Upload folder using huggingface_hub

README.md

@@ -0,0 +1,153 @@
+---
+license: apache-2.0
+base_model: mistralai/Mistral-7B-v0.1
+tags:
+- generated_from_trainer
+model-index:
+- name: mistral-7B-MedText-epochs-3-lr-0002
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl)
+<details><summary>See axolotl config</summary>
+
+axolotl version: `0.4.0`
+```yaml
+base_model: mistralai/Mistral-7B-v0.1
+model_type: MistralForCausalLM
+tokenizer_type: LlamaTokenizer
+is_mistral_derived_model: true
+
+load_in_8bit: false
+load_in_4bit: false
+strict: false
+
+datasets:
+  - path: utrgvseniorproject/medtext
+    type: completion
+dataset_prepared_path: last_run_prepared
+val_set_size: 0.05
+output_dir: ./mistral-7B-MedText-epochs-3-lr-0002
+
+sequence_len: 4096
+sample_packing: true
+pad_to_sequence_len: true
+
+adapter:
+lora_model_dir:
+lora_r:
+lora_alpha:
+lora_dropout:
+lora_target_linear:
+lora_fan_in_fan_out:
+
+wandb_project: mistral-7B-MedText
+wandb_entity: utrgvmedai
+wandb_watch:
+wandb_name: mistral-7B-MedText-epochs-3-lr-0002
+wandb_log_model:
+
+gradient_accumulation_steps: 1
+micro_batch_size: 1
+num_epochs: 3
+optimizer: adamw_bnb_8bit
+lr_scheduler: cosine
+learning_rate: 0.0002
+
+train_on_inputs: true
+group_by_length: false
+bf16: auto
+fp16:
+tf32: false
+
+gradient_checkpointing: true
+early_stopping_patience:
+#resume_from_checkpoint: true
+local_rank:
+logging_steps: 1
+xformers_attention:
+flash_attention: true
+flash_attn_cross_entropy: false
+flash_attn_rms_norm: true
+flash_attn_fuse_qkv: false
+flash_attn_fuse_mlp: true
+
+warmup_steps: 100
+evals_per_epoch: 4
+eval_table_size:
+eval_sample_packing: False
+saves_per_epoch: 1
+debug:
+deepspeed: /home/josegomez15/axolotl/deepspeed_configs/zero2.json
+weight_decay: 0.1
+fsdp:
+fsdp_config:
+special_tokens:
+
+```
+
+</details><br>
+
+# mistral-7B-MedText-epochs-3-lr-0002
+
+This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 7.2922
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 0.0002
+- train_batch_size: 1
+- eval_batch_size: 1
+- seed: 42
+- distributed_type: multi-GPU
+- num_devices: 4
+- total_train_batch_size: 4
+- total_eval_batch_size: 4
+- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
+- lr_scheduler_type: cosine
+- lr_scheduler_warmup_steps: 100
+- num_epochs: 3
+
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss |
+|:-------------:|:-----:|:----:|:---------------:|
+| 1.3985        | 0.01  | 1    | 1.5677          |
+| 1.4776        | 0.25  | 22   | 1.8568          |
+| 10.1246       | 0.51  | 44   | 8.7590          |
+| 8.1284        | 0.76  | 66   | 8.0049          |
+| 7.3967        | 1.01  | 88   | 7.4614          |
+| 7.2567        | 1.23  | 110  | 7.2993          |
+| 7.3329        | 1.48  | 132  | 7.3749          |
+| 7.0671        | 1.74  | 154  | 7.3365          |
+| 7.4786        | 1.99  | 176  | 7.3194          |
+| 7.3548        | 2.22  | 198  | 7.3092          |
+| 7.1782        | 2.47  | 220  | 7.2964          |
+| 7.2729        | 2.72  | 242  | 7.2922          |
+
+
+### Framework versions
+
+- Transformers 4.38.0
+- Pytorch 2.0.1+cu117
+- Datasets 2.17.0
+- Tokenizers 0.15.0

checkpoint-174/config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "mistralai/Mistral-7B-v0.1",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.38.0",
+  "use_cache": false,
+  "vocab_size": 32000
+}

checkpoint-174/generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": 2,
+  "transformers_version": "4.38.0"
+}

checkpoint-174/global_step174/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03d6725388def79326de49d69d8c6b94170b534dd31f6a030b7a2cbc2d601c94
+size 21725204471

checkpoint-174/global_step174/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9849abfddf8ba736da45584934d8c464b6e43f73e0dab39ae34b2420c7668144
+size 21725205111

checkpoint-174/global_step174/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:013acab701355320fcbd84283d45579ded93da9287d18209908e3313d973915f
+size 21725205111

checkpoint-174/global_step174/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:107c21efe5b9cb5ee1ea8f2e3e646ca7683d8b5b2ac2d693c020a4c9e0435282
+size 21725204535

checkpoint-174/global_step174/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ec4f15bbbe81efd524cf417a820f9118838524077c76a700f3567a758312b8a
+size 14483551747

checkpoint-174/latest

@@ -0,0 +1 @@
+global_step174

checkpoint-174/model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3de77ae8eae772f363eeb8943d44cccb53d98891ac8469da8e9f370750a5d155
+size 4943162336

checkpoint-174/model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45f0c04c0135c3b2c9d387fe2334075a0a1b40b1bdc5d5126a65b86e2085fbf1
+size 4999819336

checkpoint-174/model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a574bf1ffd420f13cf2f6d14a3d032ba8c604cb738a0266920a361ac5418eb00
+size 4540516344

checkpoint-174/model.safetensors.index.json

@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 14483464192
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00003-of-00003.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.norm.weight": "model-00003-of-00003.safetensors"
+  }
+}

checkpoint-174/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4aa38064359163a436b682a76922629be0d47715cc93798f53c30584786df380
+size 17655

checkpoint-174/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:386882d88441fb270dac4c0a02c6b12d9125a96b7fa06d4456448d9d1b9da975
+size 17655

checkpoint-174/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c2119f0f2530fd325c172a94b131c42f068b4b61340e0b39e5188ebdebae961
+size 17655

checkpoint-174/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b265d61c7eae8db516afc3f2f413284853452f9b41a53ad49381bf25c87ed7a9
+size 17655

checkpoint-174/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ba61684322025b96b5a644148a042abe6fbab4ac762f0d5bcae8f648817ea4e
+size 627

checkpoint-174/trainer_state.json

@@ -0,0 +1,1303 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9655172413793105,
+  "eval_steps": 22,
+  "global_step": 174,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "grad_norm": 19.59183260405101,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 1.3985,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 1.5676738023757935,
+      "eval_runtime": 3.9227,
+      "eval_samples_per_second": 4.844,
+      "eval_steps_per_second": 1.275,
+      "step": 1
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 25.857640141956836,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.6073,
+      "step": 2
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 19.53747244629775,
+      "learning_rate": 6e-06,
+      "loss": 1.5784,
+      "step": 3
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 32.1611390841105,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 1.4924,
+      "step": 4
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 25.68491589990429,
+      "learning_rate": 1e-05,
+      "loss": 1.5215,
+      "step": 5
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 21.800722806654328,
+      "learning_rate": 1.2e-05,
+      "loss": 1.5508,
+      "step": 6
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 17.19812526688047,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 1.205,
+      "step": 7
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 33.24510948526866,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 1.6376,
+      "step": 8
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 14.986129451187,
+      "learning_rate": 1.8e-05,
+      "loss": 1.454,
+      "step": 9
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 17.517194747344856,
+      "learning_rate": 2e-05,
+      "loss": 1.6559,
+      "step": 10
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 18.22606604955052,
+      "learning_rate": 2.2000000000000003e-05,
+      "loss": 1.5816,
+      "step": 11
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 26.577328122239592,
+      "learning_rate": 2.4e-05,
+      "loss": 1.7884,
+      "step": 12
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 17.563463503652706,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 1.4405,
+      "step": 13
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 19.98739837372538,
+      "learning_rate": 2.8000000000000003e-05,
+      "loss": 1.7512,
+      "step": 14
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 16.25364292099236,
+      "learning_rate": 3e-05,
+      "loss": 1.7037,
+      "step": 15
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 22.328347783615648,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 1.5476,
+      "step": 16
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 20.71265405702363,
+      "learning_rate": 3.4000000000000007e-05,
+      "loss": 1.8108,
+      "step": 17
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 22.425069186679085,
+      "learning_rate": 3.6e-05,
+      "loss": 1.6751,
+      "step": 18
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 18.646553829771403,
+      "learning_rate": 3.8e-05,
+      "loss": 1.7696,
+      "step": 19
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 16.74801305116847,
+      "learning_rate": 4e-05,
+      "loss": 1.7699,
+      "step": 20
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 30.99364381661585,
+      "learning_rate": 4.2e-05,
+      "loss": 2.1827,
+      "step": 21
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 19.761245037191404,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 1.4776,
+      "step": 22
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 1.856759786605835,
+      "eval_runtime": 1.4309,
+      "eval_samples_per_second": 13.279,
+      "eval_steps_per_second": 3.494,
+      "step": 22
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 27.821861354666428,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 1.8677,
+      "step": 23
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 22.80354998875929,
+      "learning_rate": 4.8e-05,
+      "loss": 1.7424,
+      "step": 24
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 19.920321630615994,
+      "learning_rate": 5e-05,
+      "loss": 1.769,
+      "step": 25
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 23.82621534355921,
+      "learning_rate": 5.2000000000000004e-05,
+      "loss": 1.8615,
+      "step": 26
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 27.63826215875828,
+      "learning_rate": 5.4000000000000005e-05,
+      "loss": 1.7164,
+      "step": 27
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 30.810269474286084,
+      "learning_rate": 5.6000000000000006e-05,
+      "loss": 1.8922,
+      "step": 28
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 23.279443617002265,
+      "learning_rate": 5.8e-05,
+      "loss": 1.8507,
+      "step": 29
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 18.091756285883704,
+      "learning_rate": 6e-05,
+      "loss": 1.6537,
+      "step": 30
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 23.292568819411724,
+      "learning_rate": 6.2e-05,
+      "loss": 1.9125,
+      "step": 31
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 52.21913805211888,
+      "learning_rate": 6.400000000000001e-05,
+      "loss": 1.8456,
+      "step": 32
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 30.771918199992527,
+      "learning_rate": 6.6e-05,
+      "loss": 1.9398,
+      "step": 33
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 40.19880102087157,
+      "learning_rate": 6.800000000000001e-05,
+      "loss": 2.3763,
+      "step": 34
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 28.33779955799431,
+      "learning_rate": 7e-05,
+      "loss": 2.0815,
+      "step": 35
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 54.379783823327905,
+      "learning_rate": 7.2e-05,
+      "loss": 2.1032,
+      "step": 36
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 662.6041427579516,
+      "learning_rate": 7.4e-05,
+      "loss": 3.5917,
+      "step": 37
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 4011.501090614335,
+      "learning_rate": 7.6e-05,
+      "loss": 7.3933,
+      "step": 38
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 753.5816146908044,
+      "learning_rate": 7.800000000000001e-05,
+      "loss": 15.8371,
+      "step": 39
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 291.86609997137384,
+      "learning_rate": 8e-05,
+      "loss": 9.0249,
+      "step": 40
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 922.0326051718562,
+      "learning_rate": 8.2e-05,
+      "loss": 9.8922,
+      "step": 41
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 115.68135873083658,
+      "learning_rate": 8.4e-05,
+      "loss": 8.2719,
+      "step": 42
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 693.1702757980898,
+      "learning_rate": 8.6e-05,
+      "loss": 24.1699,
+      "step": 43
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 219.78828591396768,
+      "learning_rate": 8.800000000000001e-05,
+      "loss": 10.1246,
+      "step": 44
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 8.758976936340332,
+      "eval_runtime": 1.3819,
+      "eval_samples_per_second": 13.749,
+      "eval_steps_per_second": 3.618,
+      "step": 44
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 101.32256273283853,
+      "learning_rate": 9e-05,
+      "loss": 9.1756,
+      "step": 45
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 11.958039992364316,
+      "learning_rate": 9.200000000000001e-05,
+      "loss": 7.7273,
+      "step": 46
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 3232.8188937829473,
+      "learning_rate": 9.4e-05,
+      "loss": 8.7486,
+      "step": 47
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 34.97640980952823,
+      "learning_rate": 9.6e-05,
+      "loss": 8.0394,
+      "step": 48
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 28.785233378323184,
+      "learning_rate": 9.8e-05,
+      "loss": 7.6241,
+      "step": 49
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 734.1733446536996,
+      "learning_rate": 0.0001,
+      "loss": 7.7813,
+      "step": 50
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 59.66290191764372,
+      "learning_rate": 0.00010200000000000001,
+      "loss": 7.6734,
+      "step": 51
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 18.23606863385719,
+      "learning_rate": 0.00010400000000000001,
+      "loss": 7.8783,
+      "step": 52
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 15.152500341579357,
+      "learning_rate": 0.00010600000000000002,
+      "loss": 7.6016,
+      "step": 53
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 25.1062172290986,
+      "learning_rate": 0.00010800000000000001,
+      "loss": 7.7206,
+      "step": 54
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 30.75760910048317,
+      "learning_rate": 0.00011000000000000002,
+      "loss": 7.7526,
+      "step": 55
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 18.354297835205724,
+      "learning_rate": 0.00011200000000000001,
+      "loss": 7.5551,
+      "step": 56
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 32.48061320811231,
+      "learning_rate": 0.00011399999999999999,
+      "loss": 7.6922,
+      "step": 57
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 26.624723764338736,
+      "learning_rate": 0.000116,
+      "loss": 7.5791,
+      "step": 58
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 13.92226298912218,
+      "learning_rate": 0.000118,
+      "loss": 7.3692,
+      "step": 59
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 15.428105634893914,
+      "learning_rate": 0.00012,
+      "loss": 7.564,
+      "step": 60
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 23.02235056392551,
+      "learning_rate": 0.000122,
+      "loss": 7.3948,
+      "step": 61
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 16.470374126805776,
+      "learning_rate": 0.000124,
+      "loss": 7.4646,
+      "step": 62
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 19.574553557429,
+      "learning_rate": 0.000126,
+      "loss": 7.3902,
+      "step": 63
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 25.549182808604662,
+      "learning_rate": 0.00012800000000000002,
+      "loss": 7.5461,
+      "step": 64
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 618.1147749407063,
+      "learning_rate": 0.00013000000000000002,
+      "loss": 8.157,
+      "step": 65
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 66.0610660913446,
+      "learning_rate": 0.000132,
+      "loss": 8.1284,
+      "step": 66
+    },
+    {
+      "epoch": 0.76,
+      "eval_loss": 8.004860877990723,
+      "eval_runtime": 1.3748,
+      "eval_samples_per_second": 13.82,
+      "eval_steps_per_second": 3.637,
+      "step": 66
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 41.635960690924975,
+      "learning_rate": 0.000134,
+      "loss": 8.0602,
+      "step": 67
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 16.679869546723175,
+      "learning_rate": 0.00013600000000000003,
+      "loss": 7.8268,
+      "step": 68
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 33.70142265056291,
+      "learning_rate": 0.000138,
+      "loss": 8.1206,
+      "step": 69
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 43.29315748881759,
+      "learning_rate": 0.00014,
+      "loss": 7.7882,
+      "step": 70
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 28.3799834727588,
+      "learning_rate": 0.000142,
+      "loss": 7.5454,
+      "step": 71
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 30.911781055251613,
+      "learning_rate": 0.000144,
+      "loss": 7.7616,
+      "step": 72
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 56.741836296758365,
+      "learning_rate": 0.000146,
+      "loss": 8.1145,
+      "step": 73
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 13.967249574234195,
+      "learning_rate": 0.000148,
+      "loss": 7.6758,
+      "step": 74
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 42.28071588262047,
+      "learning_rate": 0.00015000000000000001,
+      "loss": 7.5507,
+      "step": 75
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 27.40577229308579,
+      "learning_rate": 0.000152,
+      "loss": 7.6852,
+      "step": 76
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 21.422615778584905,
+      "learning_rate": 0.000154,
+      "loss": 7.6071,
+      "step": 77
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 28.22569647336005,
+      "learning_rate": 0.00015600000000000002,
+      "loss": 7.5556,
+      "step": 78
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 20.075502929468385,
+      "learning_rate": 0.00015800000000000002,
+      "loss": 7.529,
+      "step": 79
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 14.896626626967638,
+      "learning_rate": 0.00016,
+      "loss": 7.5352,
+      "step": 80
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 16.98565517398293,
+      "learning_rate": 0.000162,
+      "loss": 7.5734,
+      "step": 81
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 17.76803489316823,
+      "learning_rate": 0.000164,
+      "loss": 7.591,
+      "step": 82
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 15.601674972992182,
+      "learning_rate": 0.000166,
+      "loss": 7.4251,
+      "step": 83
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 21.36580777081032,
+      "learning_rate": 0.000168,
+      "loss": 7.728,
+      "step": 84
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 19.73380900316605,
+      "learning_rate": 0.00017,
+      "loss": 7.3548,
+      "step": 85
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 16.40313028374367,
+      "learning_rate": 0.000172,
+      "loss": 7.3652,
+      "step": 86
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 11.846265862818466,
+      "learning_rate": 0.000174,
+      "loss": 7.4363,
+      "step": 87
+    },
+    {
+      "epoch": 1.01,
+      "grad_norm": 11.82176337757903,
+      "learning_rate": 0.00017600000000000002,
+      "loss": 7.3967,
+      "step": 88
+    },
+    {
+      "epoch": 1.01,
+      "eval_loss": 7.461406707763672,
+      "eval_runtime": 1.3867,
+      "eval_samples_per_second": 13.702,
+      "eval_steps_per_second": 3.606,
+      "step": 88
+    },
+    {
+      "epoch": 1.02,
+      "grad_norm": 14.25806078550941,
+      "learning_rate": 0.00017800000000000002,
+      "loss": 7.3814,
+      "step": 89
+    },
+    {
+      "epoch": 1.03,
+      "grad_norm": 10.077089336320658,
+      "learning_rate": 0.00018,
+      "loss": 7.2599,
+      "step": 90
+    },
+    {
+      "epoch": 1.01,
+      "grad_norm": 15.159363102480643,
+      "learning_rate": 0.000182,
+      "loss": 7.5719,
+      "step": 91
+    },
+    {
+      "epoch": 1.02,
+      "grad_norm": 14.678683877513562,
+      "learning_rate": 0.00018400000000000003,
+      "loss": 7.4794,
+      "step": 92
+    },
+    {
+      "epoch": 1.03,
+      "grad_norm": 16.396549401486602,
+      "learning_rate": 0.00018600000000000002,
+      "loss": 7.6211,
+      "step": 93
+    },
+    {
+      "epoch": 1.05,
+      "grad_norm": 15.078668930441678,
+      "learning_rate": 0.000188,
+      "loss": 7.3625,
+      "step": 94
+    },
+    {
+      "epoch": 1.06,
+      "grad_norm": 13.546080322892589,
+      "learning_rate": 0.00019,
+      "loss": 7.3016,
+      "step": 95
+    },
+    {
+      "epoch": 1.07,
+      "grad_norm": 18.231949750348118,
+      "learning_rate": 0.000192,
+      "loss": 7.386,
+      "step": 96
+    },
+    {
+      "epoch": 1.08,
+      "grad_norm": 12.444086372431334,
+      "learning_rate": 0.000194,
+      "loss": 7.5534,
+      "step": 97
+    },
+    {
+      "epoch": 1.09,
+      "grad_norm": 9.994459143045797,
+      "learning_rate": 0.000196,
+      "loss": 7.4254,
+      "step": 98
+    },
+    {
+      "epoch": 1.1,
+      "grad_norm": 14.889247706503676,
+      "learning_rate": 0.00019800000000000002,
+      "loss": 7.3938,
+      "step": 99
+    },
+    {
+      "epoch": 1.11,
+      "grad_norm": 19.194847464605004,
+      "learning_rate": 0.0002,
+      "loss": 7.3875,
+      "step": 100
+    },
+    {
+      "epoch": 1.13,
+      "grad_norm": 19.234057674068023,
+      "learning_rate": 0.00019998096274980728,
+      "loss": 7.348,
+      "step": 101
+    },
+    {
+      "epoch": 1.14,
+      "grad_norm": 10.248320279074052,
+      "learning_rate": 0.000199923858247567,
+      "loss": 7.2365,
+      "step": 102
+    },
+    {
+      "epoch": 1.15,
+      "grad_norm": 15.579201434411983,
+      "learning_rate": 0.00019982870823553308,
+      "loss": 7.2351,
+      "step": 103
+    },
+    {
+      "epoch": 1.16,
+      "grad_norm": 10.951732160571428,
+      "learning_rate": 0.00019969554894159723,
+      "loss": 7.4413,
+      "step": 104
+    },
+    {
+      "epoch": 1.17,
+      "grad_norm": 12.034336556190219,
+      "learning_rate": 0.00019952443106549533,
+      "loss": 7.4256,
+      "step": 105
+    },
+    {
+      "epoch": 1.18,
+      "grad_norm": 11.555864069154051,
+      "learning_rate": 0.00019931541975950378,
+      "loss": 7.0947,
+      "step": 106
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 8.079125584036722,
+      "learning_rate": 0.00019906859460363307,
+      "loss": 7.3727,
+      "step": 107
+    },
+    {
+      "epoch": 1.21,
+      "grad_norm": 10.901301706963714,
+      "learning_rate": 0.00019878404957532814,
+      "loss": 7.2419,
+      "step": 108
+    },
+    {
+      "epoch": 1.22,
+      "grad_norm": 9.771076993744128,
+      "learning_rate": 0.0001984618930136869,
+      "loss": 7.2896,
+      "step": 109
+    },
+    {
+      "epoch": 1.23,
+      "grad_norm": 11.395240298688034,
+      "learning_rate": 0.00019810224757821064,
+      "loss": 7.2567,
+      "step": 110
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 7.299312591552734,
+      "eval_runtime": 1.381,
+      "eval_samples_per_second": 13.758,
+      "eval_steps_per_second": 3.621,
+      "step": 110
+    },
+    {
+      "epoch": 1.24,
+      "grad_norm": 11.839939762683946,
+      "learning_rate": 0.00019770525020210204,
+      "loss": 7.3145,
+      "step": 111
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 13.454044820108352,
+      "learning_rate": 0.0001972710520401287,
+      "loss": 7.3279,
+      "step": 112
+    },
+    {
+      "epoch": 1.26,
+      "grad_norm": 10.349510041217494,
+      "learning_rate": 0.0001967998184110713,
+      "loss": 7.0995,
+      "step": 113
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 11.299695940906268,
+      "learning_rate": 0.00019629172873477995,
+      "loss": 7.1545,
+      "step": 114
+    },
+    {
+      "epoch": 1.29,
+      "grad_norm": 10.300523126150598,
+      "learning_rate": 0.00019574697646386027,
+      "loss": 7.2619,
+      "step": 115
+    },
+    {
+      "epoch": 1.3,
+      "grad_norm": 12.280692750358627,
+      "learning_rate": 0.0001951657690100178,
+      "loss": 7.4703,
+      "step": 116
+    },
+    {
+      "epoch": 1.31,
+      "grad_norm": 11.308344462386193,
+      "learning_rate": 0.0001945483276650868,
+      "loss": 7.2927,
+      "step": 117
+    },
+    {
+      "epoch": 1.32,
+      "grad_norm": 9.46182612207882,
+      "learning_rate": 0.0001938948875167745,
+      "loss": 7.1097,
+      "step": 118
+    },
+    {
+      "epoch": 1.33,
+      "grad_norm": 8.495214573764272,
+      "learning_rate": 0.00019320569735915271,
+      "loss": 7.2199,
+      "step": 119
+    },
+    {
+      "epoch": 1.34,
+      "grad_norm": 10.888318204738358,
+      "learning_rate": 0.00019248101959793066,
+      "loss": 7.4214,
+      "step": 120
+    },
+    {
+      "epoch": 1.36,
+      "grad_norm": 15.655909856272627,
+      "learning_rate": 0.00019172113015054532,
+      "loss": 7.4141,
+      "step": 121
+    },
+    {
+      "epoch": 1.37,
+      "grad_norm": 14.348570302527095,
+      "learning_rate": 0.00019092631834110723,
+      "loss": 7.2034,
+      "step": 122
+    },
+    {
+      "epoch": 1.38,
+      "grad_norm": 11.216620878113865,
+      "learning_rate": 0.0001900968867902419,
+      "loss": 7.1925,
+      "step": 123
+    },
+    {
+      "epoch": 1.39,
+      "grad_norm": 15.214529842718143,
+      "learning_rate": 0.00018923315129986835,
+      "loss": 7.204,
+      "step": 124
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 24.933551829531577,
+      "learning_rate": 0.00018833544073295917,
+      "loss": 7.5231,
+      "step": 125
+    },
+    {
+      "epoch": 1.41,
+      "grad_norm": 12.291980045440198,
+      "learning_rate": 0.00018740409688832764,
+      "loss": 7.2323,
+      "step": 126
+    },
+    {
+      "epoch": 1.43,
+      "grad_norm": 19.503396545387695,
+      "learning_rate": 0.00018643947437048944,
+      "loss": 7.3214,
+      "step": 127
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 13.982404823681808,
+      "learning_rate": 0.00018544194045464886,
+      "loss": 7.4008,
+      "step": 128
+    },
+    {
+      "epoch": 1.45,
+      "grad_norm": 13.775851829485163,
+      "learning_rate": 0.00018441187494686053,
+      "loss": 7.2854,
+      "step": 129
+    },
+    {
+      "epoch": 1.46,
+      "grad_norm": 16.44073199273777,
+      "learning_rate": 0.0001833496700394202,
+      "loss": 7.4218,
+      "step": 130
+    },
+    {
+      "epoch": 1.47,
+      "grad_norm": 21.66328419423237,
+      "learning_rate": 0.00018225573016153945,
+      "loss": 7.3589,
+      "step": 131
+    },
+    {
+      "epoch": 1.48,
+      "grad_norm": 14.027278482513156,
+      "learning_rate": 0.00018113047182536127,
+      "loss": 7.3329,
+      "step": 132
+    },
+    {
+      "epoch": 1.48,
+      "eval_loss": 7.374873161315918,
+      "eval_runtime": 1.3815,
+      "eval_samples_per_second": 13.753,
+      "eval_steps_per_second": 3.619,
+      "step": 132
+    },
+    {
+      "epoch": 1.49,
+      "grad_norm": 10.89412121909016,
+      "learning_rate": 0.00017997432346737524,
+      "loss": 7.3277,
+      "step": 133
+    },
+    {
+      "epoch": 1.51,
+      "grad_norm": 18.182727305484867,
+      "learning_rate": 0.00017878772528529232,
+      "loss": 7.3505,
+      "step": 134
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 18.738647279231998,
+      "learning_rate": 0.000177571129070442,
+      "loss": 7.4395,
+      "step": 135
+    },
+    {
+      "epoch": 1.53,
+      "grad_norm": 8.544634593385059,
+      "learning_rate": 0.00017632499803575474,
+      "loss": 7.3132,
+      "step": 136
+    },
+    {
+      "epoch": 1.54,
+      "grad_norm": 23.787945119488832,
+      "learning_rate": 0.00017504980663939613,
+      "loss": 7.3635,
+      "step": 137
+    },
+    {
+      "epoch": 1.55,
+      "grad_norm": 15.988837162794528,
+      "learning_rate": 0.00017374604040411935,
+      "loss": 7.298,
+      "step": 138
+    },
+    {
+      "epoch": 1.56,
+      "grad_norm": 8.594141503866256,
+      "learning_rate": 0.00017241419573240462,
+      "loss": 7.2263,
+      "step": 139
+    },
+    {
+      "epoch": 1.57,
+      "grad_norm": 10.834111660116264,
+      "learning_rate": 0.00017105477971745666,
+      "loss": 7.2234,
+      "step": 140
+    },
+    {
+      "epoch": 1.59,
+      "grad_norm": 20.898663112099943,
+      "learning_rate": 0.00016966830995013133,
+      "loss": 7.4312,
+      "step": 141
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 9.882602517532073,
+      "learning_rate": 0.00016825531432186543,
+      "loss": 7.3389,
+      "step": 142
+    },
+    {
+      "epoch": 1.61,
+      "grad_norm": 7.435249821686469,
+      "learning_rate": 0.00016681633082368498,
+      "loss": 7.2015,
+      "step": 143
+    },
+    {
+      "epoch": 1.62,
+      "grad_norm": 13.215068331486576,
+      "learning_rate": 0.0001653519073413675,
+      "loss": 7.1425,
+      "step": 144
+    },
+    {
+      "epoch": 1.63,
+      "grad_norm": 11.754841557793458,
+      "learning_rate": 0.00016386260144683745,
+      "loss": 7.3101,
+      "step": 145
+    },
+    {
+      "epoch": 1.64,
+      "grad_norm": 7.674786980372276,
+      "learning_rate": 0.00016234898018587337,
+      "loss": 7.1703,
+      "step": 146
+    },
+    {
+      "epoch": 1.66,
+      "grad_norm": 8.30389279917059,
+      "learning_rate": 0.00016081161986220807,
+      "loss": 7.2516,
+      "step": 147
+    },
+    {
+      "epoch": 1.67,
+      "grad_norm": 14.325137493545546,
+      "learning_rate": 0.00015925110581810394,
+      "loss": 7.3326,
+      "step": 148
+    },
+    {
+      "epoch": 1.68,
+      "grad_norm": 9.943124774719692,
+      "learning_rate": 0.00015766803221148673,
+      "loss": 7.5501,
+      "step": 149
+    },
+    {
+      "epoch": 1.69,
+      "grad_norm": 11.710815673357324,
+      "learning_rate": 0.00015606300178972287,
+      "loss": 7.3344,
+      "step": 150
+    },
+    {
+      "epoch": 1.7,
+      "grad_norm": 8.287278539768234,
+      "learning_rate": 0.00015443662566012645,
+      "loss": 7.2426,
+      "step": 151
+    },
+    {
+      "epoch": 1.71,
+      "grad_norm": 8.045767995001754,
+      "learning_rate": 0.00015278952305728324,
+      "loss": 7.2319,
+      "step": 152
+    },
+    {
+      "epoch": 1.72,
+      "grad_norm": 16.213767636990735,
+      "learning_rate": 0.00015112232110728015,
+      "loss": 7.2226,
+      "step": 153
+    },
+    {
+      "epoch": 1.74,
+      "grad_norm": 10.115986520778595,
+      "learning_rate": 0.00014943565458893,
+      "loss": 7.0671,
+      "step": 154
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 7.336472511291504,
+      "eval_runtime": 1.3763,
+      "eval_samples_per_second": 13.806,
+      "eval_steps_per_second": 3.633,
+      "step": 154
+    },
+    {
+      "epoch": 1.75,
+      "grad_norm": 7.370162670560762,
+      "learning_rate": 0.00014773016569208283,
+      "loss": 7.2456,
+      "step": 155
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 14.854609514494987,
+      "learning_rate": 0.00014600650377311522,
+      "loss": 7.0447,
+      "step": 156
+    },
+    {
+      "epoch": 1.77,
+      "grad_norm": 9.984605765623492,
+      "learning_rate": 0.0001442653251076912,
+      "loss": 7.1948,
+      "step": 157
+    },
+    {
+      "epoch": 1.78,
+      "grad_norm": 11.053561117527067,
+      "learning_rate": 0.00014250729264088843,
+      "loss": 7.1485,
+      "step": 158
+    },
+    {
+      "epoch": 1.79,
+      "grad_norm": 11.604464862126479,
+      "learning_rate": 0.00014073307573478526,
+      "loss": 7.4198,
+      "step": 159
+    },
+    {
+      "epoch": 1.8,
+      "grad_norm": 8.62267592810566,
+      "learning_rate": 0.00013894334991360448,
+      "loss": 7.3045,
+      "step": 160
+    },
+    {
+      "epoch": 1.82,
+      "grad_norm": 10.961255591651343,
+      "learning_rate": 0.00013713879660651068,
+      "loss": 6.9983,
+      "step": 161
+    },
+    {
+      "epoch": 1.83,
+      "grad_norm": 11.12104950103157,
+      "learning_rate": 0.0001353201028881598,
+      "loss": 7.1046,
+      "step": 162
+    },
+    {
+      "epoch": 1.84,
+      "grad_norm": 8.807345567053673,
+      "learning_rate": 0.00013348796121709862,
+      "loss": 7.3378,
+      "step": 163
+    },
+    {
+      "epoch": 1.85,
+      "grad_norm": 10.666053356976384,
+      "learning_rate": 0.00013164306917211476,
+      "loss": 7.0008,
+      "step": 164
+    },
+    {
+      "epoch": 1.86,
+      "grad_norm": 7.835224761952429,
+      "learning_rate": 0.000129786129186637,
+      "loss": 6.9881,
+      "step": 165
+    },
+    {
+      "epoch": 1.87,
+      "grad_norm": 19.4062633667929,
+      "learning_rate": 0.00012791784828128724,
+      "loss": 7.2579,
+      "step": 166
+    },
+    {
+      "epoch": 1.89,
+      "grad_norm": 10.113263144537674,
+      "learning_rate": 0.00012603893779468604,
+      "loss": 7.3091,
+      "step": 167
+    },
+    {
+      "epoch": 1.9,
+      "grad_norm": 10.870503515462726,
+      "learning_rate": 0.0001241501131126138,
+      "loss": 7.4527,
+      "step": 168
+    },
+    {
+      "epoch": 1.91,
+      "grad_norm": 8.533887294828766,
+      "learning_rate": 0.00012225209339563145,
+      "loss": 7.2855,
+      "step": 169
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 13.486746604884923,
+      "learning_rate": 0.0001203456013052634,
+      "loss": 7.2705,
+      "step": 170
+    },
+    {
+      "epoch": 1.93,
+      "grad_norm": 10.69251762526038,
+      "learning_rate": 0.00011843136272884794,
+      "loss": 7.0932,
+      "step": 171
+    },
+    {
+      "epoch": 1.94,
+      "grad_norm": 9.124923914761991,
+      "learning_rate": 0.00011651010650315923,
+      "loss": 7.3754,
+      "step": 172
+    },
+    {
+      "epoch": 1.95,
+      "grad_norm": 13.481473855966252,
+      "learning_rate": 0.00011458256413690633,
+      "loss": 7.3104,
+      "step": 173
+    },
+    {
+      "epoch": 1.97,
+      "grad_norm": 13.187202052506544,
+      "learning_rate": 0.00011264946953221496,
+      "loss": 7.3614,
+      "step": 174
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 261,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 87,
+  "total_flos": 1.2162710143343002e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-174/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b50b6bb5c490e8ef22b44507d2914da60c8e4f8c0f5671692c3b6485d1e90608
+size 6395

checkpoint-174/zero_to_fp32.py

@@ -0,0 +1,592 @@
+#!/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_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    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 _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+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() if _has_callable(shape, 'numel') else math.prod(shape)
+            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("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    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, tag=args.tag)

checkpoint-261/config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "mistralai/Mistral-7B-v0.1",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.38.0",
+  "use_cache": false,
+  "vocab_size": 32000
+}

checkpoint-261/generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": 2,
+  "transformers_version": "4.38.0"
+}

checkpoint-261/global_step261/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fb9e894ca467d0c73041d1e16de366fb1601d4b7f663b96f9dee73cdcd36217c
+size 21725204471

checkpoint-261/global_step261/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a9099d294f6558914ce9c10932240bb346ee76d393d1029f814cfda3ac23c26
+size 21725205111

checkpoint-261/global_step261/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d0da12e540f5d78b442356be0e626094299e094fe03205d373dc868379a92aa
+size 21725205111

checkpoint-261/global_step261/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc93f64204a586e12cb28e6a2cdbf812997662a01611a7719b67d4639c8cb628
+size 21725204535

checkpoint-261/global_step261/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0779635d4d5861f77003df69765a486d96cf572c41481e59430cce5866f30f22
+size 14483551747

checkpoint-261/latest

@@ -0,0 +1 @@
+global_step261

checkpoint-261/model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31c532f5ad94151c66aece624fdd79259c5d5f80540419098a7ed105e64a18a1
+size 4943162336

checkpoint-261/model-00002-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d76b471758354a6480eee81cb699f14ef71295c9d75ef0cefb9b197e04d1e82
+size 4999819336

checkpoint-261/model-00003-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f0b233e5eda8c7f368f67ddcfd233c19e82631b65ebaeadcd31deb162db3080
+size 4540516344

checkpoint-261/model.safetensors.index.json

@@ -0,0 +1,298 @@
+{
+  "metadata": {
+    "total_size": 14483464192
+  },
+  "weight_map": {
+    "lm_head.weight": "model-00003-of-00003.safetensors",
+    "model.embed_tokens.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.0.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.1.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.10.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.10.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.11.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.11.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.12.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.13.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.14.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.15.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.16.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.17.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.18.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.19.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.2.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.2.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.20.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.20.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.input_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.down_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.gate_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.mlp.up_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.post_attention_layernorm.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.21.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.22.self_attn.k_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.o_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.q_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.22.self_attn.v_proj.weight": "model-00002-of-00003.safetensors",
+    "model.layers.23.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.23.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.24.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.25.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.26.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.27.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.28.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.29.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.3.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.3.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.30.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.30.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.input_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.down_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.gate_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.mlp.up_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.post_attention_layernorm.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.k_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.o_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.q_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.31.self_attn.v_proj.weight": "model-00003-of-00003.safetensors",
+    "model.layers.4.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.4.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.5.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.6.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.7.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.8.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.input_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.down_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.gate_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.mlp.up_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.post_attention_layernorm.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.k_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.o_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.q_proj.weight": "model-00001-of-00003.safetensors",
+    "model.layers.9.self_attn.v_proj.weight": "model-00001-of-00003.safetensors",
+    "model.norm.weight": "model-00003-of-00003.safetensors"
+  }
+}

checkpoint-261/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:926b70f6be2914cad151dafd0e3fa08f8277adb10199e8a7f2ff8aa53afc3223
+size 17655

checkpoint-261/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0dcd6b2a15400db4e6e93af7cb8ba6e4e7e71f455c0122d4e3d19a0fef82b0a
+size 17655

checkpoint-261/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:110606a79dab49d13dae27afec5afb7c2980f1440e4f9e7e66001b14e016dd7f
+size 17655

checkpoint-261/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9cba41cf1ae1502a93c8ba0d7ee52d2096158a0ee2d07bf8a606129bb127516a
+size 17655

checkpoint-261/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67b1fbfe77e71f3535424c43de07cc0ffaacb05756dac4b0d55425ed3a2df41f
+size 627

checkpoint-261/trainer_state.json

@@ -0,0 +1,1944 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.942528735632184,
+  "eval_steps": 22,
+  "global_step": 261,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "grad_norm": 19.59183260405101,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 1.3985,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "eval_loss": 1.5676738023757935,
+      "eval_runtime": 3.9227,
+      "eval_samples_per_second": 4.844,
+      "eval_steps_per_second": 1.275,
+      "step": 1
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 25.857640141956836,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 1.6073,
+      "step": 2
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 19.53747244629775,
+      "learning_rate": 6e-06,
+      "loss": 1.5784,
+      "step": 3
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 32.1611390841105,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 1.4924,
+      "step": 4
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 25.68491589990429,
+      "learning_rate": 1e-05,
+      "loss": 1.5215,
+      "step": 5
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 21.800722806654328,
+      "learning_rate": 1.2e-05,
+      "loss": 1.5508,
+      "step": 6
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 17.19812526688047,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 1.205,
+      "step": 7
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 33.24510948526866,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 1.6376,
+      "step": 8
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 14.986129451187,
+      "learning_rate": 1.8e-05,
+      "loss": 1.454,
+      "step": 9
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 17.517194747344856,
+      "learning_rate": 2e-05,
+      "loss": 1.6559,
+      "step": 10
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 18.22606604955052,
+      "learning_rate": 2.2000000000000003e-05,
+      "loss": 1.5816,
+      "step": 11
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 26.577328122239592,
+      "learning_rate": 2.4e-05,
+      "loss": 1.7884,
+      "step": 12
+    },
+    {
+      "epoch": 0.15,
+      "grad_norm": 17.563463503652706,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 1.4405,
+      "step": 13
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 19.98739837372538,
+      "learning_rate": 2.8000000000000003e-05,
+      "loss": 1.7512,
+      "step": 14
+    },
+    {
+      "epoch": 0.17,
+      "grad_norm": 16.25364292099236,
+      "learning_rate": 3e-05,
+      "loss": 1.7037,
+      "step": 15
+    },
+    {
+      "epoch": 0.18,
+      "grad_norm": 22.328347783615648,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 1.5476,
+      "step": 16
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 20.71265405702363,
+      "learning_rate": 3.4000000000000007e-05,
+      "loss": 1.8108,
+      "step": 17
+    },
+    {
+      "epoch": 0.21,
+      "grad_norm": 22.425069186679085,
+      "learning_rate": 3.6e-05,
+      "loss": 1.6751,
+      "step": 18
+    },
+    {
+      "epoch": 0.22,
+      "grad_norm": 18.646553829771403,
+      "learning_rate": 3.8e-05,
+      "loss": 1.7696,
+      "step": 19
+    },
+    {
+      "epoch": 0.23,
+      "grad_norm": 16.74801305116847,
+      "learning_rate": 4e-05,
+      "loss": 1.7699,
+      "step": 20
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 30.99364381661585,
+      "learning_rate": 4.2e-05,
+      "loss": 2.1827,
+      "step": 21
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 19.761245037191404,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 1.4776,
+      "step": 22
+    },
+    {
+      "epoch": 0.25,
+      "eval_loss": 1.856759786605835,
+      "eval_runtime": 1.4309,
+      "eval_samples_per_second": 13.279,
+      "eval_steps_per_second": 3.494,
+      "step": 22
+    },
+    {
+      "epoch": 0.26,
+      "grad_norm": 27.821861354666428,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 1.8677,
+      "step": 23
+    },
+    {
+      "epoch": 0.28,
+      "grad_norm": 22.80354998875929,
+      "learning_rate": 4.8e-05,
+      "loss": 1.7424,
+      "step": 24
+    },
+    {
+      "epoch": 0.29,
+      "grad_norm": 19.920321630615994,
+      "learning_rate": 5e-05,
+      "loss": 1.769,
+      "step": 25
+    },
+    {
+      "epoch": 0.3,
+      "grad_norm": 23.82621534355921,
+      "learning_rate": 5.2000000000000004e-05,
+      "loss": 1.8615,
+      "step": 26
+    },
+    {
+      "epoch": 0.31,
+      "grad_norm": 27.63826215875828,
+      "learning_rate": 5.4000000000000005e-05,
+      "loss": 1.7164,
+      "step": 27
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 30.810269474286084,
+      "learning_rate": 5.6000000000000006e-05,
+      "loss": 1.8922,
+      "step": 28
+    },
+    {
+      "epoch": 0.33,
+      "grad_norm": 23.279443617002265,
+      "learning_rate": 5.8e-05,
+      "loss": 1.8507,
+      "step": 29
+    },
+    {
+      "epoch": 0.34,
+      "grad_norm": 18.091756285883704,
+      "learning_rate": 6e-05,
+      "loss": 1.6537,
+      "step": 30
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 23.292568819411724,
+      "learning_rate": 6.2e-05,
+      "loss": 1.9125,
+      "step": 31
+    },
+    {
+      "epoch": 0.37,
+      "grad_norm": 52.21913805211888,
+      "learning_rate": 6.400000000000001e-05,
+      "loss": 1.8456,
+      "step": 32
+    },
+    {
+      "epoch": 0.38,
+      "grad_norm": 30.771918199992527,
+      "learning_rate": 6.6e-05,
+      "loss": 1.9398,
+      "step": 33
+    },
+    {
+      "epoch": 0.39,
+      "grad_norm": 40.19880102087157,
+      "learning_rate": 6.800000000000001e-05,
+      "loss": 2.3763,
+      "step": 34
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 28.33779955799431,
+      "learning_rate": 7e-05,
+      "loss": 2.0815,
+      "step": 35
+    },
+    {
+      "epoch": 0.41,
+      "grad_norm": 54.379783823327905,
+      "learning_rate": 7.2e-05,
+      "loss": 2.1032,
+      "step": 36
+    },
+    {
+      "epoch": 0.43,
+      "grad_norm": 662.6041427579516,
+      "learning_rate": 7.4e-05,
+      "loss": 3.5917,
+      "step": 37
+    },
+    {
+      "epoch": 0.44,
+      "grad_norm": 4011.501090614335,
+      "learning_rate": 7.6e-05,
+      "loss": 7.3933,
+      "step": 38
+    },
+    {
+      "epoch": 0.45,
+      "grad_norm": 753.5816146908044,
+      "learning_rate": 7.800000000000001e-05,
+      "loss": 15.8371,
+      "step": 39
+    },
+    {
+      "epoch": 0.46,
+      "grad_norm": 291.86609997137384,
+      "learning_rate": 8e-05,
+      "loss": 9.0249,
+      "step": 40
+    },
+    {
+      "epoch": 0.47,
+      "grad_norm": 922.0326051718562,
+      "learning_rate": 8.2e-05,
+      "loss": 9.8922,
+      "step": 41
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 115.68135873083658,
+      "learning_rate": 8.4e-05,
+      "loss": 8.2719,
+      "step": 42
+    },
+    {
+      "epoch": 0.49,
+      "grad_norm": 693.1702757980898,
+      "learning_rate": 8.6e-05,
+      "loss": 24.1699,
+      "step": 43
+    },
+    {
+      "epoch": 0.51,
+      "grad_norm": 219.78828591396768,
+      "learning_rate": 8.800000000000001e-05,
+      "loss": 10.1246,
+      "step": 44
+    },
+    {
+      "epoch": 0.51,
+      "eval_loss": 8.758976936340332,
+      "eval_runtime": 1.3819,
+      "eval_samples_per_second": 13.749,
+      "eval_steps_per_second": 3.618,
+      "step": 44
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 101.32256273283853,
+      "learning_rate": 9e-05,
+      "loss": 9.1756,
+      "step": 45
+    },
+    {
+      "epoch": 0.53,
+      "grad_norm": 11.958039992364316,
+      "learning_rate": 9.200000000000001e-05,
+      "loss": 7.7273,
+      "step": 46
+    },
+    {
+      "epoch": 0.54,
+      "grad_norm": 3232.8188937829473,
+      "learning_rate": 9.4e-05,
+      "loss": 8.7486,
+      "step": 47
+    },
+    {
+      "epoch": 0.55,
+      "grad_norm": 34.97640980952823,
+      "learning_rate": 9.6e-05,
+      "loss": 8.0394,
+      "step": 48
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 28.785233378323184,
+      "learning_rate": 9.8e-05,
+      "loss": 7.6241,
+      "step": 49
+    },
+    {
+      "epoch": 0.57,
+      "grad_norm": 734.1733446536996,
+      "learning_rate": 0.0001,
+      "loss": 7.7813,
+      "step": 50
+    },
+    {
+      "epoch": 0.59,
+      "grad_norm": 59.66290191764372,
+      "learning_rate": 0.00010200000000000001,
+      "loss": 7.6734,
+      "step": 51
+    },
+    {
+      "epoch": 0.6,
+      "grad_norm": 18.23606863385719,
+      "learning_rate": 0.00010400000000000001,
+      "loss": 7.8783,
+      "step": 52
+    },
+    {
+      "epoch": 0.61,
+      "grad_norm": 15.152500341579357,
+      "learning_rate": 0.00010600000000000002,
+      "loss": 7.6016,
+      "step": 53
+    },
+    {
+      "epoch": 0.62,
+      "grad_norm": 25.1062172290986,
+      "learning_rate": 0.00010800000000000001,
+      "loss": 7.7206,
+      "step": 54
+    },
+    {
+      "epoch": 0.63,
+      "grad_norm": 30.75760910048317,
+      "learning_rate": 0.00011000000000000002,
+      "loss": 7.7526,
+      "step": 55
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 18.354297835205724,
+      "learning_rate": 0.00011200000000000001,
+      "loss": 7.5551,
+      "step": 56
+    },
+    {
+      "epoch": 0.66,
+      "grad_norm": 32.48061320811231,
+      "learning_rate": 0.00011399999999999999,
+      "loss": 7.6922,
+      "step": 57
+    },
+    {
+      "epoch": 0.67,
+      "grad_norm": 26.624723764338736,
+      "learning_rate": 0.000116,
+      "loss": 7.5791,
+      "step": 58
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 13.92226298912218,
+      "learning_rate": 0.000118,
+      "loss": 7.3692,
+      "step": 59
+    },
+    {
+      "epoch": 0.69,
+      "grad_norm": 15.428105634893914,
+      "learning_rate": 0.00012,
+      "loss": 7.564,
+      "step": 60
+    },
+    {
+      "epoch": 0.7,
+      "grad_norm": 23.02235056392551,
+      "learning_rate": 0.000122,
+      "loss": 7.3948,
+      "step": 61
+    },
+    {
+      "epoch": 0.71,
+      "grad_norm": 16.470374126805776,
+      "learning_rate": 0.000124,
+      "loss": 7.4646,
+      "step": 62
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 19.574553557429,
+      "learning_rate": 0.000126,
+      "loss": 7.3902,
+      "step": 63
+    },
+    {
+      "epoch": 0.74,
+      "grad_norm": 25.549182808604662,
+      "learning_rate": 0.00012800000000000002,
+      "loss": 7.5461,
+      "step": 64
+    },
+    {
+      "epoch": 0.75,
+      "grad_norm": 618.1147749407063,
+      "learning_rate": 0.00013000000000000002,
+      "loss": 8.157,
+      "step": 65
+    },
+    {
+      "epoch": 0.76,
+      "grad_norm": 66.0610660913446,
+      "learning_rate": 0.000132,
+      "loss": 8.1284,
+      "step": 66
+    },
+    {
+      "epoch": 0.76,
+      "eval_loss": 8.004860877990723,
+      "eval_runtime": 1.3748,
+      "eval_samples_per_second": 13.82,
+      "eval_steps_per_second": 3.637,
+      "step": 66
+    },
+    {
+      "epoch": 0.77,
+      "grad_norm": 41.635960690924975,
+      "learning_rate": 0.000134,
+      "loss": 8.0602,
+      "step": 67
+    },
+    {
+      "epoch": 0.78,
+      "grad_norm": 16.679869546723175,
+      "learning_rate": 0.00013600000000000003,
+      "loss": 7.8268,
+      "step": 68
+    },
+    {
+      "epoch": 0.79,
+      "grad_norm": 33.70142265056291,
+      "learning_rate": 0.000138,
+      "loss": 8.1206,
+      "step": 69
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 43.29315748881759,
+      "learning_rate": 0.00014,
+      "loss": 7.7882,
+      "step": 70
+    },
+    {
+      "epoch": 0.82,
+      "grad_norm": 28.3799834727588,
+      "learning_rate": 0.000142,
+      "loss": 7.5454,
+      "step": 71
+    },
+    {
+      "epoch": 0.83,
+      "grad_norm": 30.911781055251613,
+      "learning_rate": 0.000144,
+      "loss": 7.7616,
+      "step": 72
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 56.741836296758365,
+      "learning_rate": 0.000146,
+      "loss": 8.1145,
+      "step": 73
+    },
+    {
+      "epoch": 0.85,
+      "grad_norm": 13.967249574234195,
+      "learning_rate": 0.000148,
+      "loss": 7.6758,
+      "step": 74
+    },
+    {
+      "epoch": 0.86,
+      "grad_norm": 42.28071588262047,
+      "learning_rate": 0.00015000000000000001,
+      "loss": 7.5507,
+      "step": 75
+    },
+    {
+      "epoch": 0.87,
+      "grad_norm": 27.40577229308579,
+      "learning_rate": 0.000152,
+      "loss": 7.6852,
+      "step": 76
+    },
+    {
+      "epoch": 0.89,
+      "grad_norm": 21.422615778584905,
+      "learning_rate": 0.000154,
+      "loss": 7.6071,
+      "step": 77
+    },
+    {
+      "epoch": 0.9,
+      "grad_norm": 28.22569647336005,
+      "learning_rate": 0.00015600000000000002,
+      "loss": 7.5556,
+      "step": 78
+    },
+    {
+      "epoch": 0.91,
+      "grad_norm": 20.075502929468385,
+      "learning_rate": 0.00015800000000000002,
+      "loss": 7.529,
+      "step": 79
+    },
+    {
+      "epoch": 0.92,
+      "grad_norm": 14.896626626967638,
+      "learning_rate": 0.00016,
+      "loss": 7.5352,
+      "step": 80
+    },
+    {
+      "epoch": 0.93,
+      "grad_norm": 16.98565517398293,
+      "learning_rate": 0.000162,
+      "loss": 7.5734,
+      "step": 81
+    },
+    {
+      "epoch": 0.94,
+      "grad_norm": 17.76803489316823,
+      "learning_rate": 0.000164,
+      "loss": 7.591,
+      "step": 82
+    },
+    {
+      "epoch": 0.95,
+      "grad_norm": 15.601674972992182,
+      "learning_rate": 0.000166,
+      "loss": 7.4251,
+      "step": 83
+    },
+    {
+      "epoch": 0.97,
+      "grad_norm": 21.36580777081032,
+      "learning_rate": 0.000168,
+      "loss": 7.728,
+      "step": 84
+    },
+    {
+      "epoch": 0.98,
+      "grad_norm": 19.73380900316605,
+      "learning_rate": 0.00017,
+      "loss": 7.3548,
+      "step": 85
+    },
+    {
+      "epoch": 0.99,
+      "grad_norm": 16.40313028374367,
+      "learning_rate": 0.000172,
+      "loss": 7.3652,
+      "step": 86
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 11.846265862818466,
+      "learning_rate": 0.000174,
+      "loss": 7.4363,
+      "step": 87
+    },
+    {
+      "epoch": 1.01,
+      "grad_norm": 11.82176337757903,
+      "learning_rate": 0.00017600000000000002,
+      "loss": 7.3967,
+      "step": 88
+    },
+    {
+      "epoch": 1.01,
+      "eval_loss": 7.461406707763672,
+      "eval_runtime": 1.3867,
+      "eval_samples_per_second": 13.702,
+      "eval_steps_per_second": 3.606,
+      "step": 88
+    },
+    {
+      "epoch": 1.02,
+      "grad_norm": 14.25806078550941,
+      "learning_rate": 0.00017800000000000002,
+      "loss": 7.3814,
+      "step": 89
+    },
+    {
+      "epoch": 1.03,
+      "grad_norm": 10.077089336320658,
+      "learning_rate": 0.00018,
+      "loss": 7.2599,
+      "step": 90
+    },
+    {
+      "epoch": 1.01,
+      "grad_norm": 15.159363102480643,
+      "learning_rate": 0.000182,
+      "loss": 7.5719,
+      "step": 91
+    },
+    {
+      "epoch": 1.02,
+      "grad_norm": 14.678683877513562,
+      "learning_rate": 0.00018400000000000003,
+      "loss": 7.4794,
+      "step": 92
+    },
+    {
+      "epoch": 1.03,
+      "grad_norm": 16.396549401486602,
+      "learning_rate": 0.00018600000000000002,
+      "loss": 7.6211,
+      "step": 93
+    },
+    {
+      "epoch": 1.05,
+      "grad_norm": 15.078668930441678,
+      "learning_rate": 0.000188,
+      "loss": 7.3625,
+      "step": 94
+    },
+    {
+      "epoch": 1.06,
+      "grad_norm": 13.546080322892589,
+      "learning_rate": 0.00019,
+      "loss": 7.3016,
+      "step": 95
+    },
+    {
+      "epoch": 1.07,
+      "grad_norm": 18.231949750348118,
+      "learning_rate": 0.000192,
+      "loss": 7.386,
+      "step": 96
+    },
+    {
+      "epoch": 1.08,
+      "grad_norm": 12.444086372431334,
+      "learning_rate": 0.000194,
+      "loss": 7.5534,
+      "step": 97
+    },
+    {
+      "epoch": 1.09,
+      "grad_norm": 9.994459143045797,
+      "learning_rate": 0.000196,
+      "loss": 7.4254,
+      "step": 98
+    },
+    {
+      "epoch": 1.1,
+      "grad_norm": 14.889247706503676,
+      "learning_rate": 0.00019800000000000002,
+      "loss": 7.3938,
+      "step": 99
+    },
+    {
+      "epoch": 1.11,
+      "grad_norm": 19.194847464605004,
+      "learning_rate": 0.0002,
+      "loss": 7.3875,
+      "step": 100
+    },
+    {
+      "epoch": 1.13,
+      "grad_norm": 19.234057674068023,
+      "learning_rate": 0.00019998096274980728,
+      "loss": 7.348,
+      "step": 101
+    },
+    {
+      "epoch": 1.14,
+      "grad_norm": 10.248320279074052,
+      "learning_rate": 0.000199923858247567,
+      "loss": 7.2365,
+      "step": 102
+    },
+    {
+      "epoch": 1.15,
+      "grad_norm": 15.579201434411983,
+      "learning_rate": 0.00019982870823553308,
+      "loss": 7.2351,
+      "step": 103
+    },
+    {
+      "epoch": 1.16,
+      "grad_norm": 10.951732160571428,
+      "learning_rate": 0.00019969554894159723,
+      "loss": 7.4413,
+      "step": 104
+    },
+    {
+      "epoch": 1.17,
+      "grad_norm": 12.034336556190219,
+      "learning_rate": 0.00019952443106549533,
+      "loss": 7.4256,
+      "step": 105
+    },
+    {
+      "epoch": 1.18,
+      "grad_norm": 11.555864069154051,
+      "learning_rate": 0.00019931541975950378,
+      "loss": 7.0947,
+      "step": 106
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 8.079125584036722,
+      "learning_rate": 0.00019906859460363307,
+      "loss": 7.3727,
+      "step": 107
+    },
+    {
+      "epoch": 1.21,
+      "grad_norm": 10.901301706963714,
+      "learning_rate": 0.00019878404957532814,
+      "loss": 7.2419,
+      "step": 108
+    },
+    {
+      "epoch": 1.22,
+      "grad_norm": 9.771076993744128,
+      "learning_rate": 0.0001984618930136869,
+      "loss": 7.2896,
+      "step": 109
+    },
+    {
+      "epoch": 1.23,
+      "grad_norm": 11.395240298688034,
+      "learning_rate": 0.00019810224757821064,
+      "loss": 7.2567,
+      "step": 110
+    },
+    {
+      "epoch": 1.23,
+      "eval_loss": 7.299312591552734,
+      "eval_runtime": 1.381,
+      "eval_samples_per_second": 13.758,
+      "eval_steps_per_second": 3.621,
+      "step": 110
+    },
+    {
+      "epoch": 1.24,
+      "grad_norm": 11.839939762683946,
+      "learning_rate": 0.00019770525020210204,
+      "loss": 7.3145,
+      "step": 111
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 13.454044820108352,
+      "learning_rate": 0.0001972710520401287,
+      "loss": 7.3279,
+      "step": 112
+    },
+    {
+      "epoch": 1.26,
+      "grad_norm": 10.349510041217494,
+      "learning_rate": 0.0001967998184110713,
+      "loss": 7.0995,
+      "step": 113
+    },
+    {
+      "epoch": 1.28,
+      "grad_norm": 11.299695940906268,
+      "learning_rate": 0.00019629172873477995,
+      "loss": 7.1545,
+      "step": 114
+    },
+    {
+      "epoch": 1.29,
+      "grad_norm": 10.300523126150598,
+      "learning_rate": 0.00019574697646386027,
+      "loss": 7.2619,
+      "step": 115
+    },
+    {
+      "epoch": 1.3,
+      "grad_norm": 12.280692750358627,
+      "learning_rate": 0.0001951657690100178,
+      "loss": 7.4703,
+      "step": 116
+    },
+    {
+      "epoch": 1.31,
+      "grad_norm": 11.308344462386193,
+      "learning_rate": 0.0001945483276650868,
+      "loss": 7.2927,
+      "step": 117
+    },
+    {
+      "epoch": 1.32,
+      "grad_norm": 9.46182612207882,
+      "learning_rate": 0.0001938948875167745,
+      "loss": 7.1097,
+      "step": 118
+    },
+    {
+      "epoch": 1.33,
+      "grad_norm": 8.495214573764272,
+      "learning_rate": 0.00019320569735915271,
+      "loss": 7.2199,
+      "step": 119
+    },
+    {
+      "epoch": 1.34,
+      "grad_norm": 10.888318204738358,
+      "learning_rate": 0.00019248101959793066,
+      "loss": 7.4214,
+      "step": 120
+    },
+    {
+      "epoch": 1.36,
+      "grad_norm": 15.655909856272627,
+      "learning_rate": 0.00019172113015054532,
+      "loss": 7.4141,
+      "step": 121
+    },
+    {
+      "epoch": 1.37,
+      "grad_norm": 14.348570302527095,
+      "learning_rate": 0.00019092631834110723,
+      "loss": 7.2034,
+      "step": 122
+    },
+    {
+      "epoch": 1.38,
+      "grad_norm": 11.216620878113865,
+      "learning_rate": 0.0001900968867902419,
+      "loss": 7.1925,
+      "step": 123
+    },
+    {
+      "epoch": 1.39,
+      "grad_norm": 15.214529842718143,
+      "learning_rate": 0.00018923315129986835,
+      "loss": 7.204,
+      "step": 124
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 24.933551829531577,
+      "learning_rate": 0.00018833544073295917,
+      "loss": 7.5231,
+      "step": 125
+    },
+    {
+      "epoch": 1.41,
+      "grad_norm": 12.291980045440198,
+      "learning_rate": 0.00018740409688832764,
+      "loss": 7.2323,
+      "step": 126
+    },
+    {
+      "epoch": 1.43,
+      "grad_norm": 19.503396545387695,
+      "learning_rate": 0.00018643947437048944,
+      "loss": 7.3214,
+      "step": 127
+    },
+    {
+      "epoch": 1.44,
+      "grad_norm": 13.982404823681808,
+      "learning_rate": 0.00018544194045464886,
+      "loss": 7.4008,
+      "step": 128
+    },
+    {
+      "epoch": 1.45,
+      "grad_norm": 13.775851829485163,
+      "learning_rate": 0.00018441187494686053,
+      "loss": 7.2854,
+      "step": 129
+    },
+    {
+      "epoch": 1.46,
+      "grad_norm": 16.44073199273777,
+      "learning_rate": 0.0001833496700394202,
+      "loss": 7.4218,
+      "step": 130
+    },
+    {
+      "epoch": 1.47,
+      "grad_norm": 21.66328419423237,
+      "learning_rate": 0.00018225573016153945,
+      "loss": 7.3589,
+      "step": 131
+    },
+    {
+      "epoch": 1.48,
+      "grad_norm": 14.027278482513156,
+      "learning_rate": 0.00018113047182536127,
+      "loss": 7.3329,
+      "step": 132
+    },
+    {
+      "epoch": 1.48,
+      "eval_loss": 7.374873161315918,
+      "eval_runtime": 1.3815,
+      "eval_samples_per_second": 13.753,
+      "eval_steps_per_second": 3.619,
+      "step": 132
+    },
+    {
+      "epoch": 1.49,
+      "grad_norm": 10.89412121909016,
+      "learning_rate": 0.00017997432346737524,
+      "loss": 7.3277,
+      "step": 133
+    },
+    {
+      "epoch": 1.51,
+      "grad_norm": 18.182727305484867,
+      "learning_rate": 0.00017878772528529232,
+      "loss": 7.3505,
+      "step": 134
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 18.738647279231998,
+      "learning_rate": 0.000177571129070442,
+      "loss": 7.4395,
+      "step": 135
+    },
+    {
+      "epoch": 1.53,
+      "grad_norm": 8.544634593385059,
+      "learning_rate": 0.00017632499803575474,
+      "loss": 7.3132,
+      "step": 136
+    },
+    {
+      "epoch": 1.54,
+      "grad_norm": 23.787945119488832,
+      "learning_rate": 0.00017504980663939613,
+      "loss": 7.3635,
+      "step": 137
+    },
+    {
+      "epoch": 1.55,
+      "grad_norm": 15.988837162794528,
+      "learning_rate": 0.00017374604040411935,
+      "loss": 7.298,
+      "step": 138
+    },
+    {
+      "epoch": 1.56,
+      "grad_norm": 8.594141503866256,
+      "learning_rate": 0.00017241419573240462,
+      "loss": 7.2263,
+      "step": 139
+    },
+    {
+      "epoch": 1.57,
+      "grad_norm": 10.834111660116264,
+      "learning_rate": 0.00017105477971745666,
+      "loss": 7.2234,
+      "step": 140
+    },
+    {
+      "epoch": 1.59,
+      "grad_norm": 20.898663112099943,
+      "learning_rate": 0.00016966830995013133,
+      "loss": 7.4312,
+      "step": 141
+    },
+    {
+      "epoch": 1.6,
+      "grad_norm": 9.882602517532073,
+      "learning_rate": 0.00016825531432186543,
+      "loss": 7.3389,
+      "step": 142
+    },
+    {
+      "epoch": 1.61,
+      "grad_norm": 7.435249821686469,
+      "learning_rate": 0.00016681633082368498,
+      "loss": 7.2015,
+      "step": 143
+    },
+    {
+      "epoch": 1.62,
+      "grad_norm": 13.215068331486576,
+      "learning_rate": 0.0001653519073413675,
+      "loss": 7.1425,
+      "step": 144
+    },
+    {
+      "epoch": 1.63,
+      "grad_norm": 11.754841557793458,
+      "learning_rate": 0.00016386260144683745,
+      "loss": 7.3101,
+      "step": 145
+    },
+    {
+      "epoch": 1.64,
+      "grad_norm": 7.674786980372276,
+      "learning_rate": 0.00016234898018587337,
+      "loss": 7.1703,
+      "step": 146
+    },
+    {
+      "epoch": 1.66,
+      "grad_norm": 8.30389279917059,
+      "learning_rate": 0.00016081161986220807,
+      "loss": 7.2516,
+      "step": 147
+    },
+    {
+      "epoch": 1.67,
+      "grad_norm": 14.325137493545546,
+      "learning_rate": 0.00015925110581810394,
+      "loss": 7.3326,
+      "step": 148
+    },
+    {
+      "epoch": 1.68,
+      "grad_norm": 9.943124774719692,
+      "learning_rate": 0.00015766803221148673,
+      "loss": 7.5501,
+      "step": 149
+    },
+    {
+      "epoch": 1.69,
+      "grad_norm": 11.710815673357324,
+      "learning_rate": 0.00015606300178972287,
+      "loss": 7.3344,
+      "step": 150
+    },
+    {
+      "epoch": 1.7,
+      "grad_norm": 8.287278539768234,
+      "learning_rate": 0.00015443662566012645,
+      "loss": 7.2426,
+      "step": 151
+    },
+    {
+      "epoch": 1.71,
+      "grad_norm": 8.045767995001754,
+      "learning_rate": 0.00015278952305728324,
+      "loss": 7.2319,
+      "step": 152
+    },
+    {
+      "epoch": 1.72,
+      "grad_norm": 16.213767636990735,
+      "learning_rate": 0.00015112232110728015,
+      "loss": 7.2226,
+      "step": 153
+    },
+    {
+      "epoch": 1.74,
+      "grad_norm": 10.115986520778595,
+      "learning_rate": 0.00014943565458893,
+      "loss": 7.0671,
+      "step": 154
+    },
+    {
+      "epoch": 1.74,
+      "eval_loss": 7.336472511291504,
+      "eval_runtime": 1.3763,
+      "eval_samples_per_second": 13.806,
+      "eval_steps_per_second": 3.633,
+      "step": 154
+    },
+    {
+      "epoch": 1.75,
+      "grad_norm": 7.370162670560762,
+      "learning_rate": 0.00014773016569208283,
+      "loss": 7.2456,
+      "step": 155
+    },
+    {
+      "epoch": 1.76,
+      "grad_norm": 14.854609514494987,
+      "learning_rate": 0.00014600650377311522,
+      "loss": 7.0447,
+      "step": 156
+    },
+    {
+      "epoch": 1.77,
+      "grad_norm": 9.984605765623492,
+      "learning_rate": 0.0001442653251076912,
+      "loss": 7.1948,
+      "step": 157
+    },
+    {
+      "epoch": 1.78,
+      "grad_norm": 11.053561117527067,
+      "learning_rate": 0.00014250729264088843,
+      "loss": 7.1485,
+      "step": 158
+    },
+    {
+      "epoch": 1.79,
+      "grad_norm": 11.604464862126479,
+      "learning_rate": 0.00014073307573478526,
+      "loss": 7.4198,
+      "step": 159
+    },
+    {
+      "epoch": 1.8,
+      "grad_norm": 8.62267592810566,
+      "learning_rate": 0.00013894334991360448,
+      "loss": 7.3045,
+      "step": 160
+    },
+    {
+      "epoch": 1.82,
+      "grad_norm": 10.961255591651343,
+      "learning_rate": 0.00013713879660651068,
+      "loss": 6.9983,
+      "step": 161
+    },
+    {
+      "epoch": 1.83,
+      "grad_norm": 11.12104950103157,
+      "learning_rate": 0.0001353201028881598,
+      "loss": 7.1046,
+      "step": 162
+    },
+    {
+      "epoch": 1.84,
+      "grad_norm": 8.807345567053673,
+      "learning_rate": 0.00013348796121709862,
+      "loss": 7.3378,
+      "step": 163
+    },
+    {
+      "epoch": 1.85,
+      "grad_norm": 10.666053356976384,
+      "learning_rate": 0.00013164306917211476,
+      "loss": 7.0008,
+      "step": 164
+    },
+    {
+      "epoch": 1.86,
+      "grad_norm": 7.835224761952429,
+      "learning_rate": 0.000129786129186637,
+      "loss": 6.9881,
+      "step": 165
+    },
+    {
+      "epoch": 1.87,
+      "grad_norm": 19.4062633667929,
+      "learning_rate": 0.00012791784828128724,
+      "loss": 7.2579,
+      "step": 166
+    },
+    {
+      "epoch": 1.89,
+      "grad_norm": 10.113263144537674,
+      "learning_rate": 0.00012603893779468604,
+      "loss": 7.3091,
+      "step": 167
+    },
+    {
+      "epoch": 1.9,
+      "grad_norm": 10.870503515462726,
+      "learning_rate": 0.0001241501131126138,
+      "loss": 7.4527,
+      "step": 168
+    },
+    {
+      "epoch": 1.91,
+      "grad_norm": 8.533887294828766,
+      "learning_rate": 0.00012225209339563145,
+      "loss": 7.2855,
+      "step": 169
+    },
+    {
+      "epoch": 1.92,
+      "grad_norm": 13.486746604884923,
+      "learning_rate": 0.0001203456013052634,
+      "loss": 7.2705,
+      "step": 170
+    },
+    {
+      "epoch": 1.93,
+      "grad_norm": 10.69251762526038,
+      "learning_rate": 0.00011843136272884794,
+      "loss": 7.0932,
+      "step": 171
+    },
+    {
+      "epoch": 1.94,
+      "grad_norm": 9.124923914761991,
+      "learning_rate": 0.00011651010650315923,
+      "loss": 7.3754,
+      "step": 172
+    },
+    {
+      "epoch": 1.95,
+      "grad_norm": 13.481473855966252,
+      "learning_rate": 0.00011458256413690633,
+      "loss": 7.3104,
+      "step": 173
+    },
+    {
+      "epoch": 1.97,
+      "grad_norm": 13.187202052506544,
+      "learning_rate": 0.00011264946953221496,
+      "loss": 7.3614,
+      "step": 174
+    },
+    {
+      "epoch": 1.98,
+      "grad_norm": 13.527193725705766,
+      "learning_rate": 0.00011071155870519777,
+      "loss": 7.2734,
+      "step": 175
+    },
+    {
+      "epoch": 1.99,
+      "grad_norm": 8.023369511183457,
+      "learning_rate": 0.00010876956950572006,
+      "loss": 7.4786,
+      "step": 176
+    },
+    {
+      "epoch": 1.99,
+      "eval_loss": 7.319369316101074,
+      "eval_runtime": 1.3976,
+      "eval_samples_per_second": 13.595,
+      "eval_steps_per_second": 3.578,
+      "step": 176
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 11.41077205951204,
+      "learning_rate": 0.0001068242413364671,
+      "loss": 7.2116,
+      "step": 177
+    },
+    {
+      "epoch": 2.01,
+      "grad_norm": 9.551865991569315,
+      "learning_rate": 0.00010487631487142017,
+      "loss": 7.2411,
+      "step": 178
+    },
+    {
+      "epoch": 2.02,
+      "grad_norm": 11.576046488111642,
+      "learning_rate": 0.00010292653177384876,
+      "loss": 7.4401,
+      "step": 179
+    },
+    {
+      "epoch": 2.01,
+      "grad_norm": 9.356673167513755,
+      "learning_rate": 0.00010097563441392581,
+      "loss": 7.3184,
+      "step": 180
+    },
+    {
+      "epoch": 2.02,
+      "grad_norm": 12.119641565241189,
+      "learning_rate": 9.90243655860742e-05,
+      "loss": 7.2285,
+      "step": 181
+    },
+    {
+      "epoch": 2.03,
+      "grad_norm": 11.06643382673938,
+      "learning_rate": 9.707346822615128e-05,
+      "loss": 7.1729,
+      "step": 182
+    },
+    {
+      "epoch": 2.05,
+      "grad_norm": 9.198991504624212,
+      "learning_rate": 9.512368512857984e-05,
+      "loss": 7.2619,
+      "step": 183
+    },
+    {
+      "epoch": 2.06,
+      "grad_norm": 14.240944645441582,
+      "learning_rate": 9.317575866353292e-05,
+      "loss": 7.5778,
+      "step": 184
+    },
+    {
+      "epoch": 2.07,
+      "grad_norm": 8.473455381717486,
+      "learning_rate": 9.123043049427995e-05,
+      "loss": 7.2432,
+      "step": 185
+    },
+    {
+      "epoch": 2.08,
+      "grad_norm": 12.027303786674432,
+      "learning_rate": 8.928844129480227e-05,
+      "loss": 7.4036,
+      "step": 186
+    },
+    {
+      "epoch": 2.09,
+      "grad_norm": 10.45148339945213,
+      "learning_rate": 8.735053046778506e-05,
+      "loss": 7.2052,
+      "step": 187
+    },
+    {
+      "epoch": 2.1,
+      "grad_norm": 12.816375890508084,
+      "learning_rate": 8.541743586309365e-05,
+      "loss": 7.1882,
+      "step": 188
+    },
+    {
+      "epoch": 2.11,
+      "grad_norm": 12.872264386256504,
+      "learning_rate": 8.348989349684076e-05,
+      "loss": 7.1129,
+      "step": 189
+    },
+    {
+      "epoch": 2.13,
+      "grad_norm": 8.712185045895534,
+      "learning_rate": 8.156863727115211e-05,
+      "loss": 7.3354,
+      "step": 190
+    },
+    {
+      "epoch": 2.14,
+      "grad_norm": 7.169962044034168,
+      "learning_rate": 7.965439869473664e-05,
+      "loss": 7.1493,
+      "step": 191
+    },
+    {
+      "epoch": 2.15,
+      "grad_norm": 9.20643667535514,
+      "learning_rate": 7.774790660436858e-05,
+      "loss": 7.1867,
+      "step": 192
+    },
+    {
+      "epoch": 2.16,
+      "grad_norm": 9.71231661001151,
+      "learning_rate": 7.584988688738622e-05,
+      "loss": 7.1941,
+      "step": 193
+    },
+    {
+      "epoch": 2.17,
+      "grad_norm": 9.869652181292283,
+      "learning_rate": 7.396106220531398e-05,
+      "loss": 7.2606,
+      "step": 194
+    },
+    {
+      "epoch": 2.18,
+      "grad_norm": 7.820315914553451,
+      "learning_rate": 7.208215171871277e-05,
+      "loss": 7.1575,
+      "step": 195
+    },
+    {
+      "epoch": 2.2,
+      "grad_norm": 8.668852408202328,
+      "learning_rate": 7.021387081336301e-05,
+      "loss": 7.3339,
+      "step": 196
+    },
+    {
+      "epoch": 2.21,
+      "grad_norm": 6.97843034949997,
+      "learning_rate": 6.835693082788525e-05,
+      "loss": 7.307,
+      "step": 197
+    },
+    {
+      "epoch": 2.22,
+      "grad_norm": 16.571225038227503,
+      "learning_rate": 6.651203878290139e-05,
+      "loss": 7.3548,
+      "step": 198
+    },
+    {
+      "epoch": 2.22,
+      "eval_loss": 7.309223175048828,
+      "eval_runtime": 1.3716,
+      "eval_samples_per_second": 13.852,
+      "eval_steps_per_second": 3.645,
+      "step": 198
+    },
+    {
+      "epoch": 2.23,
+      "grad_norm": 14.645439276497052,
+      "learning_rate": 6.46798971118402e-05,
+      "loss": 7.1198,
+      "step": 199
+    },
+    {
+      "epoch": 2.24,
+      "grad_norm": 11.180148474401758,
+      "learning_rate": 6.286120339348935e-05,
+      "loss": 7.383,
+      "step": 200
+    },
+    {
+      "epoch": 2.25,
+      "grad_norm": 8.297433608048966,
+      "learning_rate": 6.105665008639557e-05,
+      "loss": 7.0708,
+      "step": 201
+    },
+    {
+      "epoch": 2.26,
+      "grad_norm": 7.35350265603106,
+      "learning_rate": 5.926692426521474e-05,
+      "loss": 7.2686,
+      "step": 202
+    },
+    {
+      "epoch": 2.28,
+      "grad_norm": 10.931447125375616,
+      "learning_rate": 5.749270735911158e-05,
+      "loss": 7.0715,
+      "step": 203
+    },
+    {
+      "epoch": 2.29,
+      "grad_norm": 12.580996736421003,
+      "learning_rate": 5.573467489230879e-05,
+      "loss": 7.3576,
+      "step": 204
+    },
+    {
+      "epoch": 2.3,
+      "grad_norm": 12.648533406639174,
+      "learning_rate": 5.399349622688479e-05,
+      "loss": 7.3546,
+      "step": 205
+    },
+    {
+      "epoch": 2.31,
+      "grad_norm": 8.448979459452485,
+      "learning_rate": 5.226983430791722e-05,
+      "loss": 7.0968,
+      "step": 206
+    },
+    {
+      "epoch": 2.32,
+      "grad_norm": 14.75106031841705,
+      "learning_rate": 5.0564345411070025e-05,
+      "loss": 7.2726,
+      "step": 207
+    },
+    {
+      "epoch": 2.33,
+      "grad_norm": 7.623885276483663,
+      "learning_rate": 4.8877678892719866e-05,
+      "loss": 7.2646,
+      "step": 208
+    },
+    {
+      "epoch": 2.34,
+      "grad_norm": 11.112819218207525,
+      "learning_rate": 4.721047694271676e-05,
+      "loss": 7.1452,
+      "step": 209
+    },
+    {
+      "epoch": 2.36,
+      "grad_norm": 15.00591466641024,
+      "learning_rate": 4.556337433987359e-05,
+      "loss": 6.9634,
+      "step": 210
+    },
+    {
+      "epoch": 2.37,
+      "grad_norm": 9.000881151933429,
+      "learning_rate": 4.393699821027716e-05,
+      "loss": 7.2109,
+      "step": 211
+    },
+    {
+      "epoch": 2.38,
+      "grad_norm": 8.677018023760695,
+      "learning_rate": 4.2331967788513295e-05,
+      "loss": 7.1976,
+      "step": 212
+    },
+    {
+      "epoch": 2.39,
+      "grad_norm": 7.021254016561991,
+      "learning_rate": 4.074889418189608e-05,
+      "loss": 7.2308,
+      "step": 213
+    },
+    {
+      "epoch": 2.4,
+      "grad_norm": 14.097624320239541,
+      "learning_rate": 3.9188380137791936e-05,
+      "loss": 7.2678,
+      "step": 214
+    },
+    {
+      "epoch": 2.41,
+      "grad_norm": 7.660009091922155,
+      "learning_rate": 3.7651019814126654e-05,
+      "loss": 7.2609,
+      "step": 215
+    },
+    {
+      "epoch": 2.43,
+      "grad_norm": 6.879429170747529,
+      "learning_rate": 3.613739855316257e-05,
+      "loss": 7.1545,
+      "step": 216
+    },
+    {
+      "epoch": 2.44,
+      "grad_norm": 9.342549224552343,
+      "learning_rate": 3.46480926586325e-05,
+      "loss": 7.122,
+      "step": 217
+    },
+    {
+      "epoch": 2.45,
+      "grad_norm": 9.584840694689616,
+      "learning_rate": 3.3183669176315045e-05,
+      "loss": 7.2049,
+      "step": 218
+    },
+    {
+      "epoch": 2.46,
+      "grad_norm": 14.37215919620074,
+      "learning_rate": 3.174468567813461e-05,
+      "loss": 7.4374,
+      "step": 219
+    },
+    {
+      "epoch": 2.47,
+      "grad_norm": 6.257502820833114,
+      "learning_rate": 3.033169004986873e-05,
+      "loss": 7.1782,
+      "step": 220
+    },
+    {
+      "epoch": 2.47,
+      "eval_loss": 7.296360492706299,
+      "eval_runtime": 1.3751,
+      "eval_samples_per_second": 13.818,
+      "eval_steps_per_second": 3.636,
+      "step": 220
+    },
+    {
+      "epoch": 2.48,
+      "grad_norm": 7.202404193022492,
+      "learning_rate": 2.894522028254334e-05,
+      "loss": 7.1248,
+      "step": 221
+    },
+    {
+      "epoch": 2.49,
+      "grad_norm": 11.03004064225623,
+      "learning_rate": 2.7585804267595384e-05,
+      "loss": 7.4726,
+      "step": 222
+    },
+    {
+      "epoch": 2.51,
+      "grad_norm": 10.208828744090669,
+      "learning_rate": 2.6253959595880673e-05,
+      "loss": 7.1886,
+      "step": 223
+    },
+    {
+      "epoch": 2.52,
+      "grad_norm": 7.194649499092706,
+      "learning_rate": 2.495019336060387e-05,
+      "loss": 7.2576,
+      "step": 224
+    },
+    {
+      "epoch": 2.53,
+      "grad_norm": 15.063508439238962,
+      "learning_rate": 2.367500196424529e-05,
+      "loss": 6.924,
+      "step": 225
+    },
+    {
+      "epoch": 2.54,
+      "grad_norm": 9.528135297299904,
+      "learning_rate": 2.242887092955801e-05,
+      "loss": 7.1941,
+      "step": 226
+    },
+    {
+      "epoch": 2.55,
+      "grad_norm": 7.719290154588222,
+      "learning_rate": 2.121227471470768e-05,
+      "loss": 7.3391,
+      "step": 227
+    },
+    {
+      "epoch": 2.56,
+      "grad_norm": 7.714834107984481,
+      "learning_rate": 2.002567653262479e-05,
+      "loss": 7.1969,
+      "step": 228
+    },
+    {
+      "epoch": 2.57,
+      "grad_norm": 8.228923346547797,
+      "learning_rate": 1.8869528174638752e-05,
+      "loss": 7.1543,
+      "step": 229
+    },
+    {
+      "epoch": 2.59,
+      "grad_norm": 6.001756411009184,
+      "learning_rate": 1.774426983846058e-05,
+      "loss": 7.2465,
+      "step": 230
+    },
+    {
+      "epoch": 2.6,
+      "grad_norm": 10.434576481704424,
+      "learning_rate": 1.6650329960579792e-05,
+      "loss": 7.2623,
+      "step": 231
+    },
+    {
+      "epoch": 2.61,
+      "grad_norm": 8.095148267701724,
+      "learning_rate": 1.5588125053139468e-05,
+      "loss": 7.1761,
+      "step": 232
+    },
+    {
+      "epoch": 2.62,
+      "grad_norm": 9.145578669056132,
+      "learning_rate": 1.4558059545351143e-05,
+      "loss": 7.1295,
+      "step": 233
+    },
+    {
+      "epoch": 2.63,
+      "grad_norm": 7.413949304325555,
+      "learning_rate": 1.3560525629510568e-05,
+      "loss": 7.1572,
+      "step": 234
+    },
+    {
+      "epoch": 2.64,
+      "grad_norm": 8.238266895583749,
+      "learning_rate": 1.259590311167238e-05,
+      "loss": 7.2317,
+      "step": 235
+    },
+    {
+      "epoch": 2.66,
+      "grad_norm": 7.150286325977236,
+      "learning_rate": 1.166455926704082e-05,
+      "loss": 7.3381,
+      "step": 236
+    },
+    {
+      "epoch": 2.67,
+      "grad_norm": 10.918607427652104,
+      "learning_rate": 1.0766848700131648e-05,
+      "loss": 7.3925,
+      "step": 237
+    },
+    {
+      "epoch": 2.68,
+      "grad_norm": 6.6969446531973285,
+      "learning_rate": 9.903113209758096e-06,
+      "loss": 7.2799,
+      "step": 238
+    },
+    {
+      "epoch": 2.69,
+      "grad_norm": 13.477825461653845,
+      "learning_rate": 9.073681658892775e-06,
+      "loss": 6.901,
+      "step": 239
+    },
+    {
+      "epoch": 2.7,
+      "grad_norm": 6.783585528560174,
+      "learning_rate": 8.278869849454718e-06,
+      "loss": 7.2819,
+      "step": 240
+    },
+    {
+      "epoch": 2.71,
+      "grad_norm": 9.548057270672551,
+      "learning_rate": 7.5189804020693536e-06,
+      "loss": 7.1192,
+      "step": 241
+    },
+    {
+      "epoch": 2.72,
+      "grad_norm": 7.742180355619181,
+      "learning_rate": 6.794302640847294e-06,
+      "loss": 7.2729,
+      "step": 242
+    },
+    {
+      "epoch": 2.72,
+      "eval_loss": 7.292217254638672,
+      "eval_runtime": 1.3839,
+      "eval_samples_per_second": 13.729,
+      "eval_steps_per_second": 3.613,
+      "step": 242
+    },
+    {
+      "epoch": 2.74,
+      "grad_norm": 10.845656873156827,
+      "learning_rate": 6.1051124832254944e-06,
+      "loss": 6.9693,
+      "step": 243
+    },
+    {
+      "epoch": 2.75,
+      "grad_norm": 10.323147542445835,
+      "learning_rate": 5.451672334913216e-06,
+      "loss": 7.3785,
+      "step": 244
+    },
+    {
+      "epoch": 2.76,
+      "grad_norm": 10.41796325752105,
+      "learning_rate": 4.834230989982213e-06,
+      "loss": 7.3632,
+      "step": 245
+    },
+    {
+      "epoch": 2.77,
+      "grad_norm": 6.590462249149956,
+      "learning_rate": 4.253023536139733e-06,
+      "loss": 7.279,
+      "step": 246
+    },
+    {
+      "epoch": 2.78,
+      "grad_norm": 9.637387713657647,
+      "learning_rate": 3.7082712652200867e-06,
+      "loss": 7.2857,
+      "step": 247
+    },
+    {
+      "epoch": 2.79,
+      "grad_norm": 5.983998578137723,
+      "learning_rate": 3.2001815889286856e-06,
+      "loss": 7.2269,
+      "step": 248
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 6.750577266168699,
+      "learning_rate": 2.728947959871353e-06,
+      "loss": 7.2672,
+      "step": 249
+    },
+    {
+      "epoch": 2.82,
+      "grad_norm": 8.828265865856226,
+      "learning_rate": 2.294749797897955e-06,
+      "loss": 7.1355,
+      "step": 250
+    },
+    {
+      "epoch": 2.83,
+      "grad_norm": 8.092540989425222,
+      "learning_rate": 1.8977524217893783e-06,
+      "loss": 7.079,
+      "step": 251
+    },
+    {
+      "epoch": 2.84,
+      "grad_norm": 14.298932570519032,
+      "learning_rate": 1.5381069863131037e-06,
+      "loss": 7.1462,
+      "step": 252
+    },
+    {
+      "epoch": 2.85,
+      "grad_norm": 6.9722905081773465,
+      "learning_rate": 1.2159504246718522e-06,
+      "loss": 7.1766,
+      "step": 253
+    },
+    {
+      "epoch": 2.86,
+      "grad_norm": 8.429744971276516,
+      "learning_rate": 9.314053963669245e-07,
+      "loss": 7.3844,
+      "step": 254
+    },
+    {
+      "epoch": 2.87,
+      "grad_norm": 7.0927677230675386,
+      "learning_rate": 6.845802404962243e-07,
+      "loss": 7.3318,
+      "step": 255
+    },
+    {
+      "epoch": 2.89,
+      "grad_norm": 6.6033524293092265,
+      "learning_rate": 4.7556893450466653e-07,
+      "loss": 7.1159,
+      "step": 256
+    },
+    {
+      "epoch": 2.9,
+      "grad_norm": 9.59870072155561,
+      "learning_rate": 3.044510584027771e-07,
+      "loss": 7.3733,
+      "step": 257
+    },
+    {
+      "epoch": 2.91,
+      "grad_norm": 7.255692976305313,
+      "learning_rate": 1.7129176446692984e-07,
+      "loss": 7.2345,
+      "step": 258
+    },
+    {
+      "epoch": 2.92,
+      "grad_norm": 7.239401601436061,
+      "learning_rate": 7.614175243301213e-08,
+      "loss": 7.2672,
+      "step": 259
+    },
+    {
+      "epoch": 2.93,
+      "grad_norm": 6.342843427851908,
+      "learning_rate": 1.9037250192732726e-08,
+      "loss": 7.2741,
+      "step": 260
+    },
+    {
+      "epoch": 2.94,
+      "grad_norm": 8.76817744282957,
+      "learning_rate": 0.0,
+      "loss": 7.2543,
+      "step": 261
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 261,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 87,
+  "total_flos": 1.8244065215014502e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-261/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b50b6bb5c490e8ef22b44507d2914da60c8e4f8c0f5671692c3b6485d1e90608
+size 6395

checkpoint-261/zero_to_fp32.py

@@ -0,0 +1,592 @@
+#!/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_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    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 _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+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() if _has_callable(shape, 'numel') else math.prod(shape)
+            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("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    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, tag=args.tag)

checkpoint-87/config.json

@@ -0,0 +1,26 @@
+{
+  "_name_or_path": "mistralai/Mistral-7B-v0.1",
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 10000.0,
+  "sliding_window": 4096,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.38.0",
+  "use_cache": false,
+  "vocab_size": 32000
+}

checkpoint-87/generation_config.json

@@ -0,0 +1,7 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 1,
+  "do_sample": true,
+  "eos_token_id": 2,
+  "transformers_version": "4.38.0"
+}

checkpoint-87/global_step87/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa77e1c15fd13f4b39e2cb678b0bcc5bea3143286be846509a06dd79902b1ad9
+size 21725204471

checkpoint-87/global_step87/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25db455cc5e6b2b7fc43cee07c88abf4b434ae7348970c9f02208f724fbdbb4c
+size 21725205111

checkpoint-87/global_step87/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f15eb234b5d2535de1c90ebaa15957108ebb58dd18d19e4160f8f6e6f413492
+size 21725205111

checkpoint-87/global_step87/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a2c61628b310c63934ceb4cec5b83577d0227f9f766867eb3b59b25bdf3d4d62
+size 21725204535

checkpoint-87/global_step87/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c57109fa782567ba8ae332992d54c6d5fc3ef338c1a4af21ca98a5c0912a07a
+size 14483551747

checkpoint-87/latest

@@ -0,0 +1 @@
+global_step87

checkpoint-87/model-00001-of-00003.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58d3ad8b1fe0901cec33dc9668265daa215b322b436d771ea44d37b63589ee4b
+size 4943162336