Upload folder using huggingface_hub

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "openbuddy-f180-pre",
+  "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:fb58bc0c3db57ad40c8863978db0d1f9ba8b527938128f8a7d65560bf24aa763
+size 9213960603

pytorch_model-00002-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2f3a4433fa3046244dd569f607a4e42ac0fc3c23343d4653949f0c251c54efd
+size 8879584803

pytorch_model-00003-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78b87ca59bb3b4c37b979d8d6c3e02a63394725ee01f434ed3e745e363f95344
+size 8879584803

pytorch_model-00004-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f54f9d8c4b9f0f79cd22a3b7d81e19a32eaf2dfc79493cb4f47c80b1495a1c7
+size 8879584803

pytorch_model-00005-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd2e3f82cbea52dd924ee67c1d942731a08c2209e7acb87e5785b5918d789382
+size 8879584803

pytorch_model-00006-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20b13eeaa07640119125407e58e6a3b523e26c046a2db96d3d08ab4c154f06f0
+size 8879584803

pytorch_model-00007-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c4e485587d115014210c659629bcf5f265eebc2fe4b892f39ddda40404be4f3
+size 8879584803

pytorch_model-00008-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8788d79d9a7fbd7b4aaeb6f7515ccf76cf00b5d27538e0d9cefb9232bd7c7549
+size 8879584803

pytorch_model-00009-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d146b3945d57780ccd00b43f3e42fbc027cef740f76e42a7b829e60130965fb
+size 8879584803

pytorch_model-00010-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8fcb899da95eefb8bef1b22cb631fa68576186afe3f7aaab3b0aec6d8c3f8e37
+size 8879584803

pytorch_model-00011-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d2bfd05c51ad726fdb36f50370f84c8db93e617270b3567c1a9b572a5b0d783
+size 8879584803

pytorch_model-00012-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9b5645ceb11cbf97889182875df86eda19cd1024f8ea55afeeb1c305ef6a7fc
+size 8879584803

pytorch_model-00013-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bafac8115890343f6c83f3906f8fc104cafa2ff679298f032720a44d9e11fab
+size 8879584803

pytorch_model-00014-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a5519ca9a22d7ed5280379bc9083ffa601a4d84796a8f7c92f3ecff7c5eddc8
+size 8879584803

pytorch_model-00015-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef1ea037b0eb150caea35c06708304ba301d808a0d7d2d5a9687ae0343ff4191
+size 8879584803

pytorch_model-00016-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3014a2bb76503fe606859038f3f570cf10ba82a520a6119d5e94279d7f9ea6cf
+size 8879584803

pytorch_model-00017-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0046449836367837cc983695309ba915fa967b0771c8248e9ac60026801ea248
+size 8879584803

pytorch_model-00018-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77cb78098495c063cc8f069ece034f6745419f2f2ff7105c79abf1ada41b3a46
+size 8879584803

pytorch_model-00019-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43c9045645ebf73de69a0d08d59e1311f39b07e5075b2975778ea9e73bd337cc
+size 8879584803

pytorch_model-00020-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b59e351f03540843766fdded8893f7ee81c4a394da633ed67ae6922c22bdcc8e
+size 8879584803

pytorch_model-00021-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31c0e59721580afb59a193bff2dab247eb18ca74a937989eb9012c40bbc9f8d5
+size 8879584803

pytorch_model-00022-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:020cc025f21100c10af9d8132743d764cae793c5d29fa51ed72fd28053c97bb0
+size 8879584803

pytorch_model-00023-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78cb3bd1ee75daf4c5b81e23ae725ddc39638e693efc07d7f3ddb6e5ec9a47c4
+size 8879584803

pytorch_model-00024-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15668193d484ca1c1c1db884fbb4961383751ac7c433a4290d9ea5c361d502ad
+size 8879584803

pytorch_model-00025-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e1d08da46f2aff05fda1514bb1ee8eeaa04a4b6be7b23d9de60c1762e812d54
+size 8879584803

pytorch_model-00026-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c12565719d5f7d0b73f9487539bc136a287d187976bf775f639cedebd0069da
+size 8879584803

pytorch_model-00027-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0986b3b512e8cacde7b01f5098d68e62dd7f1f0a4d11d105c6eaf020a7c6671f
+size 8879584803

pytorch_model-00028-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cebfb89dde058c4b333885529797f6368d7d12d0ff8ab8d1dd647d9deb2a0e49
+size 8879584803

pytorch_model-00029-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7ead03d22a9a9c3f3b744dfc1b47480037bd4047846e6227cc1cf03b6df2b913
+size 8879584803

pytorch_model-00030-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6feb772db32c71c30fb620cd1e699455bff3e5da7051b5272012bbc8f89f5f4a
+size 8879584803

pytorch_model-00031-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8fa4c749ba303ba4289ab12ca41f371e370dc2d3633dd5b9c0603e61e7dc3af
+size 8879584803

pytorch_model-00032-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0efe6890d16829a1609f73970bcc5f3e90fd4b3cf3fa68466c2dd6138bae6f5
+size 8879584803

pytorch_model-00033-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59cf7287aa4dcf984ff4407e8b656c20f19b466f0a152fe76bc98a6711e7feeb
+size 8879584803

pytorch_model-00034-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0bfb3746d0a845a302f882a993559aa45e16069d637c2437c5c54c28e005a212
+size 8879584803

pytorch_model-00035-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47d17685dd6fa5df139c09e33a647c77c09b94b3c64d2418ccaa34942558af44
+size 8879584803

pytorch_model-00036-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7fa2c4c8779b1f59110cb51b64327ffed25cc2033ea0680331549493dcfcc53
+size 8879584803

pytorch_model-00037-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba84e9fbf6f033dccc51f45cacf3e08db1066a444bd83f20666a156a0dc7d3bf
+size 8879584803

pytorch_model-00038-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb0afc73224e316e2406df12c73a102cba272653eb7bb845c363534ea833cd1c
+size 8879584803

pytorch_model-00039-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b6a727b42491fb1bfbedc5c0462a1cea1bd3ed5ee91d4d03141eae3a22be9af
+size 8879584803

pytorch_model-00040-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da54038ff42b674717a3d16fee4de78ab1f288d3cfb44b0cb4a79734467875cd
+size 8879584803

pytorch_model-00041-of-00041.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f709575c55ef58bedfb1ddde703f16defe890537d8341981099a6efca19e1890
+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"
+}