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.5.ln_2.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.5.mlp.c_fc.bias": "model-00001-of-00014.safetensors",
+    "transformer.h.5.mlp.c_fc.weight": "model-00001-of-00014.safetensors",
+    "transformer.h.5.mlp.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.5.mlp.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.50.attn.c_attn.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.50.attn.c_attn.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.50.attn.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.50.attn.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.50.ln_1.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.50.ln_1.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.50.ln_2.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.50.ln_2.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.50.mlp.c_fc.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.50.mlp.c_fc.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.50.mlp.c_proj.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.50.mlp.c_proj.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.51.attn.c_attn.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.51.attn.c_attn.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.51.attn.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.51.attn.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.51.ln_1.bias": "model-00008-of-00014.safetensors",
+    "transformer.h.51.ln_1.weight": "model-00008-of-00014.safetensors",
+    "transformer.h.51.ln_2.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.51.ln_2.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.51.mlp.c_fc.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.51.mlp.c_fc.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.51.mlp.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.51.mlp.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.52.attn.c_attn.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.52.attn.c_attn.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.52.attn.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.52.attn.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.52.ln_1.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.52.ln_1.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.52.ln_2.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.52.ln_2.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.52.mlp.c_fc.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.52.mlp.c_fc.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.52.mlp.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.52.mlp.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.53.attn.c_attn.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.53.attn.c_attn.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.53.attn.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.53.attn.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.53.ln_1.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.53.ln_1.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.53.ln_2.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.53.ln_2.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.53.mlp.c_fc.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.53.mlp.c_fc.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.53.mlp.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.53.mlp.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.54.attn.c_attn.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.54.attn.c_attn.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.54.attn.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.54.attn.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.54.ln_1.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.54.ln_1.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.54.ln_2.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.54.ln_2.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.54.mlp.c_fc.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.54.mlp.c_fc.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.54.mlp.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.54.mlp.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.55.attn.c_attn.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.55.attn.c_attn.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.55.attn.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.55.attn.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.55.ln_1.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.55.ln_1.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.55.ln_2.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.55.ln_2.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.55.mlp.c_fc.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.55.mlp.c_fc.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.55.mlp.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.55.mlp.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.56.attn.c_attn.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.56.attn.c_attn.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.56.attn.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.56.attn.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.56.ln_1.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.56.ln_1.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.56.ln_2.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.56.ln_2.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.56.mlp.c_fc.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.56.mlp.c_fc.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.56.mlp.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.56.mlp.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.57.attn.c_attn.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.57.attn.c_attn.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.57.attn.c_proj.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.57.attn.c_proj.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.57.ln_1.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.57.ln_1.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.57.ln_2.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.57.ln_2.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.57.mlp.c_fc.bias": "model-00009-of-00014.safetensors",
+    "transformer.h.57.mlp.c_fc.weight": "model-00009-of-00014.safetensors",
+    "transformer.h.57.mlp.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.57.mlp.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.58.attn.c_attn.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.58.attn.c_attn.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.58.attn.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.58.attn.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.58.ln_1.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.58.ln_1.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.58.ln_2.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.58.ln_2.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.58.mlp.c_fc.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.58.mlp.c_fc.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.58.mlp.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.58.mlp.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.59.attn.c_attn.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.59.attn.c_attn.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.59.attn.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.59.attn.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.59.ln_1.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.59.ln_1.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.59.ln_2.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.59.ln_2.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.59.mlp.c_fc.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.59.mlp.c_fc.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.59.mlp.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.59.mlp.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.6.attn.c_attn.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.6.attn.c_attn.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.6.attn.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.6.attn.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.6.ln_1.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.6.ln_1.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.6.ln_2.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.6.ln_2.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.6.mlp.c_fc.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.6.mlp.c_fc.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.6.mlp.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.6.mlp.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.60.attn.c_attn.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.60.attn.c_attn.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.60.attn.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.60.attn.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.60.ln_1.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.60.ln_1.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.60.ln_2.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.60.ln_2.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.60.mlp.c_fc.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.60.mlp.c_fc.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.60.mlp.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.60.mlp.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.61.attn.c_attn.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.61.attn.c_attn.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.61.attn.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.61.attn.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.61.ln_1.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.61.ln_1.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.61.ln_2.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.61.ln_2.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.61.mlp.c_fc.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.61.mlp.c_fc.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.61.mlp.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.61.mlp.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.62.attn.c_attn.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.62.attn.c_attn.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.62.attn.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.62.attn.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.62.ln_1.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.62.ln_1.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.62.ln_2.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.62.ln_2.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.62.mlp.c_fc.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.62.mlp.c_fc.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.62.mlp.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.62.mlp.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.63.attn.c_attn.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.63.attn.c_attn.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.63.attn.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.63.attn.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.63.ln_1.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.63.ln_1.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.63.ln_2.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.63.ln_2.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.63.mlp.c_fc.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.63.mlp.c_fc.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.63.mlp.c_proj.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.63.mlp.c_proj.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.64.attn.c_attn.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.64.attn.c_attn.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.64.attn.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.64.attn.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.64.ln_1.bias": "model-00010-of-00014.safetensors",
+    "transformer.h.64.ln_1.weight": "model-00010-of-00014.safetensors",
+    "transformer.h.64.ln_2.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.64.ln_2.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.64.mlp.c_fc.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.64.mlp.c_fc.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.64.mlp.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.64.mlp.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.65.attn.c_attn.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.65.attn.c_attn.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.65.attn.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.65.attn.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.65.ln_1.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.65.ln_1.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.65.ln_2.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.65.ln_2.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.65.mlp.c_fc.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.65.mlp.c_fc.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.65.mlp.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.65.mlp.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.66.attn.c_attn.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.66.attn.c_attn.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.66.attn.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.66.attn.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.66.ln_1.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.66.ln_1.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.66.ln_2.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.66.ln_2.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.66.mlp.c_fc.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.66.mlp.c_fc.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.66.mlp.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.66.mlp.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.67.attn.c_attn.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.67.attn.c_attn.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.67.attn.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.67.attn.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.67.ln_1.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.67.ln_1.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.67.ln_2.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.67.ln_2.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.67.mlp.c_fc.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.67.mlp.c_fc.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.67.mlp.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.67.mlp.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.68.attn.c_attn.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.68.attn.c_attn.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.68.attn.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.68.attn.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.68.ln_1.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.68.ln_1.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.68.ln_2.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.68.ln_2.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.68.mlp.c_fc.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.68.mlp.c_fc.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.68.mlp.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.68.mlp.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.69.attn.c_attn.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.69.attn.c_attn.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.69.attn.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.69.attn.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.69.ln_1.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.69.ln_1.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.69.ln_2.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.69.ln_2.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.69.mlp.c_fc.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.69.mlp.c_fc.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.69.mlp.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.69.mlp.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.7.attn.c_attn.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.7.attn.c_attn.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.7.attn.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.7.attn.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.7.ln_1.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.7.ln_1.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.7.ln_2.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.7.ln_2.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.7.mlp.c_fc.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.7.mlp.c_fc.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.7.mlp.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.7.mlp.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.70.attn.c_attn.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.70.attn.c_attn.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.70.attn.c_proj.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.70.attn.c_proj.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.70.ln_1.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.70.ln_1.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.70.ln_2.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.70.ln_2.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.70.mlp.c_fc.bias": "model-00011-of-00014.safetensors",
+    "transformer.h.70.mlp.c_fc.weight": "model-00011-of-00014.safetensors",
+    "transformer.h.70.mlp.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.70.mlp.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.71.attn.c_attn.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.71.attn.c_attn.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.71.attn.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.71.attn.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.71.ln_1.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.71.ln_1.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.71.ln_2.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.71.ln_2.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.71.mlp.c_fc.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.71.mlp.c_fc.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.71.mlp.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.71.mlp.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.72.attn.c_attn.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.72.attn.c_attn.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.72.attn.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.72.attn.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.72.ln_1.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.72.ln_1.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.72.ln_2.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.72.ln_2.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.72.mlp.c_fc.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.72.mlp.c_fc.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.72.mlp.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.72.mlp.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.73.attn.c_attn.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.73.attn.c_attn.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.73.attn.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.73.attn.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.73.ln_1.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.73.ln_1.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.73.ln_2.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.73.ln_2.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.73.mlp.c_fc.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.73.mlp.c_fc.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.73.mlp.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.73.mlp.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.74.attn.c_attn.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.74.attn.c_attn.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.74.attn.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.74.attn.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.74.ln_1.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.74.ln_1.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.74.ln_2.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.74.ln_2.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.74.mlp.c_fc.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.74.mlp.c_fc.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.74.mlp.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.74.mlp.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.75.attn.c_attn.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.75.attn.c_attn.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.75.attn.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.75.attn.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.75.ln_1.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.75.ln_1.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.75.ln_2.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.75.ln_2.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.75.mlp.c_fc.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.75.mlp.c_fc.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.75.mlp.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.75.mlp.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.76.attn.c_attn.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.76.attn.c_attn.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.76.attn.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.76.attn.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.76.ln_1.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.76.ln_1.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.76.ln_2.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.76.ln_2.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.76.mlp.c_fc.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.76.mlp.c_fc.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.76.mlp.c_proj.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.76.mlp.c_proj.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.77.attn.c_attn.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.77.attn.c_attn.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.77.attn.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.77.attn.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.77.ln_1.bias": "model-00012-of-00014.safetensors",
+    "transformer.h.77.ln_1.weight": "model-00012-of-00014.safetensors",
+    "transformer.h.77.ln_2.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.77.ln_2.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.77.mlp.c_fc.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.77.mlp.c_fc.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.77.mlp.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.77.mlp.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.78.attn.c_attn.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.78.attn.c_attn.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.78.attn.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.78.attn.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.78.ln_1.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.78.ln_1.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.78.ln_2.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.78.ln_2.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.78.mlp.c_fc.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.78.mlp.c_fc.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.78.mlp.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.78.mlp.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.79.attn.c_attn.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.79.attn.c_attn.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.79.attn.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.79.attn.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.79.ln_1.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.79.ln_1.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.79.ln_2.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.79.ln_2.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.79.mlp.c_fc.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.79.mlp.c_fc.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.79.mlp.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.79.mlp.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.8.attn.c_attn.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.8.attn.c_attn.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.8.attn.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.8.attn.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.8.ln_1.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.8.ln_1.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.8.ln_2.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.8.ln_2.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.8.mlp.c_fc.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.8.mlp.c_fc.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.8.mlp.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.8.mlp.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.80.attn.c_attn.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.80.attn.c_attn.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.80.attn.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.80.attn.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.80.ln_1.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.80.ln_1.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.80.ln_2.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.80.ln_2.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.80.mlp.c_fc.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.80.mlp.c_fc.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.80.mlp.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.80.mlp.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.81.attn.c_attn.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.81.attn.c_attn.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.81.attn.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.81.attn.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.81.ln_1.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.81.ln_1.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.81.ln_2.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.81.ln_2.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.81.mlp.c_fc.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.81.mlp.c_fc.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.81.mlp.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.81.mlp.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.82.attn.c_attn.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.82.attn.c_attn.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.82.attn.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.82.attn.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.82.ln_1.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.82.ln_1.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.82.ln_2.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.82.ln_2.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.82.mlp.c_fc.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.82.mlp.c_fc.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.82.mlp.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.82.mlp.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.83.attn.c_attn.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.83.attn.c_attn.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.83.attn.c_proj.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.83.attn.c_proj.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.83.ln_1.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.83.ln_1.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.83.ln_2.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.83.ln_2.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.83.mlp.c_fc.bias": "model-00013-of-00014.safetensors",
+    "transformer.h.83.mlp.c_fc.weight": "model-00013-of-00014.safetensors",
+    "transformer.h.83.mlp.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.83.mlp.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.84.attn.c_attn.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.84.attn.c_attn.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.84.attn.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.84.attn.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.84.ln_1.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.84.ln_1.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.84.ln_2.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.84.ln_2.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.84.mlp.c_fc.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.84.mlp.c_fc.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.84.mlp.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.84.mlp.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.85.attn.c_attn.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.85.attn.c_attn.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.85.attn.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.85.attn.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.85.ln_1.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.85.ln_1.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.85.ln_2.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.85.ln_2.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.85.mlp.c_fc.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.85.mlp.c_fc.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.85.mlp.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.85.mlp.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.86.attn.c_attn.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.86.attn.c_attn.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.86.attn.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.86.attn.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.86.ln_1.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.86.ln_1.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.86.ln_2.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.86.ln_2.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.86.mlp.c_fc.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.86.mlp.c_fc.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.86.mlp.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.86.mlp.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.87.attn.c_attn.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.87.attn.c_attn.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.87.attn.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.87.attn.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.87.ln_1.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.87.ln_1.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.87.ln_2.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.87.ln_2.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.87.mlp.c_fc.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.87.mlp.c_fc.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.87.mlp.c_proj.bias": "model-00014-of-00014.safetensors",
+    "transformer.h.87.mlp.c_proj.weight": "model-00014-of-00014.safetensors",
+    "transformer.h.9.attn.c_attn.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.9.attn.c_attn.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.9.attn.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.9.attn.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.9.ln_1.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.9.ln_1.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.9.ln_2.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.9.ln_2.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.9.mlp.c_fc.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.9.mlp.c_fc.weight": "model-00002-of-00014.safetensors",
+    "transformer.h.9.mlp.c_proj.bias": "model-00002-of-00014.safetensors",
+    "transformer.h.9.mlp.c_proj.weight": "model-00002-of-00014.safetensors",
+    "transformer.ln_f.bias": "model-00014-of-00014.safetensors",
+    "transformer.ln_f.weight": "model-00014-of-00014.safetensors",
+    "transformer.wpe.weight": "model-00001-of-00014.safetensors",
+    "transformer.wte.weight": "model-00001-of-00014.safetensors"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,51 @@
+{
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<fim_prefix>",
+    "<fim_middle>",
+    "<fim_suffix>",
+    "<fim_pad>",
+    "<filename>",
+    "<gh_stars>",
+    "<issue_start>",
+    "<issue_comment>",
+    "<issue_closed>",
+    "<jupyter_start>",
+    "<jupyter_text>",
+    "<jupyter_code>",
+    "<jupyter_output>",
+    "<empty_output>",
+    "<commit_before>",
+    "<commit_msg>",
+    "<commit_after>",
+    "<reponame>"
+  ],
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer_config.json

@@ -0,0 +1,187 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<fim_prefix>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "<fim_middle>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<fim_suffix>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "<fim_pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "5": {
+      "content": "<filename>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "6": {
+      "content": "<gh_stars>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "7": {
+      "content": "<issue_start>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "8": {
+      "content": "<issue_comment>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "9": {
+      "content": "<issue_closed>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "10": {
+      "content": "<jupyter_start>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "11": {
+      "content": "<jupyter_text>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "12": {
+      "content": "<jupyter_code>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "13": {
+      "content": "<jupyter_output>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "14": {
+      "content": "<empty_output>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "15": {
+      "content": "<commit_before>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "16": {
+      "content": "<commit_msg>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "17": {
+      "content": "<commit_after>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "18": {
+      "content": "<reponame>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<fim_prefix>",
+    "<fim_middle>",
+    "<fim_suffix>",
+    "<fim_pad>",
+    "<filename>",
+    "<gh_stars>",
+    "<issue_start>",
+    "<issue_comment>",
+    "<issue_closed>",
+    "<jupyter_start>",
+    "<jupyter_text>",
+    "<jupyter_code>",
+    "<jupyter_output>",
+    "<empty_output>",
+    "<commit_before>",
+    "<commit_msg>",
+    "<commit_after>",
+    "<reponame>"
+  ],
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 9223372036854775807,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>",
+  "vocab_size": 49152
+}