Upload folder using huggingface_hub

checkpoint-342/config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "toanbku/oa-pythia-12b-rlhf-df",
+  "architectures": [
+    "GPTNeoXForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "classifier_dropout": 0.1,
+  "eos_token_id": 0,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 20480,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "gpt_neox",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 36,
+  "rope_scaling": null,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 0.25,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.31.0.dev0",
+  "use_cache": true,
+  "use_parallel_residual": true,
+  "vocab_size": 50288
+}

checkpoint-342/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "transformers_version": "4.31.0.dev0"
+}

checkpoint-342/global_step342/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55bdd1476345c2f04a842ffb43ac564728d467e0746c508ef61375fa392a97fd
+size 23684097854

checkpoint-342/global_step342/zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd9f1b055d7d4991d58c7066cb8c93a71c3e5217f9ff6707d55869ffdb342c46
+size 17762971806

checkpoint-342/global_step342/zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e990b6cff16b4612332a3a4cd8be00e5ee189d7edc3a11b9634946d166625096
+size 17762972766

checkpoint-342/global_step342/zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6d7859d2475aaea2d6b5b1056f9b1cef319b0b3856df1f948b428bf871287e8d
+size 17762972766

checkpoint-342/global_step342/zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5ec63440900ca51c1919e35f92345bfd52a53c2e803254d243369789f867215
+size 17762972446

checkpoint-342/global_step342/zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7a6c23d414f9cb60d13bf9e0f80a99d1baf375fdc8bf4639361d165d4ef6ef3
+size 17762972702

checkpoint-342/global_step342/zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56227481180a90df9fdbc784417e5b4058f1257f2358c50c865132c339429c74
+size 17762972830

checkpoint-342/global_step342/zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1eb87993b7e48d14f815055e40d35508c196304312ed96d470669d047614a82b
+size 17762972638

checkpoint-342/global_step342/zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00305e46139b956fc56b3dca838cb481fe07b38c89bb7ed9ca98fe81c47908bd
+size 17762971742

checkpoint-342/latest

@@ -0,0 +1 @@
+global_step342

checkpoint-342/pytorch_model-00001-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2dff00c45af3df3e650643894523894bdaf898f00a402f6b88345c3de17aef50
+size 9839916213

checkpoint-342/pytorch_model-00002-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2eeecbb8d8adaf51731db56963f6444b61f4df464b8c2b90c172f130f0ce2d20
+size 9648960597

checkpoint-342/pytorch_model-00003-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d9c11cc8272e808861f36c50e02fc766d5bef655773eb9035f5e75deb7a05e8
+size 9648982897

checkpoint-342/pytorch_model-00004-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cf69fc06344395b471cf37f74c1828b6e2d4db9013ffcd726f43be5ae581bf8
+size 9963617221

checkpoint-342/pytorch_model-00005-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcffc30bb91af54f63b9540ccc0f669eacda4d48327d015c0bdb353c16f7c587
+size 8266593363

checkpoint-342/pytorch_model.bin.index.json

@@ -0,0 +1,479 @@
+{
+  "metadata": {
+    "total_size": 47367907584
+  },
+  "weight_map": {
+    "embed_out.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.embed_in.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.final_layer_norm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.final_layer_norm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.0.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.10.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.14.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.14.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.15.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.2.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.20.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.22.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.22.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.22.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.22.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.23.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.3.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.30.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.31.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.4.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.8.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin"
+  }
+}

checkpoint-342/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7627a4adf319d8b3560071c74ece8be9b66a0992b0472b04ced3306a7a3e51a1
+size 21623

checkpoint-342/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:633074840bb9fdb66ade01ef38767085a06a413a9fdc9cef4f3ae65fa90eead0
+size 21623

checkpoint-342/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5e95e1661b4b4eeaa074607d4189adb1774e66e7bd55da196323d897919a95a
+size 21623

checkpoint-342/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56238fa338ce9ef74d74fc101050f9d039163f6560c23a5e82249053377f2b9a
+size 21623

checkpoint-342/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e40a0070b17b7832a9830020e66cff56f4570d0449d5ff9b85e0bc82fe22e513
+size 21623

checkpoint-342/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d5c0f5081eebeb8196e230a6a40c33e915e2c4f8fb504923a0f35702ed012de
+size 21623

checkpoint-342/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:117dd360c36456343804c467246c92bd880afd134fea1b85648b60401a37ff2f
+size 21623

checkpoint-342/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61c395ed5b2a3aa084d3353a1644d8d5c58603e2c4ca3b6eb41ca368333e44a5
+size 21623

checkpoint-342/special_tokens_map.json

@@ -0,0 +1,14 @@
+{
+  "additional_special_tokens": [
+    "<|prefix_begin|>",
+    "<|prefix_end|>",
+    "<|prompter|>",
+    "<|assistant|>",
+    "<|system|>"
+  ],
+  "bos_token": "<|endoftext|>",
+  "eos_token": "<|endoftext|>",
+  "pad_token": "<|padding|>",
+  "sep_token": "<|endoftext|>",
+  "unk_token": "<|endoftext|>"
+}

checkpoint-342/tokenizer_config.json

@@ -0,0 +1,9 @@
+{
+  "add_prefix_space": false,
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 1000000000000000019884624838656,
+  "tokenizer_class": "GPTNeoXTokenizer",
+  "unk_token": "<|endoftext|>"
+}

checkpoint-342/trainer_state.json

@@ -0,0 +1,526 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "global_step": 342,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.917808219178082e-06,
+      "loss": 0.0322,
+      "step": 10
+    },
+    {
+      "epoch": 0.06,
+      "eval_oasst_export_accuracy": 0.49051699074786653,
+      "eval_oasst_export_loss": 3.484375,
+      "eval_oasst_export_runtime": 6.0372,
+      "eval_oasst_export_samples_per_second": 181.21,
+      "eval_oasst_export_steps_per_second": 2.982,
+      "step": 10
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 5.800391389432485e-06,
+      "loss": 0.0246,
+      "step": 20
+    },
+    {
+      "epoch": 0.12,
+      "eval_oasst_export_accuracy": 0.489678997143981,
+      "eval_oasst_export_loss": 3.779296875,
+      "eval_oasst_export_runtime": 6.056,
+      "eval_oasst_export_samples_per_second": 180.647,
+      "eval_oasst_export_steps_per_second": 2.972,
+      "step": 20
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 5.682974559686888e-06,
+      "loss": 0.0203,
+      "step": 30
+    },
+    {
+      "epoch": 0.18,
+      "eval_oasst_export_accuracy": 0.4914062900826022,
+      "eval_oasst_export_loss": 3.734375,
+      "eval_oasst_export_runtime": 6.0349,
+      "eval_oasst_export_samples_per_second": 181.279,
+      "eval_oasst_export_steps_per_second": 2.983,
+      "step": 30
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 5.565557729941292e-06,
+      "loss": 0.0187,
+      "step": 40
+    },
+    {
+      "epoch": 0.23,
+      "eval_oasst_export_accuracy": 0.4879859080259265,
+      "eval_oasst_export_loss": 3.7109375,
+      "eval_oasst_export_runtime": 6.0607,
+      "eval_oasst_export_samples_per_second": 180.508,
+      "eval_oasst_export_steps_per_second": 2.97,
+      "step": 40
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 5.448140900195694e-06,
+      "loss": 0.0162,
+      "step": 50
+    },
+    {
+      "epoch": 0.29,
+      "eval_oasst_export_accuracy": 0.4880030099362099,
+      "eval_oasst_export_loss": 3.783203125,
+      "eval_oasst_export_runtime": 6.0304,
+      "eval_oasst_export_samples_per_second": 181.415,
+      "eval_oasst_export_steps_per_second": 2.985,
+      "step": 50
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 5.330724070450098e-06,
+      "loss": 0.0159,
+      "step": 60
+    },
+    {
+      "epoch": 0.35,
+      "eval_oasst_export_accuracy": 0.486292818907872,
+      "eval_oasst_export_loss": 3.7890625,
+      "eval_oasst_export_runtime": 6.0367,
+      "eval_oasst_export_samples_per_second": 181.224,
+      "eval_oasst_export_steps_per_second": 2.982,
+      "step": 60
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 5.213307240704501e-06,
+      "loss": 0.0172,
+      "step": 70
+    },
+    {
+      "epoch": 0.41,
+      "eval_oasst_export_accuracy": 0.48716501633232434,
+      "eval_oasst_export_loss": 3.806640625,
+      "eval_oasst_export_runtime": 6.0405,
+      "eval_oasst_export_samples_per_second": 181.11,
+      "eval_oasst_export_steps_per_second": 2.98,
+      "step": 70
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 5.095890410958904e-06,
+      "loss": 0.0157,
+      "step": 80
+    },
+    {
+      "epoch": 0.47,
+      "eval_oasst_export_accuracy": 0.48454842405896736,
+      "eval_oasst_export_loss": 3.791015625,
+      "eval_oasst_export_runtime": 6.0411,
+      "eval_oasst_export_samples_per_second": 181.092,
+      "eval_oasst_export_steps_per_second": 2.98,
+      "step": 80
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.978473581213307e-06,
+      "loss": 0.0163,
+      "step": 90
+    },
+    {
+      "epoch": 0.53,
+      "eval_oasst_export_accuracy": 0.49138918817231886,
+      "eval_oasst_export_loss": 3.822265625,
+      "eval_oasst_export_runtime": 5.6229,
+      "eval_oasst_export_samples_per_second": 194.561,
+      "eval_oasst_export_steps_per_second": 3.201,
+      "step": 90
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.86105675146771e-06,
+      "loss": 0.0139,
+      "step": 100
+    },
+    {
+      "epoch": 0.58,
+      "eval_oasst_export_accuracy": 0.4956817676534469,
+      "eval_oasst_export_loss": 3.826171875,
+      "eval_oasst_export_runtime": 6.0223,
+      "eval_oasst_export_samples_per_second": 181.659,
+      "eval_oasst_export_steps_per_second": 2.989,
+      "step": 100
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.7436399217221134e-06,
+      "loss": 0.0151,
+      "step": 110
+    },
+    {
+      "epoch": 0.64,
+      "eval_oasst_export_accuracy": 0.4845655259692508,
+      "eval_oasst_export_loss": 3.904296875,
+      "eval_oasst_export_runtime": 6.0257,
+      "eval_oasst_export_samples_per_second": 181.554,
+      "eval_oasst_export_steps_per_second": 2.987,
+      "step": 110
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.626223091976516e-06,
+      "loss": 0.0139,
+      "step": 120
+    },
+    {
+      "epoch": 0.7,
+      "eval_oasst_export_accuracy": 0.49049988883758316,
+      "eval_oasst_export_loss": 3.958984375,
+      "eval_oasst_export_runtime": 6.0136,
+      "eval_oasst_export_samples_per_second": 181.92,
+      "eval_oasst_export_steps_per_second": 2.993,
+      "step": 120
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.50880626223092e-06,
+      "loss": 0.0133,
+      "step": 130
+    },
+    {
+      "epoch": 0.76,
+      "eval_oasst_export_accuracy": 0.4803413541292562,
+      "eval_oasst_export_loss": 3.845703125,
+      "eval_oasst_export_runtime": 6.0221,
+      "eval_oasst_export_samples_per_second": 181.663,
+      "eval_oasst_export_steps_per_second": 2.989,
+      "step": 130
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.3913894324853226e-06,
+      "loss": 0.0142,
+      "step": 140
+    },
+    {
+      "epoch": 0.82,
+      "eval_oasst_export_accuracy": 0.4777076599456159,
+      "eval_oasst_export_loss": 3.828125,
+      "eval_oasst_export_runtime": 6.0368,
+      "eval_oasst_export_samples_per_second": 181.223,
+      "eval_oasst_export_steps_per_second": 2.982,
+      "step": 140
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.273972602739726e-06,
+      "loss": 0.0141,
+      "step": 150
+    },
+    {
+      "epoch": 0.88,
+      "eval_oasst_export_accuracy": 0.47088399774254786,
+      "eval_oasst_export_loss": 3.998046875,
+      "eval_oasst_export_runtime": 6.0502,
+      "eval_oasst_export_samples_per_second": 180.82,
+      "eval_oasst_export_steps_per_second": 2.975,
+      "step": 150
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.156555772994129e-06,
+      "loss": 0.0157,
+      "step": 160
+    },
+    {
+      "epoch": 0.94,
+      "eval_oasst_export_accuracy": 0.4726796983223026,
+      "eval_oasst_export_loss": 3.8828125,
+      "eval_oasst_export_runtime": 5.995,
+      "eval_oasst_export_samples_per_second": 182.487,
+      "eval_oasst_export_steps_per_second": 3.003,
+      "step": 160
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0391389432485325e-06,
+      "loss": 0.0139,
+      "step": 170
+    },
+    {
+      "epoch": 0.99,
+      "eval_oasst_export_accuracy": 0.4744240931712072,
+      "eval_oasst_export_loss": 3.9609375,
+      "eval_oasst_export_runtime": 6.0696,
+      "eval_oasst_export_samples_per_second": 180.244,
+      "eval_oasst_export_steps_per_second": 2.966,
+      "step": 170
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.921722113502935e-06,
+      "loss": 0.0125,
+      "step": 180
+    },
+    {
+      "epoch": 1.05,
+      "eval_oasst_export_accuracy": 0.48210285088844423,
+      "eval_oasst_export_loss": 4.03125,
+      "eval_oasst_export_runtime": 5.616,
+      "eval_oasst_export_samples_per_second": 194.8,
+      "eval_oasst_export_steps_per_second": 3.205,
+      "step": 180
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.8043052837573385e-06,
+      "loss": 0.017,
+      "step": 190
+    },
+    {
+      "epoch": 1.11,
+      "eval_oasst_export_accuracy": 0.4897303028748311,
+      "eval_oasst_export_loss": 3.970703125,
+      "eval_oasst_export_runtime": 6.03,
+      "eval_oasst_export_samples_per_second": 181.426,
+      "eval_oasst_export_steps_per_second": 2.985,
+      "step": 190
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.6868884540117416e-06,
+      "loss": 0.0137,
+      "step": 200
+    },
+    {
+      "epoch": 1.17,
+      "eval_oasst_export_accuracy": 0.4794862586150873,
+      "eval_oasst_export_loss": 3.986328125,
+      "eval_oasst_export_runtime": 6.0192,
+      "eval_oasst_export_samples_per_second": 181.751,
+      "eval_oasst_export_steps_per_second": 2.99,
+      "step": 200
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.569471624266145e-06,
+      "loss": 0.0133,
+      "step": 210
+    },
+    {
+      "epoch": 1.23,
+      "eval_oasst_export_accuracy": 0.47693807398286386,
+      "eval_oasst_export_loss": 4.140625,
+      "eval_oasst_export_runtime": 6.0148,
+      "eval_oasst_export_samples_per_second": 181.886,
+      "eval_oasst_export_steps_per_second": 2.993,
+      "step": 210
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.452054794520548e-06,
+      "loss": 0.0133,
+      "step": 220
+    },
+    {
+      "epoch": 1.29,
+      "eval_oasst_export_accuracy": 0.48713081251175755,
+      "eval_oasst_export_loss": 4.1484375,
+      "eval_oasst_export_runtime": 6.0347,
+      "eval_oasst_export_samples_per_second": 181.285,
+      "eval_oasst_export_steps_per_second": 2.983,
+      "step": 220
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.334637964774951e-06,
+      "loss": 0.0154,
+      "step": 230
+    },
+    {
+      "epoch": 1.35,
+      "eval_oasst_export_accuracy": 0.4752620867750928,
+      "eval_oasst_export_loss": 4.09375,
+      "eval_oasst_export_runtime": 5.6373,
+      "eval_oasst_export_samples_per_second": 194.065,
+      "eval_oasst_export_steps_per_second": 3.193,
+      "step": 230
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.217221135029354e-06,
+      "loss": 0.0139,
+      "step": 240
+    },
+    {
+      "epoch": 1.4,
+      "eval_oasst_export_accuracy": 0.4760658765584116,
+      "eval_oasst_export_loss": 4.09765625,
+      "eval_oasst_export_runtime": 6.055,
+      "eval_oasst_export_samples_per_second": 180.676,
+      "eval_oasst_export_steps_per_second": 2.973,
+      "step": 240
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.099804305283757e-06,
+      "loss": 0.0136,
+      "step": 250
+    },
+    {
+      "epoch": 1.46,
+      "eval_oasst_export_accuracy": 0.47102081302481486,
+      "eval_oasst_export_loss": 4.20703125,
+      "eval_oasst_export_runtime": 6.0408,
+      "eval_oasst_export_samples_per_second": 181.103,
+      "eval_oasst_export_steps_per_second": 2.98,
+      "step": 250
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.9823874755381603e-06,
+      "loss": 0.0129,
+      "step": 260
+    },
+    {
+      "epoch": 1.52,
+      "eval_oasst_export_accuracy": 0.4641458450908967,
+      "eval_oasst_export_loss": 4.2578125,
+      "eval_oasst_export_runtime": 6.0349,
+      "eval_oasst_export_samples_per_second": 181.279,
+      "eval_oasst_export_steps_per_second": 2.983,
+      "step": 260
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.8649706457925635e-06,
+      "loss": 0.0128,
+      "step": 270
+    },
+    {
+      "epoch": 1.58,
+      "eval_oasst_export_accuracy": 0.4649838386947822,
+      "eval_oasst_export_loss": 4.24609375,
+      "eval_oasst_export_runtime": 6.0355,
+      "eval_oasst_export_samples_per_second": 181.26,
+      "eval_oasst_export_steps_per_second": 2.982,
+      "step": 270
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.7475538160469666e-06,
+      "loss": 0.0133,
+      "step": 280
+    },
+    {
+      "epoch": 1.64,
+      "eval_oasst_export_accuracy": 0.46412874318061326,
+      "eval_oasst_export_loss": 4.2421875,
+      "eval_oasst_export_runtime": 6.0367,
+      "eval_oasst_export_samples_per_second": 181.224,
+      "eval_oasst_export_steps_per_second": 2.982,
+      "step": 280
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.63013698630137e-06,
+      "loss": 0.013,
+      "step": 290
+    },
+    {
+      "epoch": 1.7,
+      "eval_oasst_export_accuracy": 0.47182460280813365,
+      "eval_oasst_export_loss": 4.28125,
+      "eval_oasst_export_runtime": 6.0187,
+      "eval_oasst_export_samples_per_second": 181.768,
+      "eval_oasst_export_steps_per_second": 2.991,
+      "step": 290
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 2.512720156555773e-06,
+      "loss": 0.0133,
+      "step": 300
+    },
+    {
+      "epoch": 1.75,
+      "eval_oasst_export_accuracy": 0.47093530347339796,
+      "eval_oasst_export_loss": 4.30078125,
+      "eval_oasst_export_runtime": 6.0452,
+      "eval_oasst_export_samples_per_second": 180.969,
+      "eval_oasst_export_steps_per_second": 2.978,
+      "step": 300
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 2.395303326810176e-06,
+      "loss": 0.0117,
+      "step": 310
+    },
+    {
+      "epoch": 1.81,
+      "eval_oasst_export_accuracy": 0.4726454945017358,
+      "eval_oasst_export_loss": 4.41796875,
+      "eval_oasst_export_runtime": 6.0234,
+      "eval_oasst_export_samples_per_second": 181.624,
+      "eval_oasst_export_steps_per_second": 2.988,
+      "step": 310
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.2778864970645793e-06,
+      "loss": 0.0121,
+      "step": 320
+    },
+    {
+      "epoch": 1.87,
+      "eval_oasst_export_accuracy": 0.47096950729396475,
+      "eval_oasst_export_loss": 4.4375,
+      "eval_oasst_export_runtime": 6.0349,
+      "eval_oasst_export_samples_per_second": 181.279,
+      "eval_oasst_export_steps_per_second": 2.983,
+      "step": 320
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 2.1604696673189825e-06,
+      "loss": 0.0121,
+      "step": 330
+    },
+    {
+      "epoch": 1.93,
+      "eval_oasst_export_accuracy": 0.480375557949823,
+      "eval_oasst_export_loss": 4.421875,
+      "eval_oasst_export_runtime": 6.0242,
+      "eval_oasst_export_samples_per_second": 181.602,
+      "eval_oasst_export_steps_per_second": 2.988,
+      "step": 330
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.0430528375733853e-06,
+      "loss": 0.0128,
+      "step": 340
+    },
+    {
+      "epoch": 1.99,
+      "eval_oasst_export_accuracy": 0.4811964496434252,
+      "eval_oasst_export_loss": 4.40625,
+      "eval_oasst_export_runtime": 6.0198,
+      "eval_oasst_export_samples_per_second": 181.734,
+      "eval_oasst_export_steps_per_second": 2.99,
+      "step": 340
+    }
+  ],
+  "max_steps": 513,
+  "num_train_epochs": 3,
+  "total_flos": 2.4125538591477596e+18,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-342/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd9117c7e28390071fd5803c147bec189553c57f7a0a906d42ece025a148af8a
+size 5499

checkpoint-342/zero_to_fp32.py

@@ -0,0 +1,578 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dicts.append(torch.load(f, map_location=device))
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel()
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, args.output_file)

checkpoint-513/config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "toanbku/oa-pythia-12b-rlhf-df",
+  "architectures": [
+    "GPTNeoXForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 0,
+  "classifier_dropout": 0.1,
+  "eos_token_id": 0,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 5120,
+  "initializer_range": 0.02,
+  "intermediate_size": 20480,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "gpt_neox",
+  "num_attention_heads": 40,
+  "num_hidden_layers": 36,
+  "rope_scaling": null,
+  "rotary_emb_base": 10000,
+  "rotary_pct": 0.25,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.31.0.dev0",
+  "use_cache": true,
+  "use_parallel_residual": true,
+  "vocab_size": 50288
+}

checkpoint-513/generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 0,
+  "eos_token_id": 0,
+  "transformers_version": "4.31.0.dev0"
+}

checkpoint-513/global_step513/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c02057b4f5996f47e9848f3dcbf6797470c46585425925cac2db63632bc5c95
+size 23684097854

checkpoint-513/global_step513/zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeb503443e65f4375f0a53ca2a0e8703908df1ef073f4fd28cd77b80d8fa9095
+size 17762971806

checkpoint-513/global_step513/zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ff5885de281d970b62bdca900b50dc44441d290603425711ec7dd875b4da1f4
+size 17762972766

checkpoint-513/global_step513/zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89e7bbec5d200cced122d44295f3fe68f2955aaa5eb47d6b30a50f06dbfb1408
+size 17762972766

checkpoint-513/global_step513/zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbe34267d2118c3ce832e5d2bad89a5902605f1d490a7ab46d85da58e0a14b2c
+size 17762972446

checkpoint-513/global_step513/zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff79336950b7bcd0aa717255abb0c860f011370c8a0f2c6401a4552e22bfe56b
+size 17762972702

checkpoint-513/global_step513/zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf8a1f1ee7a0c1a79070922e5a0eff973a2967c8feb214977787ba917b55b570
+size 17762972830

checkpoint-513/global_step513/zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf35798822d734dd1d9bd316a5bc1729dd83ca376b004a4c83370aeedcd49c29
+size 17762972638

checkpoint-513/global_step513/zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b66df4d3fcc0ea5509fcbb599e8bbb1bc801f0f791d34d3426430a1b03285399
+size 17762971742

checkpoint-513/latest

@@ -0,0 +1 @@
+global_step513

checkpoint-513/pytorch_model-00001-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b44de8a5f6e3a987cb844e673119b4b6c43c634add97be72ce93cd381c8dcd5
+size 9839916213

checkpoint-513/pytorch_model-00002-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48c1f9303879a4b159ff12d7be944f3f2ae4cd5309ce403e4c5a1d41a2f0defd
+size 9648960597

checkpoint-513/pytorch_model-00003-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:744472a05950d41eeef5ff9b60243798e9781a9234a5c4efdd57da8b52869ff1
+size 9648982897

checkpoint-513/pytorch_model-00004-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44afbd599aea7be2c1b2cb81507ef3504d671bc23f195847cb1f3977a509eaa6
+size 9963617221

checkpoint-513/pytorch_model-00005-of-00005.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbaf88c791efc698a58eaf2235bcaf24b8c7868255f2fca4a043b7cb393aec73
+size 8266593363

checkpoint-513/pytorch_model.bin.index.json

@@ -0,0 +1,479 @@
+{
+  "metadata": {
+    "total_size": 47367907584
+  },
+  "weight_map": {
+    "embed_out.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.embed_in.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.final_layer_norm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.final_layer_norm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.0.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.0.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.1.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.10.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.10.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.11.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.12.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.13.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.14.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.14.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.14.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.15.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.15.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.16.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.17.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.18.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.19.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.2.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.2.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.20.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.20.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.mlp.dense_4h_to_h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.mlp.dense_4h_to_h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.mlp.dense_h_to_4h.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.mlp.dense_h_to_4h.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.21.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.dense.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.dense.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.query_key_value.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.query_key_value.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.attention.rotary_emb.inv_freq": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.input_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.input_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.22.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.22.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.22.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.22.post_attention_layernorm.bias": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.22.post_attention_layernorm.weight": "pytorch_model-00003-of-00005.bin",
+    "gpt_neox.layers.23.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.23.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.24.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.25.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.26.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.27.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.28.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.dense.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.dense.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.mlp.dense_4h_to_h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.mlp.dense_4h_to_h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.mlp.dense_h_to_4h.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.mlp.dense_h_to_4h.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.29.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.3.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.3.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.30.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.attention.query_key_value.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.attention.query_key_value.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.attention.rotary_emb.inv_freq": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.input_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.input_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.30.post_attention_layernorm.bias": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.30.post_attention_layernorm.weight": "pytorch_model-00004-of-00005.bin",
+    "gpt_neox.layers.31.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.31.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.32.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.33.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.34.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.dense.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.dense.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.query_key_value.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.query_key_value.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.attention.rotary_emb.inv_freq": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.input_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.input_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.mlp.dense_4h_to_h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.mlp.dense_4h_to_h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.mlp.dense_h_to_4h.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.mlp.dense_h_to_4h.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.post_attention_layernorm.bias": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.35.post_attention_layernorm.weight": "pytorch_model-00005-of-00005.bin",
+    "gpt_neox.layers.4.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.4.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.5.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.dense.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.dense.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.query_key_value.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.query_key_value.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.mlp.dense_4h_to_h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.mlp.dense_4h_to_h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.mlp.dense_h_to_4h.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.mlp.dense_h_to_4h.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.6.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.attention.rotary_emb.inv_freq": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.input_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.input_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.7.post_attention_layernorm.bias": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.7.post_attention_layernorm.weight": "pytorch_model-00001-of-00005.bin",
+    "gpt_neox.layers.8.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.8.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.dense.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.dense.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.query_key_value.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.query_key_value.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.attention.rotary_emb.inv_freq": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.input_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.input_layernorm.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.mlp.dense_4h_to_h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.mlp.dense_4h_to_h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.mlp.dense_h_to_4h.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.mlp.dense_h_to_4h.weight": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.post_attention_layernorm.bias": "pytorch_model-00002-of-00005.bin",
+    "gpt_neox.layers.9.post_attention_layernorm.weight": "pytorch_model-00002-of-00005.bin"
+  }
+}