Upload folder using huggingface_hub

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "openbuddy-f180-v13",
+  "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,
+  "max_position_embeddings": 2048,
+  "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,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.33.2",
+  "use_cache": true,
+  "vocab_size": 70656
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 11,
+  "eos_token_id": 11,
+  "transformers_version": "4.33.2"
+}

pytorch_model-00001-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2400302be6d78722de54faf92ba16d7fee389826b62401b2596b20ccae465b80
+size 9213960603

pytorch_model-00002-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eff1bc01d083faee3f708dc4fea1f81760c9a46bcb4b7415416cba0f72f67031
+size 8879584803

pytorch_model-00003-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3eb79cf306122e498844f756e71a5b9607fbf067daafae5fc4e6b45a616445b
+size 8879584803

pytorch_model-00004-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c8382e84b60182dcc8e293e799275e9d7c7e5eeed3e201924e8c2d25f45a51a
+size 8879584803

pytorch_model-00005-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7dda6613390423a87eb777fb3dae8b94e3a43e77ca2fda041005edb040e5bbb1
+size 8879584803

pytorch_model-00006-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df2be1643cdd078d5b93e75d6a71e3a086aa2029426eb5a1ae77f44e882c5e73
+size 8879584803

pytorch_model-00007-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21a104f7b5f821ac75c01a10b212234c093a545fb22409ba590633a10ee99aef
+size 8879584803

pytorch_model-00008-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc794d3a00f8ddf37388851e9c0eb37097158769547f213eb5f6e7e2cf1105be
+size 8879584803

pytorch_model-00009-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e06c77e022e81a75eeb537dba4c145a36b8ec812cfe382abf367ea5c14a0840c
+size 8879584803

pytorch_model-00010-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6074b7bc0bf78aaf86042d16943e449049d64dba9ab405b71776e0fabd060b95
+size 8879584803

pytorch_model-00011-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:044325179cabb3d46fae7ba0935817055326edfbc75894d707318da583a5dc7a
+size 8879584803

pytorch_model-00012-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21fefaef7884bf8b181e635d233a16b1ec0bd69e4578dfbb37bd8a3b4698994d
+size 8879584803

pytorch_model-00013-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f0a79f8574076e0045eb1d3208f5de23085ff0a8a547f5ad2f7e747df73e8af
+size 8879584803

pytorch_model-00014-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc8ce52d1b275eba54c51cc428d8c321bfbe11bc725e7cbf270647020e22b3fb
+size 8879584803

pytorch_model-00015-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2539082b6009ae5ea75146712c8c8485fc36dfe3589802eb1590ecbb9e29c3e9
+size 8879584803

pytorch_model-00016-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60223ec936c0bf8ed71ca48d04bea95c7dfe3407ff5f3840b40ac4016db5ea11
+size 8879584803

pytorch_model-00017-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba7daef62c8f1e47e87304c19bc223fce045be62277cca887dfa11d45d8b4163
+size 8879584803

pytorch_model-00018-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5965d2353fd8e9bc5005b02848ef046446d15032b19e5e2e13e84baf4a441604
+size 8879584803

pytorch_model-00019-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:70475444b1fe2aced6d0535a2f6d2f1a77185cba99c91d83e8329b0cd4407d9a
+size 8879584803

pytorch_model-00020-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30ab55830f524fa81df3bd46ccdade32283eaf36f51855b6e982146c7aacb339
+size 8879584803

pytorch_model-00021-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23d11867c86a8fb22d4e1dd884760fed718010c42cdd9d80a8dfcf038204e186
+size 8879584803

pytorch_model-00022-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6d2164a4ac43e654d9ac275ff017f7d19c463a41db6a96012f6ab5bb8ea3648
+size 8879584803

pytorch_model-00023-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:561e212968922757061961eded95c1496e895fa930672c1fe43df6c46ad74dc8
+size 8879584803

pytorch_model-00024-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ee496cb18297df7e0692705e1a3e0fa283d922f8d65f757605bcbd25773f887
+size 8879584803

pytorch_model-00025-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44475f754ced99f18182a6db633ed302a0cc485728d0e180d3dcf5885ef75175
+size 8879584803

pytorch_model-00026-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dcf0d33fe6a7d3140a80dfd8faf6390c7612f6f7f31961f112a06b72162cadb4
+size 8879584803

pytorch_model-00027-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:507c39ada7f74f617117b86dd1d5074496d98226a40b205da577a984db4692be
+size 8879584803

pytorch_model-00028-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df40dee3470bfca55a677a0652264b54a9d151d15599468657f5851cab667bf1
+size 8879584803

pytorch_model-00029-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1c14e2b2b48a9339faf262b87f9352b3ee8d341d2f66644e5af9ec9e09c9cb9
+size 8879584803

pytorch_model-00030-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9c81d84b3df975fb22f85c708a38f9f1b1e92ed7fe5491ebcad93b9af565ae0
+size 8879584803

pytorch_model-00031-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e96dc1034a9194823ddca4d861b45d9af2c708a04f853928d8d7443ac4df657c
+size 8879584803

pytorch_model-00032-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12206a93dc36424a155efae690f5beae768b8cf4f056d7179a1989962b4bab47
+size 8879584803

pytorch_model-00033-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04594a1adc78c325064de50a53043146200855b71a57d36c3b17d9b1d1524a94
+size 8879584803

pytorch_model-00034-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe2a6bcd339d77bee7da9291272dcb738c74cf783cd421d1d6da26d2f7a30440
+size 8879584803

pytorch_model-00035-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f912140e778e50de4743e0ac0ede11d5df36a48f498d15d1eaf9dfa886ccff9
+size 8879584803

pytorch_model-00036-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ebd57f9b8a16697089a6b7f1b49f56b1b570d85a415171953f853e16eefeda92
+size 8879584803

pytorch_model-00037-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6e3d890d4f962d3c57572ba44fbcad6a1059113562d34f70c34a99913eaab01
+size 8879584803

pytorch_model-00038-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1babef3df2fb7546616cf8be03e871bbe18e00d91a6dec04ab51fffaf0f5d3f0
+size 8879584803

pytorch_model-00039-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3e16d3ea9b71e93c03c8ef04e5778cc9377362d24244801683f5964797e8eb1
+size 8879584803

pytorch_model-00040-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ff1ff87c14c3a68981097fabeea84a448ac367056a184eb1d089c8fb8d7de61
+size 8879584803

pytorch_model-00041-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c8772475053900345a6c204898ba9f7e8fdca85a712378214161096f9c5b0136
+size 1763885712

pytorch_model.bin.index.json

@@ -0,0 +1,651 @@
+{
+  "metadata": {
+    "total_size": 357281425408
+  },
+  "weight_map": {
+    "lm_head.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.0.ln_attn.bias": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.0.ln_attn.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.0.ln_mlp.bias": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.0.ln_mlp.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.0.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.0.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.0.self_attention.dense.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.0.self_attention.query_key_value.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.1.ln_attn.bias": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.1.ln_attn.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.1.ln_mlp.bias": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.1.ln_mlp.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.1.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.1.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.1.self_attention.dense.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.1.self_attention.query_key_value.weight": "pytorch_model-00001-of-00041.bin",
+    "transformer.h.10.ln_attn.bias": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.10.ln_attn.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.10.ln_mlp.bias": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.10.ln_mlp.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.10.mlp.dense_4h_to_h.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.10.mlp.dense_h_to_4h.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.10.self_attention.dense.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.10.self_attention.query_key_value.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.11.ln_attn.bias": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.11.ln_attn.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.11.ln_mlp.bias": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.11.ln_mlp.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.11.mlp.dense_4h_to_h.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.11.mlp.dense_h_to_4h.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.11.self_attention.dense.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.11.self_attention.query_key_value.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.12.ln_attn.bias": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.12.ln_attn.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.12.ln_mlp.bias": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.12.ln_mlp.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.12.mlp.dense_4h_to_h.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.12.mlp.dense_h_to_4h.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.12.self_attention.dense.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.12.self_attention.query_key_value.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.13.ln_attn.bias": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.13.ln_attn.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.13.ln_mlp.bias": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.13.ln_mlp.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.13.mlp.dense_4h_to_h.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.13.mlp.dense_h_to_4h.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.13.self_attention.dense.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.13.self_attention.query_key_value.weight": "pytorch_model-00007-of-00041.bin",
+    "transformer.h.14.ln_attn.bias": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.14.ln_attn.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.14.ln_mlp.bias": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.14.ln_mlp.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.14.mlp.dense_4h_to_h.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.14.mlp.dense_h_to_4h.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.14.self_attention.dense.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.14.self_attention.query_key_value.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.15.ln_attn.bias": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.15.ln_attn.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.15.ln_mlp.bias": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.15.ln_mlp.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.15.mlp.dense_4h_to_h.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.15.mlp.dense_h_to_4h.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.15.self_attention.dense.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.15.self_attention.query_key_value.weight": "pytorch_model-00008-of-00041.bin",
+    "transformer.h.16.ln_attn.bias": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.16.ln_attn.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.16.ln_mlp.bias": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.16.ln_mlp.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.16.mlp.dense_4h_to_h.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.16.mlp.dense_h_to_4h.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.16.self_attention.dense.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.16.self_attention.query_key_value.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.17.ln_attn.bias": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.17.ln_attn.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.17.ln_mlp.bias": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.17.ln_mlp.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.17.mlp.dense_4h_to_h.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.17.mlp.dense_h_to_4h.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.17.self_attention.dense.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.17.self_attention.query_key_value.weight": "pytorch_model-00009-of-00041.bin",
+    "transformer.h.18.ln_attn.bias": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.18.ln_attn.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.18.ln_mlp.bias": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.18.ln_mlp.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.18.mlp.dense_4h_to_h.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.18.mlp.dense_h_to_4h.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.18.self_attention.dense.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.18.self_attention.query_key_value.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.19.ln_attn.bias": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.19.ln_attn.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.19.ln_mlp.bias": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.19.ln_mlp.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.19.mlp.dense_4h_to_h.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.19.mlp.dense_h_to_4h.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.19.self_attention.dense.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.19.self_attention.query_key_value.weight": "pytorch_model-00010-of-00041.bin",
+    "transformer.h.2.ln_attn.bias": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.2.ln_attn.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.2.ln_mlp.bias": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.2.ln_mlp.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.2.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.2.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.2.self_attention.dense.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.2.self_attention.query_key_value.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.20.ln_attn.bias": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.20.ln_attn.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.20.ln_mlp.bias": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.20.ln_mlp.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.20.mlp.dense_4h_to_h.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.20.mlp.dense_h_to_4h.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.20.self_attention.dense.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.20.self_attention.query_key_value.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.21.ln_attn.bias": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.21.ln_attn.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.21.ln_mlp.bias": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.21.ln_mlp.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.21.mlp.dense_4h_to_h.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.21.mlp.dense_h_to_4h.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.21.self_attention.dense.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.21.self_attention.query_key_value.weight": "pytorch_model-00011-of-00041.bin",
+    "transformer.h.22.ln_attn.bias": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.22.ln_attn.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.22.ln_mlp.bias": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.22.ln_mlp.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.22.mlp.dense_4h_to_h.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.22.mlp.dense_h_to_4h.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.22.self_attention.dense.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.22.self_attention.query_key_value.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.23.ln_attn.bias": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.23.ln_attn.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.23.ln_mlp.bias": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.23.ln_mlp.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.23.mlp.dense_4h_to_h.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.23.mlp.dense_h_to_4h.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.23.self_attention.dense.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.23.self_attention.query_key_value.weight": "pytorch_model-00012-of-00041.bin",
+    "transformer.h.24.ln_attn.bias": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.24.ln_attn.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.24.ln_mlp.bias": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.24.ln_mlp.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.24.mlp.dense_4h_to_h.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.24.mlp.dense_h_to_4h.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.24.self_attention.dense.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.24.self_attention.query_key_value.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.25.ln_attn.bias": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.25.ln_attn.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.25.ln_mlp.bias": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.25.ln_mlp.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.25.mlp.dense_4h_to_h.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.25.mlp.dense_h_to_4h.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.25.self_attention.dense.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.25.self_attention.query_key_value.weight": "pytorch_model-00013-of-00041.bin",
+    "transformer.h.26.ln_attn.bias": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.26.ln_attn.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.26.ln_mlp.bias": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.26.ln_mlp.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.26.mlp.dense_4h_to_h.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.26.mlp.dense_h_to_4h.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.26.self_attention.dense.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.26.self_attention.query_key_value.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.27.ln_attn.bias": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.27.ln_attn.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.27.ln_mlp.bias": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.27.ln_mlp.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.27.mlp.dense_4h_to_h.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.27.mlp.dense_h_to_4h.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.27.self_attention.dense.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.27.self_attention.query_key_value.weight": "pytorch_model-00014-of-00041.bin",
+    "transformer.h.28.ln_attn.bias": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.28.ln_attn.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.28.ln_mlp.bias": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.28.ln_mlp.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.28.mlp.dense_4h_to_h.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.28.mlp.dense_h_to_4h.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.28.self_attention.dense.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.28.self_attention.query_key_value.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.29.ln_attn.bias": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.29.ln_attn.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.29.ln_mlp.bias": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.29.ln_mlp.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.29.mlp.dense_4h_to_h.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.29.mlp.dense_h_to_4h.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.29.self_attention.dense.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.29.self_attention.query_key_value.weight": "pytorch_model-00015-of-00041.bin",
+    "transformer.h.3.ln_attn.bias": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.3.ln_attn.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.3.ln_mlp.bias": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.3.ln_mlp.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.3.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.3.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.3.self_attention.dense.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.3.self_attention.query_key_value.weight": "pytorch_model-00002-of-00041.bin",
+    "transformer.h.30.ln_attn.bias": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.30.ln_attn.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.30.ln_mlp.bias": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.30.ln_mlp.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.30.mlp.dense_4h_to_h.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.30.mlp.dense_h_to_4h.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.30.self_attention.dense.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.30.self_attention.query_key_value.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.31.ln_attn.bias": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.31.ln_attn.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.31.ln_mlp.bias": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.31.ln_mlp.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.31.mlp.dense_4h_to_h.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.31.mlp.dense_h_to_4h.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.31.self_attention.dense.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.31.self_attention.query_key_value.weight": "pytorch_model-00016-of-00041.bin",
+    "transformer.h.32.ln_attn.bias": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.32.ln_attn.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.32.ln_mlp.bias": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.32.ln_mlp.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.32.mlp.dense_4h_to_h.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.32.mlp.dense_h_to_4h.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.32.self_attention.dense.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.32.self_attention.query_key_value.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.33.ln_attn.bias": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.33.ln_attn.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.33.ln_mlp.bias": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.33.ln_mlp.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.33.mlp.dense_4h_to_h.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.33.mlp.dense_h_to_4h.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.33.self_attention.dense.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.33.self_attention.query_key_value.weight": "pytorch_model-00017-of-00041.bin",
+    "transformer.h.34.ln_attn.bias": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.34.ln_attn.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.34.ln_mlp.bias": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.34.ln_mlp.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.34.mlp.dense_4h_to_h.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.34.mlp.dense_h_to_4h.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.34.self_attention.dense.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.34.self_attention.query_key_value.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.35.ln_attn.bias": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.35.ln_attn.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.35.ln_mlp.bias": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.35.ln_mlp.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.35.mlp.dense_4h_to_h.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.35.mlp.dense_h_to_4h.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.35.self_attention.dense.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.35.self_attention.query_key_value.weight": "pytorch_model-00018-of-00041.bin",
+    "transformer.h.36.ln_attn.bias": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.36.ln_attn.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.36.ln_mlp.bias": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.36.ln_mlp.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.36.mlp.dense_4h_to_h.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.36.mlp.dense_h_to_4h.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.36.self_attention.dense.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.36.self_attention.query_key_value.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.37.ln_attn.bias": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.37.ln_attn.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.37.ln_mlp.bias": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.37.ln_mlp.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.37.mlp.dense_4h_to_h.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.37.mlp.dense_h_to_4h.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.37.self_attention.dense.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.37.self_attention.query_key_value.weight": "pytorch_model-00019-of-00041.bin",
+    "transformer.h.38.ln_attn.bias": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.38.ln_attn.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.38.ln_mlp.bias": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.38.ln_mlp.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.38.mlp.dense_4h_to_h.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.38.mlp.dense_h_to_4h.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.38.self_attention.dense.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.38.self_attention.query_key_value.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.39.ln_attn.bias": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.39.ln_attn.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.39.ln_mlp.bias": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.39.ln_mlp.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.39.mlp.dense_4h_to_h.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.39.mlp.dense_h_to_4h.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.39.self_attention.dense.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.39.self_attention.query_key_value.weight": "pytorch_model-00020-of-00041.bin",
+    "transformer.h.4.ln_attn.bias": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.4.ln_attn.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.4.ln_mlp.bias": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.4.ln_mlp.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.4.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.4.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.4.self_attention.dense.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.4.self_attention.query_key_value.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.40.ln_attn.bias": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.40.ln_attn.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.40.ln_mlp.bias": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.40.ln_mlp.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.40.mlp.dense_4h_to_h.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.40.mlp.dense_h_to_4h.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.40.self_attention.dense.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.40.self_attention.query_key_value.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.41.ln_attn.bias": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.41.ln_attn.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.41.ln_mlp.bias": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.41.ln_mlp.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.41.mlp.dense_4h_to_h.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.41.mlp.dense_h_to_4h.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.41.self_attention.dense.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.41.self_attention.query_key_value.weight": "pytorch_model-00021-of-00041.bin",
+    "transformer.h.42.ln_attn.bias": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.42.ln_attn.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.42.ln_mlp.bias": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.42.ln_mlp.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.42.mlp.dense_4h_to_h.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.42.mlp.dense_h_to_4h.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.42.self_attention.dense.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.42.self_attention.query_key_value.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.43.ln_attn.bias": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.43.ln_attn.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.43.ln_mlp.bias": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.43.ln_mlp.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.43.mlp.dense_4h_to_h.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.43.mlp.dense_h_to_4h.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.43.self_attention.dense.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.43.self_attention.query_key_value.weight": "pytorch_model-00022-of-00041.bin",
+    "transformer.h.44.ln_attn.bias": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.44.ln_attn.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.44.ln_mlp.bias": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.44.ln_mlp.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.44.mlp.dense_4h_to_h.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.44.mlp.dense_h_to_4h.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.44.self_attention.dense.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.44.self_attention.query_key_value.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.45.ln_attn.bias": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.45.ln_attn.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.45.ln_mlp.bias": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.45.ln_mlp.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.45.mlp.dense_4h_to_h.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.45.mlp.dense_h_to_4h.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.45.self_attention.dense.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.45.self_attention.query_key_value.weight": "pytorch_model-00023-of-00041.bin",
+    "transformer.h.46.ln_attn.bias": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.46.ln_attn.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.46.ln_mlp.bias": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.46.ln_mlp.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.46.mlp.dense_4h_to_h.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.46.mlp.dense_h_to_4h.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.46.self_attention.dense.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.46.self_attention.query_key_value.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.47.ln_attn.bias": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.47.ln_attn.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.47.ln_mlp.bias": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.47.ln_mlp.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.47.mlp.dense_4h_to_h.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.47.mlp.dense_h_to_4h.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.47.self_attention.dense.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.47.self_attention.query_key_value.weight": "pytorch_model-00024-of-00041.bin",
+    "transformer.h.48.ln_attn.bias": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.48.ln_attn.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.48.ln_mlp.bias": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.48.ln_mlp.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.48.mlp.dense_4h_to_h.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.48.mlp.dense_h_to_4h.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.48.self_attention.dense.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.48.self_attention.query_key_value.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.49.ln_attn.bias": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.49.ln_attn.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.49.ln_mlp.bias": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.49.ln_mlp.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.49.mlp.dense_4h_to_h.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.49.mlp.dense_h_to_4h.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.49.self_attention.dense.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.49.self_attention.query_key_value.weight": "pytorch_model-00025-of-00041.bin",
+    "transformer.h.5.ln_attn.bias": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.5.ln_attn.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.5.ln_mlp.bias": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.5.ln_mlp.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.5.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.5.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.5.self_attention.dense.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.5.self_attention.query_key_value.weight": "pytorch_model-00003-of-00041.bin",
+    "transformer.h.50.ln_attn.bias": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.50.ln_attn.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.50.ln_mlp.bias": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.50.ln_mlp.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.50.mlp.dense_4h_to_h.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.50.mlp.dense_h_to_4h.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.50.self_attention.dense.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.50.self_attention.query_key_value.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.51.ln_attn.bias": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.51.ln_attn.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.51.ln_mlp.bias": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.51.ln_mlp.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.51.mlp.dense_4h_to_h.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.51.mlp.dense_h_to_4h.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.51.self_attention.dense.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.51.self_attention.query_key_value.weight": "pytorch_model-00026-of-00041.bin",
+    "transformer.h.52.ln_attn.bias": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.52.ln_attn.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.52.ln_mlp.bias": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.52.ln_mlp.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.52.mlp.dense_4h_to_h.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.52.mlp.dense_h_to_4h.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.52.self_attention.dense.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.52.self_attention.query_key_value.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.53.ln_attn.bias": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.53.ln_attn.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.53.ln_mlp.bias": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.53.ln_mlp.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.53.mlp.dense_4h_to_h.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.53.mlp.dense_h_to_4h.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.53.self_attention.dense.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.53.self_attention.query_key_value.weight": "pytorch_model-00027-of-00041.bin",
+    "transformer.h.54.ln_attn.bias": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.54.ln_attn.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.54.ln_mlp.bias": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.54.ln_mlp.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.54.mlp.dense_4h_to_h.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.54.mlp.dense_h_to_4h.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.54.self_attention.dense.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.54.self_attention.query_key_value.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.55.ln_attn.bias": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.55.ln_attn.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.55.ln_mlp.bias": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.55.ln_mlp.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.55.mlp.dense_4h_to_h.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.55.mlp.dense_h_to_4h.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.55.self_attention.dense.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.55.self_attention.query_key_value.weight": "pytorch_model-00028-of-00041.bin",
+    "transformer.h.56.ln_attn.bias": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.56.ln_attn.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.56.ln_mlp.bias": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.56.ln_mlp.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.56.mlp.dense_4h_to_h.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.56.mlp.dense_h_to_4h.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.56.self_attention.dense.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.56.self_attention.query_key_value.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.57.ln_attn.bias": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.57.ln_attn.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.57.ln_mlp.bias": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.57.ln_mlp.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.57.mlp.dense_4h_to_h.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.57.mlp.dense_h_to_4h.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.57.self_attention.dense.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.57.self_attention.query_key_value.weight": "pytorch_model-00029-of-00041.bin",
+    "transformer.h.58.ln_attn.bias": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.58.ln_attn.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.58.ln_mlp.bias": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.58.ln_mlp.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.58.mlp.dense_4h_to_h.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.58.mlp.dense_h_to_4h.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.58.self_attention.dense.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.58.self_attention.query_key_value.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.59.ln_attn.bias": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.59.ln_attn.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.59.ln_mlp.bias": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.59.ln_mlp.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.59.mlp.dense_4h_to_h.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.59.mlp.dense_h_to_4h.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.59.self_attention.dense.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.59.self_attention.query_key_value.weight": "pytorch_model-00030-of-00041.bin",
+    "transformer.h.6.ln_attn.bias": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.6.ln_attn.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.6.ln_mlp.bias": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.6.ln_mlp.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.6.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.6.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.6.self_attention.dense.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.6.self_attention.query_key_value.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.60.ln_attn.bias": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.60.ln_attn.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.60.ln_mlp.bias": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.60.ln_mlp.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.60.mlp.dense_4h_to_h.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.60.mlp.dense_h_to_4h.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.60.self_attention.dense.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.60.self_attention.query_key_value.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.61.ln_attn.bias": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.61.ln_attn.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.61.ln_mlp.bias": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.61.ln_mlp.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.61.mlp.dense_4h_to_h.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.61.mlp.dense_h_to_4h.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.61.self_attention.dense.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.61.self_attention.query_key_value.weight": "pytorch_model-00031-of-00041.bin",
+    "transformer.h.62.ln_attn.bias": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.62.ln_attn.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.62.ln_mlp.bias": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.62.ln_mlp.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.62.mlp.dense_4h_to_h.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.62.mlp.dense_h_to_4h.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.62.self_attention.dense.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.62.self_attention.query_key_value.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.63.ln_attn.bias": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.63.ln_attn.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.63.ln_mlp.bias": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.63.ln_mlp.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.63.mlp.dense_4h_to_h.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.63.mlp.dense_h_to_4h.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.63.self_attention.dense.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.63.self_attention.query_key_value.weight": "pytorch_model-00032-of-00041.bin",
+    "transformer.h.64.ln_attn.bias": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.64.ln_attn.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.64.ln_mlp.bias": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.64.ln_mlp.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.64.mlp.dense_4h_to_h.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.64.mlp.dense_h_to_4h.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.64.self_attention.dense.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.64.self_attention.query_key_value.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.65.ln_attn.bias": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.65.ln_attn.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.65.ln_mlp.bias": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.65.ln_mlp.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.65.mlp.dense_4h_to_h.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.65.mlp.dense_h_to_4h.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.65.self_attention.dense.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.65.self_attention.query_key_value.weight": "pytorch_model-00033-of-00041.bin",
+    "transformer.h.66.ln_attn.bias": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.66.ln_attn.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.66.ln_mlp.bias": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.66.ln_mlp.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.66.mlp.dense_4h_to_h.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.66.mlp.dense_h_to_4h.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.66.self_attention.dense.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.66.self_attention.query_key_value.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.67.ln_attn.bias": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.67.ln_attn.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.67.ln_mlp.bias": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.67.ln_mlp.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.67.mlp.dense_4h_to_h.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.67.mlp.dense_h_to_4h.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.67.self_attention.dense.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.67.self_attention.query_key_value.weight": "pytorch_model-00034-of-00041.bin",
+    "transformer.h.68.ln_attn.bias": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.68.ln_attn.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.68.ln_mlp.bias": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.68.ln_mlp.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.68.mlp.dense_4h_to_h.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.68.mlp.dense_h_to_4h.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.68.self_attention.dense.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.68.self_attention.query_key_value.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.69.ln_attn.bias": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.69.ln_attn.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.69.ln_mlp.bias": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.69.ln_mlp.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.69.mlp.dense_4h_to_h.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.69.mlp.dense_h_to_4h.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.69.self_attention.dense.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.69.self_attention.query_key_value.weight": "pytorch_model-00035-of-00041.bin",
+    "transformer.h.7.ln_attn.bias": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.7.ln_attn.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.7.ln_mlp.bias": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.7.ln_mlp.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.7.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.7.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.7.self_attention.dense.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.7.self_attention.query_key_value.weight": "pytorch_model-00004-of-00041.bin",
+    "transformer.h.70.ln_attn.bias": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.70.ln_attn.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.70.ln_mlp.bias": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.70.ln_mlp.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.70.mlp.dense_4h_to_h.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.70.mlp.dense_h_to_4h.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.70.self_attention.dense.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.70.self_attention.query_key_value.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.71.ln_attn.bias": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.71.ln_attn.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.71.ln_mlp.bias": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.71.ln_mlp.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.71.mlp.dense_4h_to_h.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.71.mlp.dense_h_to_4h.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.71.self_attention.dense.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.71.self_attention.query_key_value.weight": "pytorch_model-00036-of-00041.bin",
+    "transformer.h.72.ln_attn.bias": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.72.ln_attn.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.72.ln_mlp.bias": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.72.ln_mlp.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.72.mlp.dense_4h_to_h.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.72.mlp.dense_h_to_4h.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.72.self_attention.dense.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.72.self_attention.query_key_value.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.73.ln_attn.bias": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.73.ln_attn.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.73.ln_mlp.bias": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.73.ln_mlp.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.73.mlp.dense_4h_to_h.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.73.mlp.dense_h_to_4h.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.73.self_attention.dense.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.73.self_attention.query_key_value.weight": "pytorch_model-00037-of-00041.bin",
+    "transformer.h.74.ln_attn.bias": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.74.ln_attn.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.74.ln_mlp.bias": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.74.ln_mlp.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.74.mlp.dense_4h_to_h.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.74.mlp.dense_h_to_4h.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.74.self_attention.dense.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.74.self_attention.query_key_value.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.75.ln_attn.bias": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.75.ln_attn.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.75.ln_mlp.bias": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.75.ln_mlp.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.75.mlp.dense_4h_to_h.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.75.mlp.dense_h_to_4h.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.75.self_attention.dense.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.75.self_attention.query_key_value.weight": "pytorch_model-00038-of-00041.bin",
+    "transformer.h.76.ln_attn.bias": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.76.ln_attn.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.76.ln_mlp.bias": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.76.ln_mlp.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.76.mlp.dense_4h_to_h.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.76.mlp.dense_h_to_4h.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.76.self_attention.dense.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.76.self_attention.query_key_value.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.77.ln_attn.bias": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.77.ln_attn.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.77.ln_mlp.bias": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.77.ln_mlp.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.77.mlp.dense_4h_to_h.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.77.mlp.dense_h_to_4h.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.77.self_attention.dense.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.77.self_attention.query_key_value.weight": "pytorch_model-00039-of-00041.bin",
+    "transformer.h.78.ln_attn.bias": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.78.ln_attn.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.78.ln_mlp.bias": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.78.ln_mlp.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.78.mlp.dense_4h_to_h.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.78.mlp.dense_h_to_4h.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.78.self_attention.dense.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.78.self_attention.query_key_value.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.79.ln_attn.bias": "pytorch_model-00041-of-00041.bin",
+    "transformer.h.79.ln_attn.weight": "pytorch_model-00041-of-00041.bin",
+    "transformer.h.79.ln_mlp.bias": "pytorch_model-00041-of-00041.bin",
+    "transformer.h.79.ln_mlp.weight": "pytorch_model-00041-of-00041.bin",
+    "transformer.h.79.mlp.dense_4h_to_h.weight": "pytorch_model-00041-of-00041.bin",
+    "transformer.h.79.mlp.dense_h_to_4h.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.79.self_attention.dense.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.79.self_attention.query_key_value.weight": "pytorch_model-00040-of-00041.bin",
+    "transformer.h.8.ln_attn.bias": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.8.ln_attn.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.8.ln_mlp.bias": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.8.ln_mlp.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.8.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.8.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.8.self_attention.dense.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.8.self_attention.query_key_value.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.9.ln_attn.bias": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.9.ln_attn.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.9.ln_mlp.bias": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.9.ln_mlp.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.9.mlp.dense_4h_to_h.weight": "pytorch_model-00006-of-00041.bin",
+    "transformer.h.9.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.9.self_attention.dense.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.h.9.self_attention.query_key_value.weight": "pytorch_model-00005-of-00041.bin",
+    "transformer.ln_f.bias": "pytorch_model-00041-of-00041.bin",
+    "transformer.ln_f.weight": "pytorch_model-00041-of-00041.bin",
+    "transformer.word_embeddings.weight": "pytorch_model-00001-of-00041.bin"
+  }
+}

special_tokens_map.json

@@ -0,0 +1,16 @@
+{
+  "additional_special_tokens": [
+    ">>TITLE<<",
+    ">>ABSTRACT<<",
+    ">>INTRODUCTION<<",
+    ">>SUMMARY<<",
+    ">>COMMENT<<",
+    ">>ANSWER<<",
+    ">>QUESTION<<",
+    ">>DOMAIN<<",
+    ">>PREFIX<<",
+    ">>SUFFIX<<",
+    ">>MIDDLE<<"
+  ],
+  "eos_token": "<|endoftext|>"
+}

tokenizer_config.json

@@ -0,0 +1,7 @@
+{
+  "add_prefix_space": false,
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 2048,
+  "tokenizer_class": "PreTrainedTokenizerFast"
+}