Add model as sharded safetensors

added_tokens.json

@@ -0,0 +1,3 @@
+{
+  "[PAD]": 49152
+}

config.json

@@ -0,0 +1,39 @@
+{
+  "_name_or_path": "/home/ubuntu/WizardCoder-15B-V1.0",
+  "activation_function": "gelu",
+  "architectures": [
+    "GPTBigCodeForCausalLM"
+  ],
+  "attention_softmax_in_fp32": true,
+  "attn_pdrop": 0.1,
+  "bos_token_id": 0,
+  "embd_pdrop": 0.1,
+  "eos_token_id": 0,
+  "inference_runner": 0,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "max_batch_size": null,
+  "max_sequence_length": null,
+  "model_type": "gpt_bigcode",
+  "multi_query": true,
+  "n_embd": 6144,
+  "n_head": 48,
+  "n_inner": 24576,
+  "n_layer": 40,
+  "n_positions": 8192,
+  "pad_key_length": true,
+  "pre_allocate_kv_cache": false,
+  "resid_pdrop": 0.1,
+  "scale_attention_softmax_in_fp32": true,
+  "scale_attn_weights": true,
+  "summary_activation": null,
+  "summary_first_dropout": 0.1,
+  "summary_proj_to_labels": true,
+  "summary_type": "cls_index",
+  "summary_use_proj": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.30.2",
+  "use_cache": false,
+  "validate_runner_input": true,
+  "vocab_size": 49153
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "transformers_version": "4.30.2"
+}

model-00001-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a1b2b514be061ecb1a45bff2f17d24f4866743d9d37c9e3717cd1450f8a3137
+size 1160910560

model-00002-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2f971ea67f7fbf8a5a5bfff54c4bb8f715bf56e263c75a696ac08b87d8013e2
+size 1214475520

model-00003-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f34527ee156aa5cb3d4e844b2a3fb882b10edb1011e2255cb859d44c0b16b636
+size 1060296640

model-00004-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a57d8a524e4d3a7c0358135b96c53ad95aff55c5cbf0301ccb8b83f946900fd7
+size 1214475520

model-00005-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61693dfcdc68cc799970a9a2a3653bb2e5c8dca17ccf1083634a887027ee2f83
+size 1060296640

model-00006-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fce3037da4be07e4707b22b15ed9ff3e3e9926ca3b1306b2b02a95f2a30dea30
+size 1214475520

model-00007-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f366f62966c93726f7b7226f5598a11c10984d56da424fea9618ec52c8f53a0
+size 1060296640

model-00008-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52aa6fa8533657653f471806aa0f52181f5bbb8f706acd31a27a5efd41aecdd2
+size 1214475520

model-00009-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:007768ec5046c24e84507d500cc50f8b9c4213f2b79684551e304b76d580c2ba
+size 1060296656

model-00010-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f15f93a38d7f413e2bc2c45c63ace2aa6b5b9bfd345d364682a7b34931835445
+size 1214475544

model-00011-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14d8579303bab97a13e7b4d610e500b230595b56f9f87639fa358f60829d92e9
+size 1060296656

model-00012-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:997ff30886912321c2e34308fd1673aa056e8e6ead141db98915ebed1d1c32db
+size 1214475544

model-00013-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44635613b08ea3254ff6667d130f101facc33be353165f3f9d9e247cb6440c74
+size 1060296656

model-00014-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:adfa82eb73b111e361c433ad5450e11eb1e2fa61d31c38d9c28d6f6025adfaeb
+size 1214475544

model-00015-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da4262807eea7af7f6db9a3eb091aad57c4dc093cc4b95aea95f0a7facda108a
+size 1060296656

model-00016-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a371c922ac5d489220951f611e3a6dadc98e19cdb89f4c2f8d17b6386ae1cb42
+size 1214475544

model-00017-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0eb71778d4bff6ab01469086c297880d358909b64c196177a77d66d93784048
+size 1060296656

model-00018-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b2bbc35c6d8482f193d19591cad0de0dcb37b7adce16826706bb595608abea6
+size 1214475544

model-00019-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80f84e9bb1384d65f4a8e981d572d631e2cede02d8fd082d12fa370788ef1317
+size 1060296656

model-00020-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:208fa65ede58a0447c5a54250869e9fc9d35024fcbbb3f73552f7000f682a964
+size 1214475544

model-00021-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f8b3ddf19b37e6e2630ee6f3180f11ee380e78c538dbaaf7f2f0d5269af200a
+size 1060296656

model-00022-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:439f0ec20316f313ccdcea514d884dfd19c5cca00aa8889a5df9aa534fb4bdae
+size 1214475544

model-00023-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf69dc9a6864d4d97c6c5fd787786ee8587a164b1e34ed8ddcef93f6ea6510bb
+size 1060296656

model-00024-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:feb77cb97a3ec906a7ef268cb265a88ce3d50b0d1535df776d8e1751b6f48144
+size 1214475544

model-00025-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1cc9a8465234884f6e3a05fcb6ea4bd9af3491afecc78c3c0c81ab278ee5edf
+size 1060296656

model-00026-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f311fe675d009ea757c545d294aaa1a9011a5a63a745a18f7b0e46c67a4604d7
+size 1214475544

model-00027-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d46e6c04c07bdb0cf8c3bdf9f2cc928a53122dce74e90e2d28ea435bf4dd14ed
+size 1060296656

model-00028-of-00028.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d56d979f0057efefa421f5f4b5522a1230e14916087ea51a69cb0433357a905
+size 302027176

model.safetensors.index.json

@@ -0,0 +1,491 @@
+{
+  "metadata": {
+    "total_size": 31034925056
+  },
+  "weight_map": {
+    "transformer.h.0.attn.c_attn.bias": "model-00001-of-00028.safetensors",
+    "transformer.h.0.attn.c_attn.weight": "model-00001-of-00028.safetensors",
+    "transformer.h.0.attn.c_proj.bias": "model-00001-of-00028.safetensors",
+    "transformer.h.0.attn.c_proj.weight": "model-00001-of-00028.safetensors",
+    "transformer.h.0.ln_1.bias": "model-00001-of-00028.safetensors",
+    "transformer.h.0.ln_1.weight": "model-00001-of-00028.safetensors",
+    "transformer.h.0.ln_2.bias": "model-00001-of-00028.safetensors",
+    "transformer.h.0.ln_2.weight": "model-00001-of-00028.safetensors",
+    "transformer.h.0.mlp.c_fc.bias": "model-00001-of-00028.safetensors",
+    "transformer.h.0.mlp.c_fc.weight": "model-00001-of-00028.safetensors",
+    "transformer.h.0.mlp.c_proj.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.0.mlp.c_proj.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.1.attn.c_attn.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.1.attn.c_attn.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.1.attn.c_proj.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.1.attn.c_proj.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.1.ln_1.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.1.ln_1.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.1.ln_2.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.1.ln_2.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.1.mlp.c_fc.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.1.mlp.c_fc.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.1.mlp.c_proj.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.1.mlp.c_proj.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.10.attn.c_attn.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.10.attn.c_attn.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.10.attn.c_proj.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.10.attn.c_proj.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.10.ln_1.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.10.ln_1.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.10.ln_2.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.10.ln_2.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.10.mlp.c_fc.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.10.mlp.c_fc.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.10.mlp.c_proj.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.10.mlp.c_proj.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.11.attn.c_attn.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.11.attn.c_attn.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.11.attn.c_proj.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.11.attn.c_proj.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.11.ln_1.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.11.ln_1.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.11.ln_2.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.11.ln_2.weight": "model-00008-of-00028.safetensors",
+    "transformer.h.11.mlp.c_fc.bias": "model-00009-of-00028.safetensors",
+    "transformer.h.11.mlp.c_fc.weight": "model-00009-of-00028.safetensors",
+    "transformer.h.11.mlp.c_proj.bias": "model-00009-of-00028.safetensors",
+    "transformer.h.11.mlp.c_proj.weight": "model-00009-of-00028.safetensors",
+    "transformer.h.12.attn.c_attn.bias": "model-00009-of-00028.safetensors",
+    "transformer.h.12.attn.c_attn.weight": "model-00009-of-00028.safetensors",
+    "transformer.h.12.attn.c_proj.bias": "model-00009-of-00028.safetensors",
+    "transformer.h.12.attn.c_proj.weight": "model-00009-of-00028.safetensors",
+    "transformer.h.12.ln_1.bias": "model-00009-of-00028.safetensors",
+    "transformer.h.12.ln_1.weight": "model-00009-of-00028.safetensors",
+    "transformer.h.12.ln_2.bias": "model-00009-of-00028.safetensors",
+    "transformer.h.12.ln_2.weight": "model-00009-of-00028.safetensors",
+    "transformer.h.12.mlp.c_fc.bias": "model-00009-of-00028.safetensors",
+    "transformer.h.12.mlp.c_fc.weight": "model-00009-of-00028.safetensors",
+    "transformer.h.12.mlp.c_proj.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.12.mlp.c_proj.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.13.attn.c_attn.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.13.attn.c_attn.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.13.attn.c_proj.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.13.attn.c_proj.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.13.ln_1.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.13.ln_1.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.13.ln_2.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.13.ln_2.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.13.mlp.c_fc.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.13.mlp.c_fc.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.13.mlp.c_proj.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.13.mlp.c_proj.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.14.attn.c_attn.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.14.attn.c_attn.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.14.attn.c_proj.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.14.attn.c_proj.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.14.ln_1.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.14.ln_1.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.14.ln_2.bias": "model-00010-of-00028.safetensors",
+    "transformer.h.14.ln_2.weight": "model-00010-of-00028.safetensors",
+    "transformer.h.14.mlp.c_fc.bias": "model-00011-of-00028.safetensors",
+    "transformer.h.14.mlp.c_fc.weight": "model-00011-of-00028.safetensors",
+    "transformer.h.14.mlp.c_proj.bias": "model-00011-of-00028.safetensors",
+    "transformer.h.14.mlp.c_proj.weight": "model-00011-of-00028.safetensors",
+    "transformer.h.15.attn.c_attn.bias": "model-00011-of-00028.safetensors",
+    "transformer.h.15.attn.c_attn.weight": "model-00011-of-00028.safetensors",
+    "transformer.h.15.attn.c_proj.bias": "model-00011-of-00028.safetensors",
+    "transformer.h.15.attn.c_proj.weight": "model-00011-of-00028.safetensors",
+    "transformer.h.15.ln_1.bias": "model-00011-of-00028.safetensors",
+    "transformer.h.15.ln_1.weight": "model-00011-of-00028.safetensors",
+    "transformer.h.15.ln_2.bias": "model-00011-of-00028.safetensors",
+    "transformer.h.15.ln_2.weight": "model-00011-of-00028.safetensors",
+    "transformer.h.15.mlp.c_fc.bias": "model-00011-of-00028.safetensors",
+    "transformer.h.15.mlp.c_fc.weight": "model-00011-of-00028.safetensors",
+    "transformer.h.15.mlp.c_proj.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.15.mlp.c_proj.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.16.attn.c_attn.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.16.attn.c_attn.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.16.attn.c_proj.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.16.attn.c_proj.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.16.ln_1.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.16.ln_1.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.16.ln_2.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.16.ln_2.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.16.mlp.c_fc.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.16.mlp.c_fc.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.16.mlp.c_proj.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.16.mlp.c_proj.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.17.attn.c_attn.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.17.attn.c_attn.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.17.attn.c_proj.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.17.attn.c_proj.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.17.ln_1.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.17.ln_1.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.17.ln_2.bias": "model-00012-of-00028.safetensors",
+    "transformer.h.17.ln_2.weight": "model-00012-of-00028.safetensors",
+    "transformer.h.17.mlp.c_fc.bias": "model-00013-of-00028.safetensors",
+    "transformer.h.17.mlp.c_fc.weight": "model-00013-of-00028.safetensors",
+    "transformer.h.17.mlp.c_proj.bias": "model-00013-of-00028.safetensors",
+    "transformer.h.17.mlp.c_proj.weight": "model-00013-of-00028.safetensors",
+    "transformer.h.18.attn.c_attn.bias": "model-00013-of-00028.safetensors",
+    "transformer.h.18.attn.c_attn.weight": "model-00013-of-00028.safetensors",
+    "transformer.h.18.attn.c_proj.bias": "model-00013-of-00028.safetensors",
+    "transformer.h.18.attn.c_proj.weight": "model-00013-of-00028.safetensors",
+    "transformer.h.18.ln_1.bias": "model-00013-of-00028.safetensors",
+    "transformer.h.18.ln_1.weight": "model-00013-of-00028.safetensors",
+    "transformer.h.18.ln_2.bias": "model-00013-of-00028.safetensors",
+    "transformer.h.18.ln_2.weight": "model-00013-of-00028.safetensors",
+    "transformer.h.18.mlp.c_fc.bias": "model-00013-of-00028.safetensors",
+    "transformer.h.18.mlp.c_fc.weight": "model-00013-of-00028.safetensors",
+    "transformer.h.18.mlp.c_proj.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.18.mlp.c_proj.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.19.attn.c_attn.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.19.attn.c_attn.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.19.attn.c_proj.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.19.attn.c_proj.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.19.ln_1.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.19.ln_1.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.19.ln_2.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.19.ln_2.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.19.mlp.c_fc.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.19.mlp.c_fc.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.19.mlp.c_proj.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.19.mlp.c_proj.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.2.attn.c_attn.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.2.attn.c_attn.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.2.attn.c_proj.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.2.attn.c_proj.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.2.ln_1.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.2.ln_1.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.2.ln_2.bias": "model-00002-of-00028.safetensors",
+    "transformer.h.2.ln_2.weight": "model-00002-of-00028.safetensors",
+    "transformer.h.2.mlp.c_fc.bias": "model-00003-of-00028.safetensors",
+    "transformer.h.2.mlp.c_fc.weight": "model-00003-of-00028.safetensors",
+    "transformer.h.2.mlp.c_proj.bias": "model-00003-of-00028.safetensors",
+    "transformer.h.2.mlp.c_proj.weight": "model-00003-of-00028.safetensors",
+    "transformer.h.20.attn.c_attn.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.20.attn.c_attn.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.20.attn.c_proj.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.20.attn.c_proj.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.20.ln_1.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.20.ln_1.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.20.ln_2.bias": "model-00014-of-00028.safetensors",
+    "transformer.h.20.ln_2.weight": "model-00014-of-00028.safetensors",
+    "transformer.h.20.mlp.c_fc.bias": "model-00015-of-00028.safetensors",
+    "transformer.h.20.mlp.c_fc.weight": "model-00015-of-00028.safetensors",
+    "transformer.h.20.mlp.c_proj.bias": "model-00015-of-00028.safetensors",
+    "transformer.h.20.mlp.c_proj.weight": "model-00015-of-00028.safetensors",
+    "transformer.h.21.attn.c_attn.bias": "model-00015-of-00028.safetensors",
+    "transformer.h.21.attn.c_attn.weight": "model-00015-of-00028.safetensors",
+    "transformer.h.21.attn.c_proj.bias": "model-00015-of-00028.safetensors",
+    "transformer.h.21.attn.c_proj.weight": "model-00015-of-00028.safetensors",
+    "transformer.h.21.ln_1.bias": "model-00015-of-00028.safetensors",
+    "transformer.h.21.ln_1.weight": "model-00015-of-00028.safetensors",
+    "transformer.h.21.ln_2.bias": "model-00015-of-00028.safetensors",
+    "transformer.h.21.ln_2.weight": "model-00015-of-00028.safetensors",
+    "transformer.h.21.mlp.c_fc.bias": "model-00015-of-00028.safetensors",
+    "transformer.h.21.mlp.c_fc.weight": "model-00015-of-00028.safetensors",
+    "transformer.h.21.mlp.c_proj.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.21.mlp.c_proj.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.22.attn.c_attn.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.22.attn.c_attn.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.22.attn.c_proj.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.22.attn.c_proj.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.22.ln_1.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.22.ln_1.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.22.ln_2.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.22.ln_2.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.22.mlp.c_fc.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.22.mlp.c_fc.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.22.mlp.c_proj.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.22.mlp.c_proj.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.23.attn.c_attn.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.23.attn.c_attn.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.23.attn.c_proj.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.23.attn.c_proj.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.23.ln_1.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.23.ln_1.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.23.ln_2.bias": "model-00016-of-00028.safetensors",
+    "transformer.h.23.ln_2.weight": "model-00016-of-00028.safetensors",
+    "transformer.h.23.mlp.c_fc.bias": "model-00017-of-00028.safetensors",
+    "transformer.h.23.mlp.c_fc.weight": "model-00017-of-00028.safetensors",
+    "transformer.h.23.mlp.c_proj.bias": "model-00017-of-00028.safetensors",
+    "transformer.h.23.mlp.c_proj.weight": "model-00017-of-00028.safetensors",
+    "transformer.h.24.attn.c_attn.bias": "model-00017-of-00028.safetensors",
+    "transformer.h.24.attn.c_attn.weight": "model-00017-of-00028.safetensors",
+    "transformer.h.24.attn.c_proj.bias": "model-00017-of-00028.safetensors",
+    "transformer.h.24.attn.c_proj.weight": "model-00017-of-00028.safetensors",
+    "transformer.h.24.ln_1.bias": "model-00017-of-00028.safetensors",
+    "transformer.h.24.ln_1.weight": "model-00017-of-00028.safetensors",
+    "transformer.h.24.ln_2.bias": "model-00017-of-00028.safetensors",
+    "transformer.h.24.ln_2.weight": "model-00017-of-00028.safetensors",
+    "transformer.h.24.mlp.c_fc.bias": "model-00017-of-00028.safetensors",
+    "transformer.h.24.mlp.c_fc.weight": "model-00017-of-00028.safetensors",
+    "transformer.h.24.mlp.c_proj.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.24.mlp.c_proj.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.25.attn.c_attn.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.25.attn.c_attn.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.25.attn.c_proj.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.25.attn.c_proj.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.25.ln_1.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.25.ln_1.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.25.ln_2.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.25.ln_2.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.25.mlp.c_fc.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.25.mlp.c_fc.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.25.mlp.c_proj.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.25.mlp.c_proj.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.26.attn.c_attn.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.26.attn.c_attn.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.26.attn.c_proj.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.26.attn.c_proj.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.26.ln_1.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.26.ln_1.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.26.ln_2.bias": "model-00018-of-00028.safetensors",
+    "transformer.h.26.ln_2.weight": "model-00018-of-00028.safetensors",
+    "transformer.h.26.mlp.c_fc.bias": "model-00019-of-00028.safetensors",
+    "transformer.h.26.mlp.c_fc.weight": "model-00019-of-00028.safetensors",
+    "transformer.h.26.mlp.c_proj.bias": "model-00019-of-00028.safetensors",
+    "transformer.h.26.mlp.c_proj.weight": "model-00019-of-00028.safetensors",
+    "transformer.h.27.attn.c_attn.bias": "model-00019-of-00028.safetensors",
+    "transformer.h.27.attn.c_attn.weight": "model-00019-of-00028.safetensors",
+    "transformer.h.27.attn.c_proj.bias": "model-00019-of-00028.safetensors",
+    "transformer.h.27.attn.c_proj.weight": "model-00019-of-00028.safetensors",
+    "transformer.h.27.ln_1.bias": "model-00019-of-00028.safetensors",
+    "transformer.h.27.ln_1.weight": "model-00019-of-00028.safetensors",
+    "transformer.h.27.ln_2.bias": "model-00019-of-00028.safetensors",
+    "transformer.h.27.ln_2.weight": "model-00019-of-00028.safetensors",
+    "transformer.h.27.mlp.c_fc.bias": "model-00019-of-00028.safetensors",
+    "transformer.h.27.mlp.c_fc.weight": "model-00019-of-00028.safetensors",
+    "transformer.h.27.mlp.c_proj.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.27.mlp.c_proj.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.28.attn.c_attn.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.28.attn.c_attn.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.28.attn.c_proj.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.28.attn.c_proj.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.28.ln_1.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.28.ln_1.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.28.ln_2.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.28.ln_2.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.28.mlp.c_fc.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.28.mlp.c_fc.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.28.mlp.c_proj.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.28.mlp.c_proj.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.29.attn.c_attn.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.29.attn.c_attn.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.29.attn.c_proj.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.29.attn.c_proj.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.29.ln_1.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.29.ln_1.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.29.ln_2.bias": "model-00020-of-00028.safetensors",
+    "transformer.h.29.ln_2.weight": "model-00020-of-00028.safetensors",
+    "transformer.h.29.mlp.c_fc.bias": "model-00021-of-00028.safetensors",
+    "transformer.h.29.mlp.c_fc.weight": "model-00021-of-00028.safetensors",
+    "transformer.h.29.mlp.c_proj.bias": "model-00021-of-00028.safetensors",
+    "transformer.h.29.mlp.c_proj.weight": "model-00021-of-00028.safetensors",
+    "transformer.h.3.attn.c_attn.bias": "model-00003-of-00028.safetensors",
+    "transformer.h.3.attn.c_attn.weight": "model-00003-of-00028.safetensors",
+    "transformer.h.3.attn.c_proj.bias": "model-00003-of-00028.safetensors",
+    "transformer.h.3.attn.c_proj.weight": "model-00003-of-00028.safetensors",
+    "transformer.h.3.ln_1.bias": "model-00003-of-00028.safetensors",
+    "transformer.h.3.ln_1.weight": "model-00003-of-00028.safetensors",
+    "transformer.h.3.ln_2.bias": "model-00003-of-00028.safetensors",
+    "transformer.h.3.ln_2.weight": "model-00003-of-00028.safetensors",
+    "transformer.h.3.mlp.c_fc.bias": "model-00003-of-00028.safetensors",
+    "transformer.h.3.mlp.c_fc.weight": "model-00003-of-00028.safetensors",
+    "transformer.h.3.mlp.c_proj.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.3.mlp.c_proj.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.30.attn.c_attn.bias": "model-00021-of-00028.safetensors",
+    "transformer.h.30.attn.c_attn.weight": "model-00021-of-00028.safetensors",
+    "transformer.h.30.attn.c_proj.bias": "model-00021-of-00028.safetensors",
+    "transformer.h.30.attn.c_proj.weight": "model-00021-of-00028.safetensors",
+    "transformer.h.30.ln_1.bias": "model-00021-of-00028.safetensors",
+    "transformer.h.30.ln_1.weight": "model-00021-of-00028.safetensors",
+    "transformer.h.30.ln_2.bias": "model-00021-of-00028.safetensors",
+    "transformer.h.30.ln_2.weight": "model-00021-of-00028.safetensors",
+    "transformer.h.30.mlp.c_fc.bias": "model-00021-of-00028.safetensors",
+    "transformer.h.30.mlp.c_fc.weight": "model-00021-of-00028.safetensors",
+    "transformer.h.30.mlp.c_proj.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.30.mlp.c_proj.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.31.attn.c_attn.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.31.attn.c_attn.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.31.attn.c_proj.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.31.attn.c_proj.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.31.ln_1.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.31.ln_1.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.31.ln_2.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.31.ln_2.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.31.mlp.c_fc.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.31.mlp.c_fc.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.31.mlp.c_proj.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.31.mlp.c_proj.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.32.attn.c_attn.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.32.attn.c_attn.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.32.attn.c_proj.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.32.attn.c_proj.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.32.ln_1.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.32.ln_1.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.32.ln_2.bias": "model-00022-of-00028.safetensors",
+    "transformer.h.32.ln_2.weight": "model-00022-of-00028.safetensors",
+    "transformer.h.32.mlp.c_fc.bias": "model-00023-of-00028.safetensors",
+    "transformer.h.32.mlp.c_fc.weight": "model-00023-of-00028.safetensors",
+    "transformer.h.32.mlp.c_proj.bias": "model-00023-of-00028.safetensors",
+    "transformer.h.32.mlp.c_proj.weight": "model-00023-of-00028.safetensors",
+    "transformer.h.33.attn.c_attn.bias": "model-00023-of-00028.safetensors",
+    "transformer.h.33.attn.c_attn.weight": "model-00023-of-00028.safetensors",
+    "transformer.h.33.attn.c_proj.bias": "model-00023-of-00028.safetensors",
+    "transformer.h.33.attn.c_proj.weight": "model-00023-of-00028.safetensors",
+    "transformer.h.33.ln_1.bias": "model-00023-of-00028.safetensors",
+    "transformer.h.33.ln_1.weight": "model-00023-of-00028.safetensors",
+    "transformer.h.33.ln_2.bias": "model-00023-of-00028.safetensors",
+    "transformer.h.33.ln_2.weight": "model-00023-of-00028.safetensors",
+    "transformer.h.33.mlp.c_fc.bias": "model-00023-of-00028.safetensors",
+    "transformer.h.33.mlp.c_fc.weight": "model-00023-of-00028.safetensors",
+    "transformer.h.33.mlp.c_proj.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.33.mlp.c_proj.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.34.attn.c_attn.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.34.attn.c_attn.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.34.attn.c_proj.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.34.attn.c_proj.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.34.ln_1.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.34.ln_1.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.34.ln_2.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.34.ln_2.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.34.mlp.c_fc.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.34.mlp.c_fc.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.34.mlp.c_proj.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.34.mlp.c_proj.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.35.attn.c_attn.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.35.attn.c_attn.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.35.attn.c_proj.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.35.attn.c_proj.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.35.ln_1.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.35.ln_1.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.35.ln_2.bias": "model-00024-of-00028.safetensors",
+    "transformer.h.35.ln_2.weight": "model-00024-of-00028.safetensors",
+    "transformer.h.35.mlp.c_fc.bias": "model-00025-of-00028.safetensors",
+    "transformer.h.35.mlp.c_fc.weight": "model-00025-of-00028.safetensors",
+    "transformer.h.35.mlp.c_proj.bias": "model-00025-of-00028.safetensors",
+    "transformer.h.35.mlp.c_proj.weight": "model-00025-of-00028.safetensors",
+    "transformer.h.36.attn.c_attn.bias": "model-00025-of-00028.safetensors",
+    "transformer.h.36.attn.c_attn.weight": "model-00025-of-00028.safetensors",
+    "transformer.h.36.attn.c_proj.bias": "model-00025-of-00028.safetensors",
+    "transformer.h.36.attn.c_proj.weight": "model-00025-of-00028.safetensors",
+    "transformer.h.36.ln_1.bias": "model-00025-of-00028.safetensors",
+    "transformer.h.36.ln_1.weight": "model-00025-of-00028.safetensors",
+    "transformer.h.36.ln_2.bias": "model-00025-of-00028.safetensors",
+    "transformer.h.36.ln_2.weight": "model-00025-of-00028.safetensors",
+    "transformer.h.36.mlp.c_fc.bias": "model-00025-of-00028.safetensors",
+    "transformer.h.36.mlp.c_fc.weight": "model-00025-of-00028.safetensors",
+    "transformer.h.36.mlp.c_proj.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.36.mlp.c_proj.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.37.attn.c_attn.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.37.attn.c_attn.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.37.attn.c_proj.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.37.attn.c_proj.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.37.ln_1.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.37.ln_1.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.37.ln_2.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.37.ln_2.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.37.mlp.c_fc.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.37.mlp.c_fc.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.37.mlp.c_proj.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.37.mlp.c_proj.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.38.attn.c_attn.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.38.attn.c_attn.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.38.attn.c_proj.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.38.attn.c_proj.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.38.ln_1.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.38.ln_1.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.38.ln_2.bias": "model-00026-of-00028.safetensors",
+    "transformer.h.38.ln_2.weight": "model-00026-of-00028.safetensors",
+    "transformer.h.38.mlp.c_fc.bias": "model-00027-of-00028.safetensors",
+    "transformer.h.38.mlp.c_fc.weight": "model-00027-of-00028.safetensors",
+    "transformer.h.38.mlp.c_proj.bias": "model-00027-of-00028.safetensors",
+    "transformer.h.38.mlp.c_proj.weight": "model-00027-of-00028.safetensors",
+    "transformer.h.39.attn.c_attn.bias": "model-00027-of-00028.safetensors",
+    "transformer.h.39.attn.c_attn.weight": "model-00027-of-00028.safetensors",
+    "transformer.h.39.attn.c_proj.bias": "model-00027-of-00028.safetensors",
+    "transformer.h.39.attn.c_proj.weight": "model-00027-of-00028.safetensors",
+    "transformer.h.39.ln_1.bias": "model-00027-of-00028.safetensors",
+    "transformer.h.39.ln_1.weight": "model-00027-of-00028.safetensors",
+    "transformer.h.39.ln_2.bias": "model-00027-of-00028.safetensors",
+    "transformer.h.39.ln_2.weight": "model-00027-of-00028.safetensors",
+    "transformer.h.39.mlp.c_fc.bias": "model-00027-of-00028.safetensors",
+    "transformer.h.39.mlp.c_fc.weight": "model-00027-of-00028.safetensors",
+    "transformer.h.39.mlp.c_proj.bias": "model-00028-of-00028.safetensors",
+    "transformer.h.39.mlp.c_proj.weight": "model-00028-of-00028.safetensors",
+    "transformer.h.4.attn.c_attn.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.4.attn.c_attn.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.4.attn.c_proj.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.4.attn.c_proj.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.4.ln_1.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.4.ln_1.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.4.ln_2.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.4.ln_2.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.4.mlp.c_fc.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.4.mlp.c_fc.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.4.mlp.c_proj.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.4.mlp.c_proj.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.5.attn.c_attn.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.5.attn.c_attn.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.5.attn.c_proj.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.5.attn.c_proj.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.5.ln_1.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.5.ln_1.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.5.ln_2.bias": "model-00004-of-00028.safetensors",
+    "transformer.h.5.ln_2.weight": "model-00004-of-00028.safetensors",
+    "transformer.h.5.mlp.c_fc.bias": "model-00005-of-00028.safetensors",
+    "transformer.h.5.mlp.c_fc.weight": "model-00005-of-00028.safetensors",
+    "transformer.h.5.mlp.c_proj.bias": "model-00005-of-00028.safetensors",
+    "transformer.h.5.mlp.c_proj.weight": "model-00005-of-00028.safetensors",
+    "transformer.h.6.attn.c_attn.bias": "model-00005-of-00028.safetensors",
+    "transformer.h.6.attn.c_attn.weight": "model-00005-of-00028.safetensors",
+    "transformer.h.6.attn.c_proj.bias": "model-00005-of-00028.safetensors",
+    "transformer.h.6.attn.c_proj.weight": "model-00005-of-00028.safetensors",
+    "transformer.h.6.ln_1.bias": "model-00005-of-00028.safetensors",
+    "transformer.h.6.ln_1.weight": "model-00005-of-00028.safetensors",
+    "transformer.h.6.ln_2.bias": "model-00005-of-00028.safetensors",
+    "transformer.h.6.ln_2.weight": "model-00005-of-00028.safetensors",
+    "transformer.h.6.mlp.c_fc.bias": "model-00005-of-00028.safetensors",
+    "transformer.h.6.mlp.c_fc.weight": "model-00005-of-00028.safetensors",
+    "transformer.h.6.mlp.c_proj.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.6.mlp.c_proj.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.7.attn.c_attn.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.7.attn.c_attn.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.7.attn.c_proj.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.7.attn.c_proj.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.7.ln_1.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.7.ln_1.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.7.ln_2.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.7.ln_2.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.7.mlp.c_fc.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.7.mlp.c_fc.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.7.mlp.c_proj.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.7.mlp.c_proj.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.8.attn.c_attn.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.8.attn.c_attn.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.8.attn.c_proj.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.8.attn.c_proj.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.8.ln_1.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.8.ln_1.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.8.ln_2.bias": "model-00006-of-00028.safetensors",
+    "transformer.h.8.ln_2.weight": "model-00006-of-00028.safetensors",
+    "transformer.h.8.mlp.c_fc.bias": "model-00007-of-00028.safetensors",
+    "transformer.h.8.mlp.c_fc.weight": "model-00007-of-00028.safetensors",
+    "transformer.h.8.mlp.c_proj.bias": "model-00007-of-00028.safetensors",
+    "transformer.h.8.mlp.c_proj.weight": "model-00007-of-00028.safetensors",
+    "transformer.h.9.attn.c_attn.bias": "model-00007-of-00028.safetensors",
+    "transformer.h.9.attn.c_attn.weight": "model-00007-of-00028.safetensors",
+    "transformer.h.9.attn.c_proj.bias": "model-00007-of-00028.safetensors",
+    "transformer.h.9.attn.c_proj.weight": "model-00007-of-00028.safetensors",
+    "transformer.h.9.ln_1.bias": "model-00007-of-00028.safetensors",
+    "transformer.h.9.ln_1.weight": "model-00007-of-00028.safetensors",
+    "transformer.h.9.ln_2.bias": "model-00007-of-00028.safetensors",
+    "transformer.h.9.ln_2.weight": "model-00007-of-00028.safetensors",
+    "transformer.h.9.mlp.c_fc.bias": "model-00007-of-00028.safetensors",
+    "transformer.h.9.mlp.c_fc.weight": "model-00007-of-00028.safetensors",
+    "transformer.h.9.mlp.c_proj.bias": "model-00008-of-00028.safetensors",
+    "transformer.h.9.mlp.c_proj.weight": "model-00008-of-00028.safetensors",
+    "transformer.ln_f.bias": "model-00028-of-00028.safetensors",
+    "transformer.ln_f.weight": "model-00028-of-00028.safetensors",
+    "transformer.wpe.weight": "model-00001-of-00028.safetensors",
+    "transformer.wte.weight": "model-00001-of-00028.safetensors"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,27 @@
+{
+  "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|>",
+  "eos_token": "<|endoftext|>",
+  "pad_token": "[PAD]",
+  "unk_token": "<|endoftext|>"
+}

tokenizer_config.json

@@ -0,0 +1,32 @@
+{
+  "add_prefix_space": false,
+  "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": 2048,
+  "padding_side": "right",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>",
+  "vocab_size": 49152
+}