Upload FalconForCausalLM

config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "falcon-180B-chat-instruct",
+  "alibi": false,
+  "architectures": [
+    "FalconForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bias": false,
+  "bos_token_id": 11,
+  "eos_token_id": 11,
+  "hidden_dropout": 0.0,
+  "hidden_size": 14848,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "model_type": "falcon",
+  "multi_query": true,
+  "new_decoder_architecture": true,
+  "num_attention_heads": 232,
+  "num_hidden_layers": 80,
+  "num_kv_heads": 8,
+  "parallel_attn": true,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.31.0",
+  "use_cache": true,
+  "vocab_size": 65024
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 11,
+  "eos_token_id": 11,
+  "transformers_version": "4.31.0"
+}

model-00001-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a0fa13734f56a302f4f65a9d477a3aea53fb845c2c6cfa1aff01d2471ff294f
+size 9046709632

model-00002-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56bff5ac95c81a6fffe084b8707ab24e34600e33e735552c7f5aaebf9f05875d
+size 8879580960

model-00003-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7669b21d905352ba95b4b5f7385335fae702f89af245f80f5b8fc262a26a2b9b
+size 8879580960

model-00004-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15a2e92a8d99ddf4b1931ecc40185f278a2c9b6383b134c08d8e19cf24491f38
+size 8879580960

model-00005-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ac9634192823166ecf134fff339f3692271378f689bb6f1036b8b5b978325f9e
+size 8879580960

model-00006-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba6200782bfaa0572f1887d7eba8b5e7fd5164073aaed06e53139922b29d45af
+size 8879580968

model-00007-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc471bdda4c708d5615d1b2d31b2163b0231e1a24915235472a3bf809e4fc12c
+size 8879580976

model-00008-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89cf8887b0bb33dd64e70a7b6a5fa7db8087688a9c406ef9570841c75f3c6dd1
+size 8879580976

model-00009-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05be12040d386f5cd62a2747cc7f7e573b5c0dbb3d02ee66efda66ac1513a3fd
+size 8879580976

model-00010-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:639e5096e16f4a4911cf51ea9a039f753c4c8c644ce042a7bb85a8425a5a4e22
+size 8879580976

model-00011-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3a2fa4a366afaaaaeb6e7941950e713b6c5b37abb79afe3311787f97e73b4b1
+size 8879580976

model-00012-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf355a21f2f44b1abd00f02e7cb0219c6767221f2b6af3c1518cd4149477c041
+size 8879580976

model-00013-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8449f93d84d1348401c25c794dc86f5a7a15db432626fe1e22afdbd3f8034fcd
+size 8879580976

model-00014-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:382f0ce96f243cd96dfb4d44cbb5aeb93d6cff68ffb3bd8eb0b940e92f815220
+size 8879580976

model-00015-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ae86f8679c0e9b5bec3a90b50f3360b2dcc24eefb24f96b114d287ba11b2fd7
+size 8879580976

model-00016-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ddfa22dc765ec7e88e34915d64209c88883109c589390ee3155d3aa780176bf
+size 8879580976

model-00017-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:79c0a85378fe31ea049552acfdecdfba9592feab89898a38239c0515f4e52414
+size 8879580976

model-00018-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc9ec74e0ce499c479d97ad83c33e9dd0f9b277693cdfc2902bb76670bc493dc
+size 8879580976

model-00019-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c6987f343119d5ff218d303cdf18f2f7c393babd74f6b2e79eb79a85e125907
+size 8879580976

model-00020-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:712995c95b2900efd9e2de1714e259c961f7ad076909bb1352ae1aa63f39d4b3
+size 8879580976

model-00021-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65792f2069b83062228686d68598a3d7a1d27531879db7267882f6dcc271904c
+size 8879580976

model-00022-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a89fab137551b6d1660b685767ad0065a3917138093a4077dd3d6e8087608b4
+size 8879580976

model-00023-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b87cf46acfcd49ba799b7dfd248c5bbf76453932773b315decfa1c26983563ca
+size 8879580976

model-00024-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06b66390bb89c45b15305c1b3feeb626cdc7a8c398932b0efae962c0e76977df
+size 8879580976

model-00025-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:964a71f9482c9a819a3aa051151590145f4ea770c7bd77a30b141dcaa9e42faa
+size 8879580976

model-00026-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f361547f53b1e0906531bae86ed95508da90077e6a293bbfec4cd868eae0c4d
+size 8879580976

model-00027-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a9cdc4af96557d3dd996cd5acd699d1b3a7a8777d2028528e798e8e3d95932e
+size 8879580976

model-00028-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f544618606e7f212a54bf83c10bac8ac9de2c26d9fd7fb2e5ec7d0247cdfd99e
+size 8879580976

model-00029-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84fb596166224dc19ef7d8e379b7d9d4054bfc0dcb2c79617d227fab2ed32b25
+size 8879580976

model-00030-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ebbe4864608f7e0605a90aa11aea5d11301ed3dac74aa99cf87d1016bd503c79
+size 8879580976

model-00031-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21beeea55f1235e7c92c816e1bdb62a1bf0663815aa0bafd9fee80402b1b94ac
+size 8879580976

model-00032-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b0c1a7db2b25809e1339defef06319a26d03af8bed24951b5698cb648783e834
+size 8879580976

model-00033-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f810799eee280e8e3f60652d3dc05b69e0a948672463e9e63c33e690b95fdb1
+size 8879580976

model-00034-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82ec3450620b199036209aeea7fc489dcc8c7c7f19fbb6f5154e8a191a950e31
+size 8879580976

model-00035-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c329b5ae35e85333893ecd3f00ff75fbe86acf13a8a0d371de20356b50b94cb
+size 8879580976

model-00036-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f28a5a3b250b3e16ca944467f3795b81cbec82e3275c875cb5a7a86fa63c7e2
+size 8879580976

model-00037-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2239c11ef9b59677eeecaab9c9f3f996081195e4226e973f0818f9914878fe66
+size 8879580976

model-00038-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cecf1eb322a8107b0b229f5b4b8275223bb42c51af7c8b2e4796b2cb9239244d
+size 8879580976

model-00039-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da04a98674b82c9362d1f5991efa8b3388c6ae62d42268252e1bd07185d075d7
+size 8879580976

model-00040-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe96cc3b14b1d0be882a0180e466e18e85c48916cdf6edf26eadec16a810afb8
+size 8879580976

model-00041-of-00041.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39a7b29896999887da12780ab62087d6a3f17ffedc0c08ed81bfa12581d36ad5
+size 1763883736

model.safetensors.index.json

@@ -0,0 +1,650 @@
+{
+  "metadata": {
+    "total_size": 357114177536
+  },
+  "weight_map": {
+    "transformer.h.0.ln_attn.bias": "model-00001-of-00041.safetensors",
+    "transformer.h.0.ln_attn.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.0.ln_mlp.bias": "model-00001-of-00041.safetensors",
+    "transformer.h.0.ln_mlp.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.0.mlp.dense_4h_to_h.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.0.mlp.dense_h_to_4h.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.0.self_attention.dense.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.0.self_attention.query_key_value.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.1.ln_attn.bias": "model-00002-of-00041.safetensors",
+    "transformer.h.1.ln_attn.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.1.ln_mlp.bias": "model-00002-of-00041.safetensors",
+    "transformer.h.1.ln_mlp.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.1.mlp.dense_4h_to_h.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.1.mlp.dense_h_to_4h.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.1.self_attention.dense.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.1.self_attention.query_key_value.weight": "model-00001-of-00041.safetensors",
+    "transformer.h.10.ln_attn.bias": "model-00006-of-00041.safetensors",
+    "transformer.h.10.ln_attn.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.10.ln_mlp.bias": "model-00006-of-00041.safetensors",
+    "transformer.h.10.ln_mlp.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.10.mlp.dense_4h_to_h.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.10.mlp.dense_h_to_4h.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.10.self_attention.dense.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.10.self_attention.query_key_value.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.11.ln_attn.bias": "model-00007-of-00041.safetensors",
+    "transformer.h.11.ln_attn.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.11.ln_mlp.bias": "model-00007-of-00041.safetensors",
+    "transformer.h.11.ln_mlp.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.11.mlp.dense_4h_to_h.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.11.mlp.dense_h_to_4h.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.11.self_attention.dense.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.11.self_attention.query_key_value.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.12.ln_attn.bias": "model-00007-of-00041.safetensors",
+    "transformer.h.12.ln_attn.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.12.ln_mlp.bias": "model-00007-of-00041.safetensors",
+    "transformer.h.12.ln_mlp.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.12.mlp.dense_4h_to_h.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.12.mlp.dense_h_to_4h.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.12.self_attention.dense.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.12.self_attention.query_key_value.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.13.ln_attn.bias": "model-00008-of-00041.safetensors",
+    "transformer.h.13.ln_attn.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.13.ln_mlp.bias": "model-00008-of-00041.safetensors",
+    "transformer.h.13.ln_mlp.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.13.mlp.dense_4h_to_h.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.13.mlp.dense_h_to_4h.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.13.self_attention.dense.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.13.self_attention.query_key_value.weight": "model-00007-of-00041.safetensors",
+    "transformer.h.14.ln_attn.bias": "model-00008-of-00041.safetensors",
+    "transformer.h.14.ln_attn.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.14.ln_mlp.bias": "model-00008-of-00041.safetensors",
+    "transformer.h.14.ln_mlp.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.14.mlp.dense_4h_to_h.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.14.mlp.dense_h_to_4h.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.14.self_attention.dense.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.14.self_attention.query_key_value.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.15.ln_attn.bias": "model-00009-of-00041.safetensors",
+    "transformer.h.15.ln_attn.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.15.ln_mlp.bias": "model-00009-of-00041.safetensors",
+    "transformer.h.15.ln_mlp.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.15.mlp.dense_4h_to_h.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.15.mlp.dense_h_to_4h.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.15.self_attention.dense.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.15.self_attention.query_key_value.weight": "model-00008-of-00041.safetensors",
+    "transformer.h.16.ln_attn.bias": "model-00009-of-00041.safetensors",
+    "transformer.h.16.ln_attn.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.16.ln_mlp.bias": "model-00009-of-00041.safetensors",
+    "transformer.h.16.ln_mlp.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.16.mlp.dense_4h_to_h.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.16.mlp.dense_h_to_4h.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.16.self_attention.dense.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.16.self_attention.query_key_value.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.17.ln_attn.bias": "model-00010-of-00041.safetensors",
+    "transformer.h.17.ln_attn.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.17.ln_mlp.bias": "model-00010-of-00041.safetensors",
+    "transformer.h.17.ln_mlp.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.17.mlp.dense_4h_to_h.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.17.mlp.dense_h_to_4h.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.17.self_attention.dense.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.17.self_attention.query_key_value.weight": "model-00009-of-00041.safetensors",
+    "transformer.h.18.ln_attn.bias": "model-00010-of-00041.safetensors",
+    "transformer.h.18.ln_attn.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.18.ln_mlp.bias": "model-00010-of-00041.safetensors",
+    "transformer.h.18.ln_mlp.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.18.mlp.dense_4h_to_h.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.18.mlp.dense_h_to_4h.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.18.self_attention.dense.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.18.self_attention.query_key_value.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.19.ln_attn.bias": "model-00011-of-00041.safetensors",
+    "transformer.h.19.ln_attn.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.19.ln_mlp.bias": "model-00011-of-00041.safetensors",
+    "transformer.h.19.ln_mlp.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.19.mlp.dense_4h_to_h.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.19.mlp.dense_h_to_4h.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.19.self_attention.dense.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.19.self_attention.query_key_value.weight": "model-00010-of-00041.safetensors",
+    "transformer.h.2.ln_attn.bias": "model-00002-of-00041.safetensors",
+    "transformer.h.2.ln_attn.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.2.ln_mlp.bias": "model-00002-of-00041.safetensors",
+    "transformer.h.2.ln_mlp.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.2.mlp.dense_4h_to_h.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.2.mlp.dense_h_to_4h.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.2.self_attention.dense.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.2.self_attention.query_key_value.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.20.ln_attn.bias": "model-00011-of-00041.safetensors",
+    "transformer.h.20.ln_attn.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.20.ln_mlp.bias": "model-00011-of-00041.safetensors",
+    "transformer.h.20.ln_mlp.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.20.mlp.dense_4h_to_h.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.20.mlp.dense_h_to_4h.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.20.self_attention.dense.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.20.self_attention.query_key_value.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.21.ln_attn.bias": "model-00012-of-00041.safetensors",
+    "transformer.h.21.ln_attn.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.21.ln_mlp.bias": "model-00012-of-00041.safetensors",
+    "transformer.h.21.ln_mlp.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.21.mlp.dense_4h_to_h.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.21.mlp.dense_h_to_4h.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.21.self_attention.dense.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.21.self_attention.query_key_value.weight": "model-00011-of-00041.safetensors",
+    "transformer.h.22.ln_attn.bias": "model-00012-of-00041.safetensors",
+    "transformer.h.22.ln_attn.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.22.ln_mlp.bias": "model-00012-of-00041.safetensors",
+    "transformer.h.22.ln_mlp.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.22.mlp.dense_4h_to_h.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.22.mlp.dense_h_to_4h.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.22.self_attention.dense.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.22.self_attention.query_key_value.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.23.ln_attn.bias": "model-00013-of-00041.safetensors",
+    "transformer.h.23.ln_attn.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.23.ln_mlp.bias": "model-00013-of-00041.safetensors",
+    "transformer.h.23.ln_mlp.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.23.mlp.dense_4h_to_h.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.23.mlp.dense_h_to_4h.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.23.self_attention.dense.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.23.self_attention.query_key_value.weight": "model-00012-of-00041.safetensors",
+    "transformer.h.24.ln_attn.bias": "model-00013-of-00041.safetensors",
+    "transformer.h.24.ln_attn.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.24.ln_mlp.bias": "model-00013-of-00041.safetensors",
+    "transformer.h.24.ln_mlp.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.24.mlp.dense_4h_to_h.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.24.mlp.dense_h_to_4h.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.24.self_attention.dense.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.24.self_attention.query_key_value.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.25.ln_attn.bias": "model-00014-of-00041.safetensors",
+    "transformer.h.25.ln_attn.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.25.ln_mlp.bias": "model-00014-of-00041.safetensors",
+    "transformer.h.25.ln_mlp.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.25.mlp.dense_4h_to_h.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.25.mlp.dense_h_to_4h.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.25.self_attention.dense.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.25.self_attention.query_key_value.weight": "model-00013-of-00041.safetensors",
+    "transformer.h.26.ln_attn.bias": "model-00014-of-00041.safetensors",
+    "transformer.h.26.ln_attn.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.26.ln_mlp.bias": "model-00014-of-00041.safetensors",
+    "transformer.h.26.ln_mlp.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.26.mlp.dense_4h_to_h.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.26.mlp.dense_h_to_4h.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.26.self_attention.dense.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.26.self_attention.query_key_value.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.27.ln_attn.bias": "model-00015-of-00041.safetensors",
+    "transformer.h.27.ln_attn.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.27.ln_mlp.bias": "model-00015-of-00041.safetensors",
+    "transformer.h.27.ln_mlp.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.27.mlp.dense_4h_to_h.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.27.mlp.dense_h_to_4h.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.27.self_attention.dense.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.27.self_attention.query_key_value.weight": "model-00014-of-00041.safetensors",
+    "transformer.h.28.ln_attn.bias": "model-00015-of-00041.safetensors",
+    "transformer.h.28.ln_attn.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.28.ln_mlp.bias": "model-00015-of-00041.safetensors",
+    "transformer.h.28.ln_mlp.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.28.mlp.dense_4h_to_h.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.28.mlp.dense_h_to_4h.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.28.self_attention.dense.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.28.self_attention.query_key_value.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.29.ln_attn.bias": "model-00016-of-00041.safetensors",
+    "transformer.h.29.ln_attn.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.29.ln_mlp.bias": "model-00016-of-00041.safetensors",
+    "transformer.h.29.ln_mlp.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.29.mlp.dense_4h_to_h.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.29.mlp.dense_h_to_4h.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.29.self_attention.dense.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.29.self_attention.query_key_value.weight": "model-00015-of-00041.safetensors",
+    "transformer.h.3.ln_attn.bias": "model-00003-of-00041.safetensors",
+    "transformer.h.3.ln_attn.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.3.ln_mlp.bias": "model-00003-of-00041.safetensors",
+    "transformer.h.3.ln_mlp.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.3.mlp.dense_4h_to_h.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.3.mlp.dense_h_to_4h.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.3.self_attention.dense.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.3.self_attention.query_key_value.weight": "model-00002-of-00041.safetensors",
+    "transformer.h.30.ln_attn.bias": "model-00016-of-00041.safetensors",
+    "transformer.h.30.ln_attn.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.30.ln_mlp.bias": "model-00016-of-00041.safetensors",
+    "transformer.h.30.ln_mlp.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.30.mlp.dense_4h_to_h.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.30.mlp.dense_h_to_4h.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.30.self_attention.dense.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.30.self_attention.query_key_value.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.31.ln_attn.bias": "model-00017-of-00041.safetensors",
+    "transformer.h.31.ln_attn.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.31.ln_mlp.bias": "model-00017-of-00041.safetensors",
+    "transformer.h.31.ln_mlp.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.31.mlp.dense_4h_to_h.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.31.mlp.dense_h_to_4h.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.31.self_attention.dense.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.31.self_attention.query_key_value.weight": "model-00016-of-00041.safetensors",
+    "transformer.h.32.ln_attn.bias": "model-00017-of-00041.safetensors",
+    "transformer.h.32.ln_attn.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.32.ln_mlp.bias": "model-00017-of-00041.safetensors",
+    "transformer.h.32.ln_mlp.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.32.mlp.dense_4h_to_h.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.32.mlp.dense_h_to_4h.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.32.self_attention.dense.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.32.self_attention.query_key_value.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.33.ln_attn.bias": "model-00018-of-00041.safetensors",
+    "transformer.h.33.ln_attn.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.33.ln_mlp.bias": "model-00018-of-00041.safetensors",
+    "transformer.h.33.ln_mlp.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.33.mlp.dense_4h_to_h.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.33.mlp.dense_h_to_4h.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.33.self_attention.dense.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.33.self_attention.query_key_value.weight": "model-00017-of-00041.safetensors",
+    "transformer.h.34.ln_attn.bias": "model-00018-of-00041.safetensors",
+    "transformer.h.34.ln_attn.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.34.ln_mlp.bias": "model-00018-of-00041.safetensors",
+    "transformer.h.34.ln_mlp.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.34.mlp.dense_4h_to_h.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.34.mlp.dense_h_to_4h.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.34.self_attention.dense.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.34.self_attention.query_key_value.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.35.ln_attn.bias": "model-00019-of-00041.safetensors",
+    "transformer.h.35.ln_attn.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.35.ln_mlp.bias": "model-00019-of-00041.safetensors",
+    "transformer.h.35.ln_mlp.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.35.mlp.dense_4h_to_h.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.35.mlp.dense_h_to_4h.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.35.self_attention.dense.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.35.self_attention.query_key_value.weight": "model-00018-of-00041.safetensors",
+    "transformer.h.36.ln_attn.bias": "model-00019-of-00041.safetensors",
+    "transformer.h.36.ln_attn.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.36.ln_mlp.bias": "model-00019-of-00041.safetensors",
+    "transformer.h.36.ln_mlp.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.36.mlp.dense_4h_to_h.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.36.mlp.dense_h_to_4h.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.36.self_attention.dense.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.36.self_attention.query_key_value.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.37.ln_attn.bias": "model-00020-of-00041.safetensors",
+    "transformer.h.37.ln_attn.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.37.ln_mlp.bias": "model-00020-of-00041.safetensors",
+    "transformer.h.37.ln_mlp.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.37.mlp.dense_4h_to_h.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.37.mlp.dense_h_to_4h.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.37.self_attention.dense.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.37.self_attention.query_key_value.weight": "model-00019-of-00041.safetensors",
+    "transformer.h.38.ln_attn.bias": "model-00020-of-00041.safetensors",
+    "transformer.h.38.ln_attn.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.38.ln_mlp.bias": "model-00020-of-00041.safetensors",
+    "transformer.h.38.ln_mlp.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.38.mlp.dense_4h_to_h.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.38.mlp.dense_h_to_4h.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.38.self_attention.dense.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.38.self_attention.query_key_value.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.39.ln_attn.bias": "model-00021-of-00041.safetensors",
+    "transformer.h.39.ln_attn.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.39.ln_mlp.bias": "model-00021-of-00041.safetensors",
+    "transformer.h.39.ln_mlp.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.39.mlp.dense_4h_to_h.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.39.mlp.dense_h_to_4h.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.39.self_attention.dense.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.39.self_attention.query_key_value.weight": "model-00020-of-00041.safetensors",
+    "transformer.h.4.ln_attn.bias": "model-00003-of-00041.safetensors",
+    "transformer.h.4.ln_attn.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.4.ln_mlp.bias": "model-00003-of-00041.safetensors",
+    "transformer.h.4.ln_mlp.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.4.mlp.dense_4h_to_h.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.4.mlp.dense_h_to_4h.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.4.self_attention.dense.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.4.self_attention.query_key_value.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.40.ln_attn.bias": "model-00021-of-00041.safetensors",
+    "transformer.h.40.ln_attn.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.40.ln_mlp.bias": "model-00021-of-00041.safetensors",
+    "transformer.h.40.ln_mlp.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.40.mlp.dense_4h_to_h.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.40.mlp.dense_h_to_4h.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.40.self_attention.dense.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.40.self_attention.query_key_value.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.41.ln_attn.bias": "model-00022-of-00041.safetensors",
+    "transformer.h.41.ln_attn.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.41.ln_mlp.bias": "model-00022-of-00041.safetensors",
+    "transformer.h.41.ln_mlp.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.41.mlp.dense_4h_to_h.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.41.mlp.dense_h_to_4h.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.41.self_attention.dense.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.41.self_attention.query_key_value.weight": "model-00021-of-00041.safetensors",
+    "transformer.h.42.ln_attn.bias": "model-00022-of-00041.safetensors",
+    "transformer.h.42.ln_attn.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.42.ln_mlp.bias": "model-00022-of-00041.safetensors",
+    "transformer.h.42.ln_mlp.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.42.mlp.dense_4h_to_h.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.42.mlp.dense_h_to_4h.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.42.self_attention.dense.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.42.self_attention.query_key_value.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.43.ln_attn.bias": "model-00023-of-00041.safetensors",
+    "transformer.h.43.ln_attn.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.43.ln_mlp.bias": "model-00023-of-00041.safetensors",
+    "transformer.h.43.ln_mlp.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.43.mlp.dense_4h_to_h.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.43.mlp.dense_h_to_4h.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.43.self_attention.dense.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.43.self_attention.query_key_value.weight": "model-00022-of-00041.safetensors",
+    "transformer.h.44.ln_attn.bias": "model-00023-of-00041.safetensors",
+    "transformer.h.44.ln_attn.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.44.ln_mlp.bias": "model-00023-of-00041.safetensors",
+    "transformer.h.44.ln_mlp.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.44.mlp.dense_4h_to_h.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.44.mlp.dense_h_to_4h.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.44.self_attention.dense.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.44.self_attention.query_key_value.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.45.ln_attn.bias": "model-00024-of-00041.safetensors",
+    "transformer.h.45.ln_attn.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.45.ln_mlp.bias": "model-00024-of-00041.safetensors",
+    "transformer.h.45.ln_mlp.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.45.mlp.dense_4h_to_h.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.45.mlp.dense_h_to_4h.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.45.self_attention.dense.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.45.self_attention.query_key_value.weight": "model-00023-of-00041.safetensors",
+    "transformer.h.46.ln_attn.bias": "model-00024-of-00041.safetensors",
+    "transformer.h.46.ln_attn.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.46.ln_mlp.bias": "model-00024-of-00041.safetensors",
+    "transformer.h.46.ln_mlp.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.46.mlp.dense_4h_to_h.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.46.mlp.dense_h_to_4h.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.46.self_attention.dense.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.46.self_attention.query_key_value.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.47.ln_attn.bias": "model-00025-of-00041.safetensors",
+    "transformer.h.47.ln_attn.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.47.ln_mlp.bias": "model-00025-of-00041.safetensors",
+    "transformer.h.47.ln_mlp.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.47.mlp.dense_4h_to_h.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.47.mlp.dense_h_to_4h.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.47.self_attention.dense.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.47.self_attention.query_key_value.weight": "model-00024-of-00041.safetensors",
+    "transformer.h.48.ln_attn.bias": "model-00025-of-00041.safetensors",
+    "transformer.h.48.ln_attn.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.48.ln_mlp.bias": "model-00025-of-00041.safetensors",
+    "transformer.h.48.ln_mlp.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.48.mlp.dense_4h_to_h.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.48.mlp.dense_h_to_4h.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.48.self_attention.dense.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.48.self_attention.query_key_value.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.49.ln_attn.bias": "model-00026-of-00041.safetensors",
+    "transformer.h.49.ln_attn.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.49.ln_mlp.bias": "model-00026-of-00041.safetensors",
+    "transformer.h.49.ln_mlp.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.49.mlp.dense_4h_to_h.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.49.mlp.dense_h_to_4h.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.49.self_attention.dense.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.49.self_attention.query_key_value.weight": "model-00025-of-00041.safetensors",
+    "transformer.h.5.ln_attn.bias": "model-00004-of-00041.safetensors",
+    "transformer.h.5.ln_attn.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.5.ln_mlp.bias": "model-00004-of-00041.safetensors",
+    "transformer.h.5.ln_mlp.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.5.mlp.dense_4h_to_h.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.5.mlp.dense_h_to_4h.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.5.self_attention.dense.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.5.self_attention.query_key_value.weight": "model-00003-of-00041.safetensors",
+    "transformer.h.50.ln_attn.bias": "model-00026-of-00041.safetensors",
+    "transformer.h.50.ln_attn.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.50.ln_mlp.bias": "model-00026-of-00041.safetensors",
+    "transformer.h.50.ln_mlp.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.50.mlp.dense_4h_to_h.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.50.mlp.dense_h_to_4h.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.50.self_attention.dense.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.50.self_attention.query_key_value.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.51.ln_attn.bias": "model-00027-of-00041.safetensors",
+    "transformer.h.51.ln_attn.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.51.ln_mlp.bias": "model-00027-of-00041.safetensors",
+    "transformer.h.51.ln_mlp.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.51.mlp.dense_4h_to_h.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.51.mlp.dense_h_to_4h.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.51.self_attention.dense.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.51.self_attention.query_key_value.weight": "model-00026-of-00041.safetensors",
+    "transformer.h.52.ln_attn.bias": "model-00027-of-00041.safetensors",
+    "transformer.h.52.ln_attn.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.52.ln_mlp.bias": "model-00027-of-00041.safetensors",
+    "transformer.h.52.ln_mlp.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.52.mlp.dense_4h_to_h.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.52.mlp.dense_h_to_4h.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.52.self_attention.dense.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.52.self_attention.query_key_value.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.53.ln_attn.bias": "model-00028-of-00041.safetensors",
+    "transformer.h.53.ln_attn.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.53.ln_mlp.bias": "model-00028-of-00041.safetensors",
+    "transformer.h.53.ln_mlp.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.53.mlp.dense_4h_to_h.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.53.mlp.dense_h_to_4h.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.53.self_attention.dense.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.53.self_attention.query_key_value.weight": "model-00027-of-00041.safetensors",
+    "transformer.h.54.ln_attn.bias": "model-00028-of-00041.safetensors",
+    "transformer.h.54.ln_attn.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.54.ln_mlp.bias": "model-00028-of-00041.safetensors",
+    "transformer.h.54.ln_mlp.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.54.mlp.dense_4h_to_h.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.54.mlp.dense_h_to_4h.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.54.self_attention.dense.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.54.self_attention.query_key_value.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.55.ln_attn.bias": "model-00029-of-00041.safetensors",
+    "transformer.h.55.ln_attn.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.55.ln_mlp.bias": "model-00029-of-00041.safetensors",
+    "transformer.h.55.ln_mlp.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.55.mlp.dense_4h_to_h.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.55.mlp.dense_h_to_4h.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.55.self_attention.dense.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.55.self_attention.query_key_value.weight": "model-00028-of-00041.safetensors",
+    "transformer.h.56.ln_attn.bias": "model-00029-of-00041.safetensors",
+    "transformer.h.56.ln_attn.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.56.ln_mlp.bias": "model-00029-of-00041.safetensors",
+    "transformer.h.56.ln_mlp.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.56.mlp.dense_4h_to_h.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.56.mlp.dense_h_to_4h.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.56.self_attention.dense.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.56.self_attention.query_key_value.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.57.ln_attn.bias": "model-00030-of-00041.safetensors",
+    "transformer.h.57.ln_attn.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.57.ln_mlp.bias": "model-00030-of-00041.safetensors",
+    "transformer.h.57.ln_mlp.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.57.mlp.dense_4h_to_h.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.57.mlp.dense_h_to_4h.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.57.self_attention.dense.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.57.self_attention.query_key_value.weight": "model-00029-of-00041.safetensors",
+    "transformer.h.58.ln_attn.bias": "model-00030-of-00041.safetensors",
+    "transformer.h.58.ln_attn.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.58.ln_mlp.bias": "model-00030-of-00041.safetensors",
+    "transformer.h.58.ln_mlp.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.58.mlp.dense_4h_to_h.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.58.mlp.dense_h_to_4h.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.58.self_attention.dense.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.58.self_attention.query_key_value.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.59.ln_attn.bias": "model-00031-of-00041.safetensors",
+    "transformer.h.59.ln_attn.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.59.ln_mlp.bias": "model-00031-of-00041.safetensors",
+    "transformer.h.59.ln_mlp.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.59.mlp.dense_4h_to_h.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.59.mlp.dense_h_to_4h.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.59.self_attention.dense.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.59.self_attention.query_key_value.weight": "model-00030-of-00041.safetensors",
+    "transformer.h.6.ln_attn.bias": "model-00004-of-00041.safetensors",
+    "transformer.h.6.ln_attn.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.6.ln_mlp.bias": "model-00004-of-00041.safetensors",
+    "transformer.h.6.ln_mlp.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.6.mlp.dense_4h_to_h.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.6.mlp.dense_h_to_4h.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.6.self_attention.dense.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.6.self_attention.query_key_value.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.60.ln_attn.bias": "model-00031-of-00041.safetensors",
+    "transformer.h.60.ln_attn.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.60.ln_mlp.bias": "model-00031-of-00041.safetensors",
+    "transformer.h.60.ln_mlp.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.60.mlp.dense_4h_to_h.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.60.mlp.dense_h_to_4h.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.60.self_attention.dense.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.60.self_attention.query_key_value.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.61.ln_attn.bias": "model-00032-of-00041.safetensors",
+    "transformer.h.61.ln_attn.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.61.ln_mlp.bias": "model-00032-of-00041.safetensors",
+    "transformer.h.61.ln_mlp.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.61.mlp.dense_4h_to_h.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.61.mlp.dense_h_to_4h.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.61.self_attention.dense.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.61.self_attention.query_key_value.weight": "model-00031-of-00041.safetensors",
+    "transformer.h.62.ln_attn.bias": "model-00032-of-00041.safetensors",
+    "transformer.h.62.ln_attn.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.62.ln_mlp.bias": "model-00032-of-00041.safetensors",
+    "transformer.h.62.ln_mlp.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.62.mlp.dense_4h_to_h.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.62.mlp.dense_h_to_4h.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.62.self_attention.dense.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.62.self_attention.query_key_value.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.63.ln_attn.bias": "model-00033-of-00041.safetensors",
+    "transformer.h.63.ln_attn.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.63.ln_mlp.bias": "model-00033-of-00041.safetensors",
+    "transformer.h.63.ln_mlp.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.63.mlp.dense_4h_to_h.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.63.mlp.dense_h_to_4h.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.63.self_attention.dense.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.63.self_attention.query_key_value.weight": "model-00032-of-00041.safetensors",
+    "transformer.h.64.ln_attn.bias": "model-00033-of-00041.safetensors",
+    "transformer.h.64.ln_attn.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.64.ln_mlp.bias": "model-00033-of-00041.safetensors",
+    "transformer.h.64.ln_mlp.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.64.mlp.dense_4h_to_h.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.64.mlp.dense_h_to_4h.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.64.self_attention.dense.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.64.self_attention.query_key_value.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.65.ln_attn.bias": "model-00034-of-00041.safetensors",
+    "transformer.h.65.ln_attn.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.65.ln_mlp.bias": "model-00034-of-00041.safetensors",
+    "transformer.h.65.ln_mlp.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.65.mlp.dense_4h_to_h.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.65.mlp.dense_h_to_4h.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.65.self_attention.dense.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.65.self_attention.query_key_value.weight": "model-00033-of-00041.safetensors",
+    "transformer.h.66.ln_attn.bias": "model-00034-of-00041.safetensors",
+    "transformer.h.66.ln_attn.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.66.ln_mlp.bias": "model-00034-of-00041.safetensors",
+    "transformer.h.66.ln_mlp.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.66.mlp.dense_4h_to_h.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.66.mlp.dense_h_to_4h.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.66.self_attention.dense.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.66.self_attention.query_key_value.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.67.ln_attn.bias": "model-00035-of-00041.safetensors",
+    "transformer.h.67.ln_attn.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.67.ln_mlp.bias": "model-00035-of-00041.safetensors",
+    "transformer.h.67.ln_mlp.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.67.mlp.dense_4h_to_h.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.67.mlp.dense_h_to_4h.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.67.self_attention.dense.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.67.self_attention.query_key_value.weight": "model-00034-of-00041.safetensors",
+    "transformer.h.68.ln_attn.bias": "model-00035-of-00041.safetensors",
+    "transformer.h.68.ln_attn.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.68.ln_mlp.bias": "model-00035-of-00041.safetensors",
+    "transformer.h.68.ln_mlp.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.68.mlp.dense_4h_to_h.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.68.mlp.dense_h_to_4h.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.68.self_attention.dense.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.68.self_attention.query_key_value.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.69.ln_attn.bias": "model-00036-of-00041.safetensors",
+    "transformer.h.69.ln_attn.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.69.ln_mlp.bias": "model-00036-of-00041.safetensors",
+    "transformer.h.69.ln_mlp.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.69.mlp.dense_4h_to_h.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.69.mlp.dense_h_to_4h.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.69.self_attention.dense.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.69.self_attention.query_key_value.weight": "model-00035-of-00041.safetensors",
+    "transformer.h.7.ln_attn.bias": "model-00005-of-00041.safetensors",
+    "transformer.h.7.ln_attn.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.7.ln_mlp.bias": "model-00005-of-00041.safetensors",
+    "transformer.h.7.ln_mlp.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.7.mlp.dense_4h_to_h.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.7.mlp.dense_h_to_4h.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.7.self_attention.dense.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.7.self_attention.query_key_value.weight": "model-00004-of-00041.safetensors",
+    "transformer.h.70.ln_attn.bias": "model-00036-of-00041.safetensors",
+    "transformer.h.70.ln_attn.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.70.ln_mlp.bias": "model-00036-of-00041.safetensors",
+    "transformer.h.70.ln_mlp.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.70.mlp.dense_4h_to_h.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.70.mlp.dense_h_to_4h.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.70.self_attention.dense.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.70.self_attention.query_key_value.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.71.ln_attn.bias": "model-00037-of-00041.safetensors",
+    "transformer.h.71.ln_attn.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.71.ln_mlp.bias": "model-00037-of-00041.safetensors",
+    "transformer.h.71.ln_mlp.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.71.mlp.dense_4h_to_h.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.71.mlp.dense_h_to_4h.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.71.self_attention.dense.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.71.self_attention.query_key_value.weight": "model-00036-of-00041.safetensors",
+    "transformer.h.72.ln_attn.bias": "model-00037-of-00041.safetensors",
+    "transformer.h.72.ln_attn.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.72.ln_mlp.bias": "model-00037-of-00041.safetensors",
+    "transformer.h.72.ln_mlp.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.72.mlp.dense_4h_to_h.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.72.mlp.dense_h_to_4h.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.72.self_attention.dense.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.72.self_attention.query_key_value.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.73.ln_attn.bias": "model-00038-of-00041.safetensors",
+    "transformer.h.73.ln_attn.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.73.ln_mlp.bias": "model-00038-of-00041.safetensors",
+    "transformer.h.73.ln_mlp.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.73.mlp.dense_4h_to_h.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.73.mlp.dense_h_to_4h.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.73.self_attention.dense.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.73.self_attention.query_key_value.weight": "model-00037-of-00041.safetensors",
+    "transformer.h.74.ln_attn.bias": "model-00038-of-00041.safetensors",
+    "transformer.h.74.ln_attn.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.74.ln_mlp.bias": "model-00038-of-00041.safetensors",
+    "transformer.h.74.ln_mlp.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.74.mlp.dense_4h_to_h.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.74.mlp.dense_h_to_4h.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.74.self_attention.dense.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.74.self_attention.query_key_value.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.75.ln_attn.bias": "model-00039-of-00041.safetensors",
+    "transformer.h.75.ln_attn.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.75.ln_mlp.bias": "model-00039-of-00041.safetensors",
+    "transformer.h.75.ln_mlp.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.75.mlp.dense_4h_to_h.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.75.mlp.dense_h_to_4h.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.75.self_attention.dense.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.75.self_attention.query_key_value.weight": "model-00038-of-00041.safetensors",
+    "transformer.h.76.ln_attn.bias": "model-00039-of-00041.safetensors",
+    "transformer.h.76.ln_attn.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.76.ln_mlp.bias": "model-00039-of-00041.safetensors",
+    "transformer.h.76.ln_mlp.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.76.mlp.dense_4h_to_h.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.76.mlp.dense_h_to_4h.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.76.self_attention.dense.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.76.self_attention.query_key_value.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.77.ln_attn.bias": "model-00040-of-00041.safetensors",
+    "transformer.h.77.ln_attn.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.77.ln_mlp.bias": "model-00040-of-00041.safetensors",
+    "transformer.h.77.ln_mlp.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.77.mlp.dense_4h_to_h.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.77.mlp.dense_h_to_4h.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.77.self_attention.dense.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.77.self_attention.query_key_value.weight": "model-00039-of-00041.safetensors",
+    "transformer.h.78.ln_attn.bias": "model-00040-of-00041.safetensors",
+    "transformer.h.78.ln_attn.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.78.ln_mlp.bias": "model-00040-of-00041.safetensors",
+    "transformer.h.78.ln_mlp.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.78.mlp.dense_4h_to_h.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.78.mlp.dense_h_to_4h.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.78.self_attention.dense.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.78.self_attention.query_key_value.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.79.ln_attn.bias": "model-00041-of-00041.safetensors",
+    "transformer.h.79.ln_attn.weight": "model-00041-of-00041.safetensors",
+    "transformer.h.79.ln_mlp.bias": "model-00041-of-00041.safetensors",
+    "transformer.h.79.ln_mlp.weight": "model-00041-of-00041.safetensors",
+    "transformer.h.79.mlp.dense_4h_to_h.weight": "model-00041-of-00041.safetensors",
+    "transformer.h.79.mlp.dense_h_to_4h.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.79.self_attention.dense.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.79.self_attention.query_key_value.weight": "model-00040-of-00041.safetensors",
+    "transformer.h.8.ln_attn.bias": "model-00005-of-00041.safetensors",
+    "transformer.h.8.ln_attn.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.8.ln_mlp.bias": "model-00005-of-00041.safetensors",
+    "transformer.h.8.ln_mlp.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.8.mlp.dense_4h_to_h.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.8.mlp.dense_h_to_4h.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.8.self_attention.dense.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.8.self_attention.query_key_value.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.9.ln_attn.bias": "model-00006-of-00041.safetensors",
+    "transformer.h.9.ln_attn.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.9.ln_mlp.bias": "model-00006-of-00041.safetensors",
+    "transformer.h.9.ln_mlp.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.9.mlp.dense_4h_to_h.weight": "model-00006-of-00041.safetensors",
+    "transformer.h.9.mlp.dense_h_to_4h.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.9.self_attention.dense.weight": "model-00005-of-00041.safetensors",
+    "transformer.h.9.self_attention.query_key_value.weight": "model-00005-of-00041.safetensors",
+    "transformer.ln_f.bias": "model-00041-of-00041.safetensors",
+    "transformer.ln_f.weight": "model-00041-of-00041.safetensors",
+    "transformer.word_embeddings.weight": "model-00001-of-00041.safetensors"
+  }
+}