Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+checkpoint-116/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-174/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-232/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+checkpoint-58/tokenizer.json filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

adapter_config.json

@@ -0,0 +1,37 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-32B-Instruct",
+  "bias": "none",
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": null,
+  "inference_mode": false,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 512,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 256,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "up_proj",
+    "o_proj",
+    "gate_proj",
+    "v_proj",
+    "q_proj",
+    "k_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoint-116/README.md

@@ -0,0 +1,202 @@
+---
+base_model: Qwen/Qwen2.5-32B-Instruct
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.14.0

checkpoint-116/adapter_config.json

@@ -0,0 +1,37 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-32B-Instruct",
+  "bias": "none",
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": null,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 512,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 256,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "k_proj",
+    "up_proj",
+    "gate_proj",
+    "o_proj",
+    "v_proj",
+    "down_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

checkpoint-116/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4aa83eacd6470e1fdb03a43533501c5efb5765a9ed1a7ea570de0fd6153c934c
+size 4295091864

checkpoint-116/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

checkpoint-116/global_step116/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3520c08e4ed88d6250827533ae77df6d5dbffa403116785c56c1c2b9afbf783c
+size 6442479132

checkpoint-116/global_step116/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cd934c152c807dddae9370fc927e14a47b50a8a7744a8b6761a29fcc686cb24
+size 6442479260

checkpoint-116/global_step116/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e5710b74ad518c8f1b38abd4dfb6bfbcfff0218358cef97583ad6295d5423fc
+size 6442479260

checkpoint-116/global_step116/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:b4f763b2ae473c5ed902c5bd1270e3648fa9a70861dde6c3ca89bb1e10fec37d
+size 6442479260

checkpoint-116/global_step116/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:90e9f196cea396bee29a90ca35ccb1d4083c9f4ff2956a7b5680d4ce4e46246e
+size 4792285125

checkpoint-116/latest

@@ -0,0 +1 @@
+global_step116

checkpoint-116/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87c86454d7f017a866f1ea178cb53c6bb42fd42ae1723eeb01b11e0600cc2aa0
+size 15024

checkpoint-116/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a8812aedbf00eed6b869c2977f57d253519c60802d532e7166a84706d7ca017
+size 15024

checkpoint-116/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7079af4cbe946f6c9e6ba5bc98412d8e341471adf43beeea2dc6aaa22f890cfb
+size 15024

checkpoint-116/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a41e5fbc547c7aa6d1a21ee147d91edb4c7e7539f03c9fce336efdc5b6ee4b7
+size 15024

checkpoint-116/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c41db4bea64b34da7fc06691a2d02adc61955b3c8d84a3a59878aeb40d48e61c
+size 1064

checkpoint-116/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-116/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

checkpoint-116/tokenizer_config.json

@@ -0,0 +1,208 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-116/trainer_state.json

@@ -0,0 +1,845 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 116,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.017241379310344827,
+      "grad_norm": 16.797964096069336,
+      "learning_rate": 5.0000000000000004e-08,
+      "loss": 3.4811,
+      "step": 1
+    },
+    {
+      "epoch": 0.034482758620689655,
+      "grad_norm": 16.552379608154297,
+      "learning_rate": 1.0000000000000001e-07,
+      "loss": 3.4256,
+      "step": 2
+    },
+    {
+      "epoch": 0.05172413793103448,
+      "grad_norm": 16.94522476196289,
+      "learning_rate": 1.5000000000000002e-07,
+      "loss": 3.4624,
+      "step": 3
+    },
+    {
+      "epoch": 0.06896551724137931,
+      "grad_norm": 16.51656723022461,
+      "learning_rate": 2.0000000000000002e-07,
+      "loss": 3.4428,
+      "step": 4
+    },
+    {
+      "epoch": 0.08620689655172414,
+      "grad_norm": 16.90593910217285,
+      "learning_rate": 2.5000000000000004e-07,
+      "loss": 3.4053,
+      "step": 5
+    },
+    {
+      "epoch": 0.10344827586206896,
+      "grad_norm": 16.83192253112793,
+      "learning_rate": 3.0000000000000004e-07,
+      "loss": 3.506,
+      "step": 6
+    },
+    {
+      "epoch": 0.1206896551724138,
+      "grad_norm": 16.592622756958008,
+      "learning_rate": 3.5000000000000004e-07,
+      "loss": 3.4454,
+      "step": 7
+    },
+    {
+      "epoch": 0.13793103448275862,
+      "grad_norm": 16.566585540771484,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 3.4268,
+      "step": 8
+    },
+    {
+      "epoch": 0.15517241379310345,
+      "grad_norm": 16.46024513244629,
+      "learning_rate": 4.5000000000000003e-07,
+      "loss": 3.4186,
+      "step": 9
+    },
+    {
+      "epoch": 0.1724137931034483,
+      "grad_norm": 16.710294723510742,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 3.4258,
+      "step": 10
+    },
+    {
+      "epoch": 0.1896551724137931,
+      "grad_norm": 16.718793869018555,
+      "learning_rate": 5.5e-07,
+      "loss": 3.3919,
+      "step": 11
+    },
+    {
+      "epoch": 0.20689655172413793,
+      "grad_norm": 15.603546142578125,
+      "learning_rate": 6.000000000000001e-07,
+      "loss": 3.3223,
+      "step": 12
+    },
+    {
+      "epoch": 0.22413793103448276,
+      "grad_norm": 16.184322357177734,
+      "learning_rate": 6.5e-07,
+      "loss": 3.3516,
+      "step": 13
+    },
+    {
+      "epoch": 0.2413793103448276,
+      "grad_norm": 15.28188419342041,
+      "learning_rate": 7.000000000000001e-07,
+      "loss": 3.2181,
+      "step": 14
+    },
+    {
+      "epoch": 0.25862068965517243,
+      "grad_norm": 14.98234748840332,
+      "learning_rate": 7.5e-07,
+      "loss": 3.1837,
+      "step": 15
+    },
+    {
+      "epoch": 0.27586206896551724,
+      "grad_norm": 15.273055076599121,
+      "learning_rate": 8.000000000000001e-07,
+      "loss": 3.2103,
+      "step": 16
+    },
+    {
+      "epoch": 0.29310344827586204,
+      "grad_norm": 14.799996376037598,
+      "learning_rate": 8.500000000000001e-07,
+      "loss": 3.1195,
+      "step": 17
+    },
+    {
+      "epoch": 0.3103448275862069,
+      "grad_norm": 13.851456642150879,
+      "learning_rate": 9.000000000000001e-07,
+      "loss": 3.0261,
+      "step": 18
+    },
+    {
+      "epoch": 0.3275862068965517,
+      "grad_norm": 12.985479354858398,
+      "learning_rate": 9.500000000000001e-07,
+      "loss": 2.8995,
+      "step": 19
+    },
+    {
+      "epoch": 0.3448275862068966,
+      "grad_norm": 12.569958686828613,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 2.8289,
+      "step": 20
+    },
+    {
+      "epoch": 0.3620689655172414,
+      "grad_norm": 11.687292098999023,
+      "learning_rate": 1.0500000000000001e-06,
+      "loss": 2.7955,
+      "step": 21
+    },
+    {
+      "epoch": 0.3793103448275862,
+      "grad_norm": 10.375764846801758,
+      "learning_rate": 1.1e-06,
+      "loss": 2.6422,
+      "step": 22
+    },
+    {
+      "epoch": 0.39655172413793105,
+      "grad_norm": 9.314571380615234,
+      "learning_rate": 1.1500000000000002e-06,
+      "loss": 2.5636,
+      "step": 23
+    },
+    {
+      "epoch": 0.41379310344827586,
+      "grad_norm": 8.794600486755371,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 2.4439,
+      "step": 24
+    },
+    {
+      "epoch": 0.43103448275862066,
+      "grad_norm": 8.352209091186523,
+      "learning_rate": 1.25e-06,
+      "loss": 2.351,
+      "step": 25
+    },
+    {
+      "epoch": 0.4482758620689655,
+      "grad_norm": 8.14919662475586,
+      "learning_rate": 1.3e-06,
+      "loss": 2.2486,
+      "step": 26
+    },
+    {
+      "epoch": 0.46551724137931033,
+      "grad_norm": 8.511932373046875,
+      "learning_rate": 1.3500000000000002e-06,
+      "loss": 2.201,
+      "step": 27
+    },
+    {
+      "epoch": 0.4827586206896552,
+      "grad_norm": 8.55715274810791,
+      "learning_rate": 1.4000000000000001e-06,
+      "loss": 2.0727,
+      "step": 28
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 8.575194358825684,
+      "learning_rate": 1.45e-06,
+      "loss": 1.9386,
+      "step": 29
+    },
+    {
+      "epoch": 0.5172413793103449,
+      "grad_norm": 8.735200881958008,
+      "learning_rate": 1.5e-06,
+      "loss": 1.8335,
+      "step": 30
+    },
+    {
+      "epoch": 0.5344827586206896,
+      "grad_norm": 8.932766914367676,
+      "learning_rate": 1.5500000000000002e-06,
+      "loss": 1.7265,
+      "step": 31
+    },
+    {
+      "epoch": 0.5517241379310345,
+      "grad_norm": 9.010940551757812,
+      "learning_rate": 1.6000000000000001e-06,
+      "loss": 1.5886,
+      "step": 32
+    },
+    {
+      "epoch": 0.5689655172413793,
+      "grad_norm": 9.089744567871094,
+      "learning_rate": 1.6500000000000003e-06,
+      "loss": 1.4379,
+      "step": 33
+    },
+    {
+      "epoch": 0.5862068965517241,
+      "grad_norm": 9.299127578735352,
+      "learning_rate": 1.7000000000000002e-06,
+      "loss": 1.2766,
+      "step": 34
+    },
+    {
+      "epoch": 0.603448275862069,
+      "grad_norm": 9.707971572875977,
+      "learning_rate": 1.75e-06,
+      "loss": 1.1075,
+      "step": 35
+    },
+    {
+      "epoch": 0.6206896551724138,
+      "grad_norm": 9.807348251342773,
+      "learning_rate": 1.8000000000000001e-06,
+      "loss": 0.9345,
+      "step": 36
+    },
+    {
+      "epoch": 0.6379310344827587,
+      "grad_norm": 9.576728820800781,
+      "learning_rate": 1.85e-06,
+      "loss": 0.7481,
+      "step": 37
+    },
+    {
+      "epoch": 0.6551724137931034,
+      "grad_norm": 8.874764442443848,
+      "learning_rate": 1.9000000000000002e-06,
+      "loss": 0.5921,
+      "step": 38
+    },
+    {
+      "epoch": 0.6724137931034483,
+      "grad_norm": 7.2561354637146,
+      "learning_rate": 1.9500000000000004e-06,
+      "loss": 0.4217,
+      "step": 39
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 4.964897155761719,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 0.3012,
+      "step": 40
+    },
+    {
+      "epoch": 0.7068965517241379,
+      "grad_norm": 3.965514659881592,
+      "learning_rate": 2.05e-06,
+      "loss": 0.2359,
+      "step": 41
+    },
+    {
+      "epoch": 0.7241379310344828,
+      "grad_norm": 3.0960378646850586,
+      "learning_rate": 2.1000000000000002e-06,
+      "loss": 0.1738,
+      "step": 42
+    },
+    {
+      "epoch": 0.7413793103448276,
+      "grad_norm": 2.1212306022644043,
+      "learning_rate": 2.15e-06,
+      "loss": 0.1364,
+      "step": 43
+    },
+    {
+      "epoch": 0.7586206896551724,
+      "grad_norm": 1.155745506286621,
+      "learning_rate": 2.2e-06,
+      "loss": 0.1012,
+      "step": 44
+    },
+    {
+      "epoch": 0.7758620689655172,
+      "grad_norm": 0.7050064206123352,
+      "learning_rate": 2.25e-06,
+      "loss": 0.095,
+      "step": 45
+    },
+    {
+      "epoch": 0.7931034482758621,
+      "grad_norm": 0.43830573558807373,
+      "learning_rate": 2.3000000000000004e-06,
+      "loss": 0.0801,
+      "step": 46
+    },
+    {
+      "epoch": 0.8103448275862069,
+      "grad_norm": 0.3802882432937622,
+      "learning_rate": 2.35e-06,
+      "loss": 0.0827,
+      "step": 47
+    },
+    {
+      "epoch": 0.8275862068965517,
+      "grad_norm": 0.3097267746925354,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 0.0762,
+      "step": 48
+    },
+    {
+      "epoch": 0.8448275862068966,
+      "grad_norm": 0.2734082341194153,
+      "learning_rate": 2.4500000000000003e-06,
+      "loss": 0.0749,
+      "step": 49
+    },
+    {
+      "epoch": 0.8620689655172413,
+      "grad_norm": 0.2831459641456604,
+      "learning_rate": 2.5e-06,
+      "loss": 0.0762,
+      "step": 50
+    },
+    {
+      "epoch": 0.8793103448275862,
+      "grad_norm": 0.2537994086742401,
+      "learning_rate": 2.55e-06,
+      "loss": 0.072,
+      "step": 51
+    },
+    {
+      "epoch": 0.896551724137931,
+      "grad_norm": 0.3006448745727539,
+      "learning_rate": 2.6e-06,
+      "loss": 0.073,
+      "step": 52
+    },
+    {
+      "epoch": 0.9137931034482759,
+      "grad_norm": 0.24919214844703674,
+      "learning_rate": 2.6500000000000005e-06,
+      "loss": 0.0699,
+      "step": 53
+    },
+    {
+      "epoch": 0.9310344827586207,
+      "grad_norm": 0.23921510577201843,
+      "learning_rate": 2.7000000000000004e-06,
+      "loss": 0.0647,
+      "step": 54
+    },
+    {
+      "epoch": 0.9482758620689655,
+      "grad_norm": 0.1967734843492508,
+      "learning_rate": 2.7500000000000004e-06,
+      "loss": 0.0711,
+      "step": 55
+    },
+    {
+      "epoch": 0.9655172413793104,
+      "grad_norm": 0.1832527369260788,
+      "learning_rate": 2.8000000000000003e-06,
+      "loss": 0.0686,
+      "step": 56
+    },
+    {
+      "epoch": 0.9827586206896551,
+      "grad_norm": 0.16995869576931,
+      "learning_rate": 2.85e-06,
+      "loss": 0.0644,
+      "step": 57
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.1822829246520996,
+      "learning_rate": 2.9e-06,
+      "loss": 0.0669,
+      "step": 58
+    },
+    {
+      "epoch": 1.0172413793103448,
+      "grad_norm": 0.17660750448703766,
+      "learning_rate": 2.95e-06,
+      "loss": 0.0668,
+      "step": 59
+    },
+    {
+      "epoch": 1.0344827586206897,
+      "grad_norm": 0.12004642933607101,
+      "learning_rate": 3e-06,
+      "loss": 0.0588,
+      "step": 60
+    },
+    {
+      "epoch": 1.0517241379310345,
+      "grad_norm": 0.15699979662895203,
+      "learning_rate": 3.05e-06,
+      "loss": 0.0631,
+      "step": 61
+    },
+    {
+      "epoch": 1.0689655172413792,
+      "grad_norm": 0.15650232136249542,
+      "learning_rate": 3.1000000000000004e-06,
+      "loss": 0.0596,
+      "step": 62
+    },
+    {
+      "epoch": 1.0862068965517242,
+      "grad_norm": 0.1188632994890213,
+      "learning_rate": 3.1500000000000003e-06,
+      "loss": 0.0614,
+      "step": 63
+    },
+    {
+      "epoch": 1.103448275862069,
+      "grad_norm": 0.11891665309667587,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 0.0636,
+      "step": 64
+    },
+    {
+      "epoch": 1.1206896551724137,
+      "grad_norm": 0.12215171754360199,
+      "learning_rate": 3.2500000000000002e-06,
+      "loss": 0.0567,
+      "step": 65
+    },
+    {
+      "epoch": 1.1379310344827587,
+      "grad_norm": 0.10582094639539719,
+      "learning_rate": 3.3000000000000006e-06,
+      "loss": 0.0588,
+      "step": 66
+    },
+    {
+      "epoch": 1.1551724137931034,
+      "grad_norm": 0.11556132137775421,
+      "learning_rate": 3.3500000000000005e-06,
+      "loss": 0.0628,
+      "step": 67
+    },
+    {
+      "epoch": 1.1724137931034484,
+      "grad_norm": 0.12296544015407562,
+      "learning_rate": 3.4000000000000005e-06,
+      "loss": 0.0623,
+      "step": 68
+    },
+    {
+      "epoch": 1.1896551724137931,
+      "grad_norm": 0.1281358152627945,
+      "learning_rate": 3.45e-06,
+      "loss": 0.0626,
+      "step": 69
+    },
+    {
+      "epoch": 1.206896551724138,
+      "grad_norm": 0.1277759075164795,
+      "learning_rate": 3.5e-06,
+      "loss": 0.0592,
+      "step": 70
+    },
+    {
+      "epoch": 1.2241379310344827,
+      "grad_norm": 0.09552627056837082,
+      "learning_rate": 3.5500000000000003e-06,
+      "loss": 0.0595,
+      "step": 71
+    },
+    {
+      "epoch": 1.2413793103448276,
+      "grad_norm": 0.11053085327148438,
+      "learning_rate": 3.6000000000000003e-06,
+      "loss": 0.0546,
+      "step": 72
+    },
+    {
+      "epoch": 1.2586206896551724,
+      "grad_norm": 0.09386970102787018,
+      "learning_rate": 3.65e-06,
+      "loss": 0.0556,
+      "step": 73
+    },
+    {
+      "epoch": 1.2758620689655173,
+      "grad_norm": 0.0934402197599411,
+      "learning_rate": 3.7e-06,
+      "loss": 0.0563,
+      "step": 74
+    },
+    {
+      "epoch": 1.293103448275862,
+      "grad_norm": 0.09964020550251007,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.0569,
+      "step": 75
+    },
+    {
+      "epoch": 1.3103448275862069,
+      "grad_norm": 0.08830665796995163,
+      "learning_rate": 3.8000000000000005e-06,
+      "loss": 0.0578,
+      "step": 76
+    },
+    {
+      "epoch": 1.3275862068965516,
+      "grad_norm": 0.13536307215690613,
+      "learning_rate": 3.85e-06,
+      "loss": 0.0569,
+      "step": 77
+    },
+    {
+      "epoch": 1.3448275862068966,
+      "grad_norm": 0.10773950815200806,
+      "learning_rate": 3.900000000000001e-06,
+      "loss": 0.0581,
+      "step": 78
+    },
+    {
+      "epoch": 1.3620689655172413,
+      "grad_norm": 0.09468439221382141,
+      "learning_rate": 3.95e-06,
+      "loss": 0.0549,
+      "step": 79
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "grad_norm": 0.08775551617145538,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 0.0594,
+      "step": 80
+    },
+    {
+      "epoch": 1.396551724137931,
+      "grad_norm": 0.12008150666952133,
+      "learning_rate": 4.05e-06,
+      "loss": 0.057,
+      "step": 81
+    },
+    {
+      "epoch": 1.4137931034482758,
+      "grad_norm": 0.12070683389902115,
+      "learning_rate": 4.1e-06,
+      "loss": 0.0549,
+      "step": 82
+    },
+    {
+      "epoch": 1.4310344827586206,
+      "grad_norm": 0.1037198081612587,
+      "learning_rate": 4.15e-06,
+      "loss": 0.0554,
+      "step": 83
+    },
+    {
+      "epoch": 1.4482758620689655,
+      "grad_norm": 0.14529870450496674,
+      "learning_rate": 4.2000000000000004e-06,
+      "loss": 0.0549,
+      "step": 84
+    },
+    {
+      "epoch": 1.4655172413793103,
+      "grad_norm": 0.0954233855009079,
+      "learning_rate": 4.25e-06,
+      "loss": 0.0556,
+      "step": 85
+    },
+    {
+      "epoch": 1.4827586206896552,
+      "grad_norm": 0.08504101634025574,
+      "learning_rate": 4.3e-06,
+      "loss": 0.0505,
+      "step": 86
+    },
+    {
+      "epoch": 1.5,
+      "grad_norm": 0.15293122828006744,
+      "learning_rate": 4.350000000000001e-06,
+      "loss": 0.0577,
+      "step": 87
+    },
+    {
+      "epoch": 1.5172413793103448,
+      "grad_norm": 0.10908783227205276,
+      "learning_rate": 4.4e-06,
+      "loss": 0.0556,
+      "step": 88
+    },
+    {
+      "epoch": 1.5344827586206895,
+      "grad_norm": 0.12018983066082001,
+      "learning_rate": 4.450000000000001e-06,
+      "loss": 0.0554,
+      "step": 89
+    },
+    {
+      "epoch": 1.5517241379310345,
+      "grad_norm": 0.1018645316362381,
+      "learning_rate": 4.5e-06,
+      "loss": 0.0548,
+      "step": 90
+    },
+    {
+      "epoch": 1.5689655172413794,
+      "grad_norm": 0.09623338282108307,
+      "learning_rate": 4.5500000000000005e-06,
+      "loss": 0.0566,
+      "step": 91
+    },
+    {
+      "epoch": 1.5862068965517242,
+      "grad_norm": 0.09007120132446289,
+      "learning_rate": 4.600000000000001e-06,
+      "loss": 0.0552,
+      "step": 92
+    },
+    {
+      "epoch": 1.603448275862069,
+      "grad_norm": 0.07549538463354111,
+      "learning_rate": 4.65e-06,
+      "loss": 0.056,
+      "step": 93
+    },
+    {
+      "epoch": 1.6206896551724137,
+      "grad_norm": 0.14191967248916626,
+      "learning_rate": 4.7e-06,
+      "loss": 0.0547,
+      "step": 94
+    },
+    {
+      "epoch": 1.6379310344827587,
+      "grad_norm": 0.13249421119689941,
+      "learning_rate": 4.75e-06,
+      "loss": 0.0529,
+      "step": 95
+    },
+    {
+      "epoch": 1.6551724137931034,
+      "grad_norm": 0.09079012274742126,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 0.0537,
+      "step": 96
+    },
+    {
+      "epoch": 1.6724137931034484,
+      "grad_norm": 0.08319538831710815,
+      "learning_rate": 4.85e-06,
+      "loss": 0.0523,
+      "step": 97
+    },
+    {
+      "epoch": 1.6896551724137931,
+      "grad_norm": 0.09330669790506363,
+      "learning_rate": 4.9000000000000005e-06,
+      "loss": 0.0515,
+      "step": 98
+    },
+    {
+      "epoch": 1.706896551724138,
+      "grad_norm": 0.1553690880537033,
+      "learning_rate": 4.95e-06,
+      "loss": 0.0539,
+      "step": 99
+    },
+    {
+      "epoch": 1.7241379310344827,
+      "grad_norm": 0.10644665360450745,
+      "learning_rate": 5e-06,
+      "loss": 0.0537,
+      "step": 100
+    },
+    {
+      "epoch": 1.7413793103448276,
+      "grad_norm": 0.10171601176261902,
+      "learning_rate": 4.999799414013322e-06,
+      "loss": 0.0526,
+      "step": 101
+    },
+    {
+      "epoch": 1.7586206896551724,
+      "grad_norm": 0.08432668447494507,
+      "learning_rate": 4.999197688241076e-06,
+      "loss": 0.0525,
+      "step": 102
+    },
+    {
+      "epoch": 1.7758620689655173,
+      "grad_norm": 0.08552516251802444,
+      "learning_rate": 4.998194919241471e-06,
+      "loss": 0.0528,
+      "step": 103
+    },
+    {
+      "epoch": 1.793103448275862,
+      "grad_norm": 0.10473164916038513,
+      "learning_rate": 4.996791267927632e-06,
+      "loss": 0.0509,
+      "step": 104
+    },
+    {
+      "epoch": 1.8103448275862069,
+      "grad_norm": 0.08742501586675644,
+      "learning_rate": 4.994986959541788e-06,
+      "loss": 0.0527,
+      "step": 105
+    },
+    {
+      "epoch": 1.8275862068965516,
+      "grad_norm": 0.07423796504735947,
+      "learning_rate": 4.9927822836191185e-06,
+      "loss": 0.0532,
+      "step": 106
+    },
+    {
+      "epoch": 1.8448275862068966,
+      "grad_norm": 0.0921708270907402,
+      "learning_rate": 4.990177593941303e-06,
+      "loss": 0.0545,
+      "step": 107
+    },
+    {
+      "epoch": 1.8620689655172413,
+      "grad_norm": 0.08664074540138245,
+      "learning_rate": 4.987173308479738e-06,
+      "loss": 0.0505,
+      "step": 108
+    },
+    {
+      "epoch": 1.8793103448275863,
+      "grad_norm": 0.09824781864881516,
+      "learning_rate": 4.9837699093284765e-06,
+      "loss": 0.053,
+      "step": 109
+    },
+    {
+      "epoch": 1.896551724137931,
+      "grad_norm": 0.0782749354839325,
+      "learning_rate": 4.9799679426268575e-06,
+      "loss": 0.0525,
+      "step": 110
+    },
+    {
+      "epoch": 1.9137931034482758,
+      "grad_norm": 0.09107070416212082,
+      "learning_rate": 4.975768018471877e-06,
+      "loss": 0.0536,
+      "step": 111
+    },
+    {
+      "epoch": 1.9310344827586206,
+      "grad_norm": 0.07815398275852203,
+      "learning_rate": 4.971170810820279e-06,
+      "loss": 0.0486,
+      "step": 112
+    },
+    {
+      "epoch": 1.9482758620689655,
+      "grad_norm": 0.08175615966320038,
+      "learning_rate": 4.966177057380409e-06,
+      "loss": 0.0545,
+      "step": 113
+    },
+    {
+      "epoch": 1.9655172413793105,
+      "grad_norm": 0.08906937390565872,
+      "learning_rate": 4.960787559493836e-06,
+      "loss": 0.0478,
+      "step": 114
+    },
+    {
+      "epoch": 1.9827586206896552,
+      "grad_norm": 0.08419270813465118,
+      "learning_rate": 4.955003182006761e-06,
+      "loss": 0.0528,
+      "step": 115
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.10093575716018677,
+      "learning_rate": 4.948824853131237e-06,
+      "loss": 0.049,
+      "step": 116
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 348,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 6,
+  "save_steps": 58,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 4.5960715144989245e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-116/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4a8bf738cfba89eb19d24178060f165bf0f68185db34cebf3b311b7c61eebc5
+size 7992

checkpoint-116/zero_to_fp32.py

@@ -0,0 +1,604 @@
+#!/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, exclude_frozen_parameters):
+    """
+    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,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+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,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _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,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _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, exclude_frozen_parameters=False):
+    """
+    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``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    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, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    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``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    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("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    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,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)

checkpoint-174/README.md

@@ -0,0 +1,202 @@
+---
+base_model: Qwen/Qwen2.5-32B-Instruct
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.14.0

checkpoint-174/adapter_config.json

@@ -0,0 +1,37 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "Qwen/Qwen2.5-32B-Instruct",
+  "bias": "none",
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": null,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 512,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 256,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "q_proj",
+    "k_proj",
+    "up_proj",
+    "gate_proj",
+    "o_proj",
+    "v_proj",
+    "down_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

checkpoint-174/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:508ce93ebd026732662a4020a72144553e1d3bb16541df03ca222ddd4a3e1b38
+size 4295091864

checkpoint-174/added_tokens.json

@@ -0,0 +1,24 @@
+{
+  "</tool_call>": 151658,
+  "<tool_call>": 151657,
+  "<|box_end|>": 151649,
+  "<|box_start|>": 151648,
+  "<|endoftext|>": 151643,
+  "<|file_sep|>": 151664,
+  "<|fim_middle|>": 151660,
+  "<|fim_pad|>": 151662,
+  "<|fim_prefix|>": 151659,
+  "<|fim_suffix|>": 151661,
+  "<|im_end|>": 151645,
+  "<|im_start|>": 151644,
+  "<|image_pad|>": 151655,
+  "<|object_ref_end|>": 151647,
+  "<|object_ref_start|>": 151646,
+  "<|quad_end|>": 151651,
+  "<|quad_start|>": 151650,
+  "<|repo_name|>": 151663,
+  "<|video_pad|>": 151656,
+  "<|vision_end|>": 151653,
+  "<|vision_pad|>": 151654,
+  "<|vision_start|>": 151652
+}

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 sha256:914f67c2093e53be18dfc6082eb8790669bffa0d6b0ef66bdbccecf58eade7b8
+size 6442479132

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:c7f103bee0083ba5128bd04e887e71beff178ab9f882b82b5621511c53f3d43c
+size 6442479260

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 sha256:52b9770bb4620ec1fd776d150fa3598f1c8e303a20871b78908ae7fe9ffd15f3
+size 6442479260

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 sha256:32030d82b87cda11f83bc628146fd45e2799541fc64ea035f3da09f7911285ab
+size 6442479260

checkpoint-174/global_step174/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35da2832cad1bde78be515f9002e7a271716ece52df59738df87c058e89e75d9
+size 4792285125

checkpoint-174/latest

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

checkpoint-174/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cd34d9292ef865c1e4c8de6fbe5ea7fe66c1c3dc127273f0abbf87f5633407b
+size 15024

checkpoint-174/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27bc17c8042c630148ea6c52c345acc67f18276d0a34b06113ef5f464b1e30de
+size 15024

checkpoint-174/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da8c1e3e9688cec2c31335cfa71384d949998d7050132f4cc5677add652cbbfb
+size 15024

checkpoint-174/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c20600f50b44e29cae0af11cac792150b3d9ee120148241badd3368aedfcc769
+size 15024

checkpoint-174/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37ddddc0c10904285aaee6ac2fac97341c7f05befc3f8aa621649bd8ac0ff104
+size 1064

checkpoint-174/special_tokens_map.json

@@ -0,0 +1,31 @@
+{
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "eos_token": {
+    "content": "<|im_end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-174/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
+size 11421896

checkpoint-174/tokenizer_config.json

@@ -0,0 +1,208 @@
+{
+  "add_bos_token": false,
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "151643": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151644": {
+      "content": "<|im_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151645": {
+      "content": "<|im_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151646": {
+      "content": "<|object_ref_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151647": {
+      "content": "<|object_ref_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151648": {
+      "content": "<|box_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151649": {
+      "content": "<|box_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151650": {
+      "content": "<|quad_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151651": {
+      "content": "<|quad_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151652": {
+      "content": "<|vision_start|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151653": {
+      "content": "<|vision_end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151654": {
+      "content": "<|vision_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151655": {
+      "content": "<|image_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151656": {
+      "content": "<|video_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "151657": {
+      "content": "<tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151658": {
+      "content": "</tool_call>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151659": {
+      "content": "<|fim_prefix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151660": {
+      "content": "<|fim_middle|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151661": {
+      "content": "<|fim_suffix|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151662": {
+      "content": "<|fim_pad|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151663": {
+      "content": "<|repo_name|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    },
+    "151664": {
+      "content": "<|file_sep|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": false
+    }
+  },
+  "additional_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "bos_token": null,
+  "chat_template": "{%- if tools %}\n    {{- '<|im_start|>system\\n' }}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- messages[0]['content'] }}\n    {%- else %}\n        {{- 'You are Qwen, created by Alibaba Cloud. You are a helpful assistant.' }}\n    {%- endif %}\n    {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n    {%- for tool in tools %}\n        {{- \"\\n\" }}\n        {{- tool | tojson }}\n    {%- endfor %}\n    {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n    {%- if messages[0]['role'] == 'system' %}\n        {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n    {%- else %}\n        {{- '<|im_start|>system\\nYou are Qwen, created by Alibaba Cloud. You are a helpful assistant.<|im_end|>\\n' }}\n    {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n    {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n        {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n    {%- elif message.role == \"assistant\" %}\n        {{- '<|im_start|>' + message.role }}\n        {%- if message.content %}\n            {{- '\\n' + message.content }}\n        {%- endif %}\n        {%- for tool_call in message.tool_calls %}\n            {%- if tool_call.function is defined %}\n                {%- set tool_call = tool_call.function %}\n            {%- endif %}\n            {{- '\\n<tool_call>\\n{\"name\": \"' }}\n            {{- tool_call.name }}\n            {{- '\", \"arguments\": ' }}\n            {{- tool_call.arguments | tojson }}\n            {{- '}\\n</tool_call>' }}\n        {%- endfor %}\n        {{- '<|im_end|>\\n' }}\n    {%- elif message.role == \"tool\" %}\n        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n            {{- '<|im_start|>user' }}\n        {%- endif %}\n        {{- '\\n<tool_response>\\n' }}\n        {{- message.content }}\n        {{- '\\n</tool_response>' }}\n        {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n            {{- '<|im_end|>\\n' }}\n        {%- endif %}\n    {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n    {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|im_end|>",
+  "errors": "replace",
+  "extra_special_tokens": {},
+  "model_max_length": 131072,
+  "pad_token": "<|endoftext|>",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}

checkpoint-174/trainer_state.json

@@ -0,0 +1,1251 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 174,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.017241379310344827,
+      "grad_norm": 16.797964096069336,
+      "learning_rate": 5.0000000000000004e-08,
+      "loss": 3.4811,
+      "step": 1
+    },
+    {
+      "epoch": 0.034482758620689655,
+      "grad_norm": 16.552379608154297,
+      "learning_rate": 1.0000000000000001e-07,
+      "loss": 3.4256,
+      "step": 2
+    },
+    {
+      "epoch": 0.05172413793103448,
+      "grad_norm": 16.94522476196289,
+      "learning_rate": 1.5000000000000002e-07,
+      "loss": 3.4624,
+      "step": 3
+    },
+    {
+      "epoch": 0.06896551724137931,
+      "grad_norm": 16.51656723022461,
+      "learning_rate": 2.0000000000000002e-07,
+      "loss": 3.4428,
+      "step": 4
+    },
+    {
+      "epoch": 0.08620689655172414,
+      "grad_norm": 16.90593910217285,
+      "learning_rate": 2.5000000000000004e-07,
+      "loss": 3.4053,
+      "step": 5
+    },
+    {
+      "epoch": 0.10344827586206896,
+      "grad_norm": 16.83192253112793,
+      "learning_rate": 3.0000000000000004e-07,
+      "loss": 3.506,
+      "step": 6
+    },
+    {
+      "epoch": 0.1206896551724138,
+      "grad_norm": 16.592622756958008,
+      "learning_rate": 3.5000000000000004e-07,
+      "loss": 3.4454,
+      "step": 7
+    },
+    {
+      "epoch": 0.13793103448275862,
+      "grad_norm": 16.566585540771484,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 3.4268,
+      "step": 8
+    },
+    {
+      "epoch": 0.15517241379310345,
+      "grad_norm": 16.46024513244629,
+      "learning_rate": 4.5000000000000003e-07,
+      "loss": 3.4186,
+      "step": 9
+    },
+    {
+      "epoch": 0.1724137931034483,
+      "grad_norm": 16.710294723510742,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 3.4258,
+      "step": 10
+    },
+    {
+      "epoch": 0.1896551724137931,
+      "grad_norm": 16.718793869018555,
+      "learning_rate": 5.5e-07,
+      "loss": 3.3919,
+      "step": 11
+    },
+    {
+      "epoch": 0.20689655172413793,
+      "grad_norm": 15.603546142578125,
+      "learning_rate": 6.000000000000001e-07,
+      "loss": 3.3223,
+      "step": 12
+    },
+    {
+      "epoch": 0.22413793103448276,
+      "grad_norm": 16.184322357177734,
+      "learning_rate": 6.5e-07,
+      "loss": 3.3516,
+      "step": 13
+    },
+    {
+      "epoch": 0.2413793103448276,
+      "grad_norm": 15.28188419342041,
+      "learning_rate": 7.000000000000001e-07,
+      "loss": 3.2181,
+      "step": 14
+    },
+    {
+      "epoch": 0.25862068965517243,
+      "grad_norm": 14.98234748840332,
+      "learning_rate": 7.5e-07,
+      "loss": 3.1837,
+      "step": 15
+    },
+    {
+      "epoch": 0.27586206896551724,
+      "grad_norm": 15.273055076599121,
+      "learning_rate": 8.000000000000001e-07,
+      "loss": 3.2103,
+      "step": 16
+    },
+    {
+      "epoch": 0.29310344827586204,
+      "grad_norm": 14.799996376037598,
+      "learning_rate": 8.500000000000001e-07,
+      "loss": 3.1195,
+      "step": 17
+    },
+    {
+      "epoch": 0.3103448275862069,
+      "grad_norm": 13.851456642150879,
+      "learning_rate": 9.000000000000001e-07,
+      "loss": 3.0261,
+      "step": 18
+    },
+    {
+      "epoch": 0.3275862068965517,
+      "grad_norm": 12.985479354858398,
+      "learning_rate": 9.500000000000001e-07,
+      "loss": 2.8995,
+      "step": 19
+    },
+    {
+      "epoch": 0.3448275862068966,
+      "grad_norm": 12.569958686828613,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 2.8289,
+      "step": 20
+    },
+    {
+      "epoch": 0.3620689655172414,
+      "grad_norm": 11.687292098999023,
+      "learning_rate": 1.0500000000000001e-06,
+      "loss": 2.7955,
+      "step": 21
+    },
+    {
+      "epoch": 0.3793103448275862,
+      "grad_norm": 10.375764846801758,
+      "learning_rate": 1.1e-06,
+      "loss": 2.6422,
+      "step": 22
+    },
+    {
+      "epoch": 0.39655172413793105,
+      "grad_norm": 9.314571380615234,
+      "learning_rate": 1.1500000000000002e-06,
+      "loss": 2.5636,
+      "step": 23
+    },
+    {
+      "epoch": 0.41379310344827586,
+      "grad_norm": 8.794600486755371,
+      "learning_rate": 1.2000000000000002e-06,
+      "loss": 2.4439,
+      "step": 24
+    },
+    {
+      "epoch": 0.43103448275862066,
+      "grad_norm": 8.352209091186523,
+      "learning_rate": 1.25e-06,
+      "loss": 2.351,
+      "step": 25
+    },
+    {
+      "epoch": 0.4482758620689655,
+      "grad_norm": 8.14919662475586,
+      "learning_rate": 1.3e-06,
+      "loss": 2.2486,
+      "step": 26
+    },
+    {
+      "epoch": 0.46551724137931033,
+      "grad_norm": 8.511932373046875,
+      "learning_rate": 1.3500000000000002e-06,
+      "loss": 2.201,
+      "step": 27
+    },
+    {
+      "epoch": 0.4827586206896552,
+      "grad_norm": 8.55715274810791,
+      "learning_rate": 1.4000000000000001e-06,
+      "loss": 2.0727,
+      "step": 28
+    },
+    {
+      "epoch": 0.5,
+      "grad_norm": 8.575194358825684,
+      "learning_rate": 1.45e-06,
+      "loss": 1.9386,
+      "step": 29
+    },
+    {
+      "epoch": 0.5172413793103449,
+      "grad_norm": 8.735200881958008,
+      "learning_rate": 1.5e-06,
+      "loss": 1.8335,
+      "step": 30
+    },
+    {
+      "epoch": 0.5344827586206896,
+      "grad_norm": 8.932766914367676,
+      "learning_rate": 1.5500000000000002e-06,
+      "loss": 1.7265,
+      "step": 31
+    },
+    {
+      "epoch": 0.5517241379310345,
+      "grad_norm": 9.010940551757812,
+      "learning_rate": 1.6000000000000001e-06,
+      "loss": 1.5886,
+      "step": 32
+    },
+    {
+      "epoch": 0.5689655172413793,
+      "grad_norm": 9.089744567871094,
+      "learning_rate": 1.6500000000000003e-06,
+      "loss": 1.4379,
+      "step": 33
+    },
+    {
+      "epoch": 0.5862068965517241,
+      "grad_norm": 9.299127578735352,
+      "learning_rate": 1.7000000000000002e-06,
+      "loss": 1.2766,
+      "step": 34
+    },
+    {
+      "epoch": 0.603448275862069,
+      "grad_norm": 9.707971572875977,
+      "learning_rate": 1.75e-06,
+      "loss": 1.1075,
+      "step": 35
+    },
+    {
+      "epoch": 0.6206896551724138,
+      "grad_norm": 9.807348251342773,
+      "learning_rate": 1.8000000000000001e-06,
+      "loss": 0.9345,
+      "step": 36
+    },
+    {
+      "epoch": 0.6379310344827587,
+      "grad_norm": 9.576728820800781,
+      "learning_rate": 1.85e-06,
+      "loss": 0.7481,
+      "step": 37
+    },
+    {
+      "epoch": 0.6551724137931034,
+      "grad_norm": 8.874764442443848,
+      "learning_rate": 1.9000000000000002e-06,
+      "loss": 0.5921,
+      "step": 38
+    },
+    {
+      "epoch": 0.6724137931034483,
+      "grad_norm": 7.2561354637146,
+      "learning_rate": 1.9500000000000004e-06,
+      "loss": 0.4217,
+      "step": 39
+    },
+    {
+      "epoch": 0.6896551724137931,
+      "grad_norm": 4.964897155761719,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 0.3012,
+      "step": 40
+    },
+    {
+      "epoch": 0.7068965517241379,
+      "grad_norm": 3.965514659881592,
+      "learning_rate": 2.05e-06,
+      "loss": 0.2359,
+      "step": 41
+    },
+    {
+      "epoch": 0.7241379310344828,
+      "grad_norm": 3.0960378646850586,
+      "learning_rate": 2.1000000000000002e-06,
+      "loss": 0.1738,
+      "step": 42
+    },
+    {
+      "epoch": 0.7413793103448276,
+      "grad_norm": 2.1212306022644043,
+      "learning_rate": 2.15e-06,
+      "loss": 0.1364,
+      "step": 43
+    },
+    {
+      "epoch": 0.7586206896551724,
+      "grad_norm": 1.155745506286621,
+      "learning_rate": 2.2e-06,
+      "loss": 0.1012,
+      "step": 44
+    },
+    {
+      "epoch": 0.7758620689655172,
+      "grad_norm": 0.7050064206123352,
+      "learning_rate": 2.25e-06,
+      "loss": 0.095,
+      "step": 45
+    },
+    {
+      "epoch": 0.7931034482758621,
+      "grad_norm": 0.43830573558807373,
+      "learning_rate": 2.3000000000000004e-06,
+      "loss": 0.0801,
+      "step": 46
+    },
+    {
+      "epoch": 0.8103448275862069,
+      "grad_norm": 0.3802882432937622,
+      "learning_rate": 2.35e-06,
+      "loss": 0.0827,
+      "step": 47
+    },
+    {
+      "epoch": 0.8275862068965517,
+      "grad_norm": 0.3097267746925354,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 0.0762,
+      "step": 48
+    },
+    {
+      "epoch": 0.8448275862068966,
+      "grad_norm": 0.2734082341194153,
+      "learning_rate": 2.4500000000000003e-06,
+      "loss": 0.0749,
+      "step": 49
+    },
+    {
+      "epoch": 0.8620689655172413,
+      "grad_norm": 0.2831459641456604,
+      "learning_rate": 2.5e-06,
+      "loss": 0.0762,
+      "step": 50
+    },
+    {
+      "epoch": 0.8793103448275862,
+      "grad_norm": 0.2537994086742401,
+      "learning_rate": 2.55e-06,
+      "loss": 0.072,
+      "step": 51
+    },
+    {
+      "epoch": 0.896551724137931,
+      "grad_norm": 0.3006448745727539,
+      "learning_rate": 2.6e-06,
+      "loss": 0.073,
+      "step": 52
+    },
+    {
+      "epoch": 0.9137931034482759,
+      "grad_norm": 0.24919214844703674,
+      "learning_rate": 2.6500000000000005e-06,
+      "loss": 0.0699,
+      "step": 53
+    },
+    {
+      "epoch": 0.9310344827586207,
+      "grad_norm": 0.23921510577201843,
+      "learning_rate": 2.7000000000000004e-06,
+      "loss": 0.0647,
+      "step": 54
+    },
+    {
+      "epoch": 0.9482758620689655,
+      "grad_norm": 0.1967734843492508,
+      "learning_rate": 2.7500000000000004e-06,
+      "loss": 0.0711,
+      "step": 55
+    },
+    {
+      "epoch": 0.9655172413793104,
+      "grad_norm": 0.1832527369260788,
+      "learning_rate": 2.8000000000000003e-06,
+      "loss": 0.0686,
+      "step": 56
+    },
+    {
+      "epoch": 0.9827586206896551,
+      "grad_norm": 0.16995869576931,
+      "learning_rate": 2.85e-06,
+      "loss": 0.0644,
+      "step": 57
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.1822829246520996,
+      "learning_rate": 2.9e-06,
+      "loss": 0.0669,
+      "step": 58
+    },
+    {
+      "epoch": 1.0172413793103448,
+      "grad_norm": 0.17660750448703766,
+      "learning_rate": 2.95e-06,
+      "loss": 0.0668,
+      "step": 59
+    },
+    {
+      "epoch": 1.0344827586206897,
+      "grad_norm": 0.12004642933607101,
+      "learning_rate": 3e-06,
+      "loss": 0.0588,
+      "step": 60
+    },
+    {
+      "epoch": 1.0517241379310345,
+      "grad_norm": 0.15699979662895203,
+      "learning_rate": 3.05e-06,
+      "loss": 0.0631,
+      "step": 61
+    },
+    {
+      "epoch": 1.0689655172413792,
+      "grad_norm": 0.15650232136249542,
+      "learning_rate": 3.1000000000000004e-06,
+      "loss": 0.0596,
+      "step": 62
+    },
+    {
+      "epoch": 1.0862068965517242,
+      "grad_norm": 0.1188632994890213,
+      "learning_rate": 3.1500000000000003e-06,
+      "loss": 0.0614,
+      "step": 63
+    },
+    {
+      "epoch": 1.103448275862069,
+      "grad_norm": 0.11891665309667587,
+      "learning_rate": 3.2000000000000003e-06,
+      "loss": 0.0636,
+      "step": 64
+    },
+    {
+      "epoch": 1.1206896551724137,
+      "grad_norm": 0.12215171754360199,
+      "learning_rate": 3.2500000000000002e-06,
+      "loss": 0.0567,
+      "step": 65
+    },
+    {
+      "epoch": 1.1379310344827587,
+      "grad_norm": 0.10582094639539719,
+      "learning_rate": 3.3000000000000006e-06,
+      "loss": 0.0588,
+      "step": 66
+    },
+    {
+      "epoch": 1.1551724137931034,
+      "grad_norm": 0.11556132137775421,
+      "learning_rate": 3.3500000000000005e-06,
+      "loss": 0.0628,
+      "step": 67
+    },
+    {
+      "epoch": 1.1724137931034484,
+      "grad_norm": 0.12296544015407562,
+      "learning_rate": 3.4000000000000005e-06,
+      "loss": 0.0623,
+      "step": 68
+    },
+    {
+      "epoch": 1.1896551724137931,
+      "grad_norm": 0.1281358152627945,
+      "learning_rate": 3.45e-06,
+      "loss": 0.0626,
+      "step": 69
+    },
+    {
+      "epoch": 1.206896551724138,
+      "grad_norm": 0.1277759075164795,
+      "learning_rate": 3.5e-06,
+      "loss": 0.0592,
+      "step": 70
+    },
+    {
+      "epoch": 1.2241379310344827,
+      "grad_norm": 0.09552627056837082,
+      "learning_rate": 3.5500000000000003e-06,
+      "loss": 0.0595,
+      "step": 71
+    },
+    {
+      "epoch": 1.2413793103448276,
+      "grad_norm": 0.11053085327148438,
+      "learning_rate": 3.6000000000000003e-06,
+      "loss": 0.0546,
+      "step": 72
+    },
+    {
+      "epoch": 1.2586206896551724,
+      "grad_norm": 0.09386970102787018,
+      "learning_rate": 3.65e-06,
+      "loss": 0.0556,
+      "step": 73
+    },
+    {
+      "epoch": 1.2758620689655173,
+      "grad_norm": 0.0934402197599411,
+      "learning_rate": 3.7e-06,
+      "loss": 0.0563,
+      "step": 74
+    },
+    {
+      "epoch": 1.293103448275862,
+      "grad_norm": 0.09964020550251007,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.0569,
+      "step": 75
+    },
+    {
+      "epoch": 1.3103448275862069,
+      "grad_norm": 0.08830665796995163,
+      "learning_rate": 3.8000000000000005e-06,
+      "loss": 0.0578,
+      "step": 76
+    },
+    {
+      "epoch": 1.3275862068965516,
+      "grad_norm": 0.13536307215690613,
+      "learning_rate": 3.85e-06,
+      "loss": 0.0569,
+      "step": 77
+    },
+    {
+      "epoch": 1.3448275862068966,
+      "grad_norm": 0.10773950815200806,
+      "learning_rate": 3.900000000000001e-06,
+      "loss": 0.0581,
+      "step": 78
+    },
+    {
+      "epoch": 1.3620689655172413,
+      "grad_norm": 0.09468439221382141,
+      "learning_rate": 3.95e-06,
+      "loss": 0.0549,
+      "step": 79
+    },
+    {
+      "epoch": 1.3793103448275863,
+      "grad_norm": 0.08775551617145538,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 0.0594,
+      "step": 80
+    },
+    {
+      "epoch": 1.396551724137931,
+      "grad_norm": 0.12008150666952133,
+      "learning_rate": 4.05e-06,
+      "loss": 0.057,
+      "step": 81
+    },
+    {
+      "epoch": 1.4137931034482758,
+      "grad_norm": 0.12070683389902115,
+      "learning_rate": 4.1e-06,
+      "loss": 0.0549,
+      "step": 82
+    },
+    {
+      "epoch": 1.4310344827586206,
+      "grad_norm": 0.1037198081612587,
+      "learning_rate": 4.15e-06,
+      "loss": 0.0554,
+      "step": 83
+    },
+    {
+      "epoch": 1.4482758620689655,
+      "grad_norm": 0.14529870450496674,
+      "learning_rate": 4.2000000000000004e-06,
+      "loss": 0.0549,
+      "step": 84
+    },
+    {
+      "epoch": 1.4655172413793103,
+      "grad_norm": 0.0954233855009079,
+      "learning_rate": 4.25e-06,
+      "loss": 0.0556,
+      "step": 85
+    },
+    {
+      "epoch": 1.4827586206896552,
+      "grad_norm": 0.08504101634025574,
+      "learning_rate": 4.3e-06,
+      "loss": 0.0505,
+      "step": 86
+    },
+    {
+      "epoch": 1.5,
+      "grad_norm": 0.15293122828006744,
+      "learning_rate": 4.350000000000001e-06,
+      "loss": 0.0577,
+      "step": 87
+    },
+    {
+      "epoch": 1.5172413793103448,
+      "grad_norm": 0.10908783227205276,
+      "learning_rate": 4.4e-06,
+      "loss": 0.0556,
+      "step": 88
+    },
+    {
+      "epoch": 1.5344827586206895,
+      "grad_norm": 0.12018983066082001,
+      "learning_rate": 4.450000000000001e-06,
+      "loss": 0.0554,
+      "step": 89
+    },
+    {
+      "epoch": 1.5517241379310345,
+      "grad_norm": 0.1018645316362381,
+      "learning_rate": 4.5e-06,
+      "loss": 0.0548,
+      "step": 90
+    },
+    {
+      "epoch": 1.5689655172413794,
+      "grad_norm": 0.09623338282108307,
+      "learning_rate": 4.5500000000000005e-06,
+      "loss": 0.0566,
+      "step": 91
+    },
+    {
+      "epoch": 1.5862068965517242,
+      "grad_norm": 0.09007120132446289,
+      "learning_rate": 4.600000000000001e-06,
+      "loss": 0.0552,
+      "step": 92
+    },
+    {
+      "epoch": 1.603448275862069,
+      "grad_norm": 0.07549538463354111,
+      "learning_rate": 4.65e-06,
+      "loss": 0.056,
+      "step": 93
+    },
+    {
+      "epoch": 1.6206896551724137,
+      "grad_norm": 0.14191967248916626,
+      "learning_rate": 4.7e-06,
+      "loss": 0.0547,
+      "step": 94
+    },
+    {
+      "epoch": 1.6379310344827587,
+      "grad_norm": 0.13249421119689941,
+      "learning_rate": 4.75e-06,
+      "loss": 0.0529,
+      "step": 95
+    },
+    {
+      "epoch": 1.6551724137931034,
+      "grad_norm": 0.09079012274742126,
+      "learning_rate": 4.800000000000001e-06,
+      "loss": 0.0537,
+      "step": 96
+    },
+    {
+      "epoch": 1.6724137931034484,
+      "grad_norm": 0.08319538831710815,
+      "learning_rate": 4.85e-06,
+      "loss": 0.0523,
+      "step": 97
+    },
+    {
+      "epoch": 1.6896551724137931,
+      "grad_norm": 0.09330669790506363,
+      "learning_rate": 4.9000000000000005e-06,
+      "loss": 0.0515,
+      "step": 98
+    },
+    {
+      "epoch": 1.706896551724138,
+      "grad_norm": 0.1553690880537033,
+      "learning_rate": 4.95e-06,
+      "loss": 0.0539,
+      "step": 99
+    },
+    {
+      "epoch": 1.7241379310344827,
+      "grad_norm": 0.10644665360450745,
+      "learning_rate": 5e-06,
+      "loss": 0.0537,
+      "step": 100
+    },
+    {
+      "epoch": 1.7413793103448276,
+      "grad_norm": 0.10171601176261902,
+      "learning_rate": 4.999799414013322e-06,
+      "loss": 0.0526,
+      "step": 101
+    },
+    {
+      "epoch": 1.7586206896551724,
+      "grad_norm": 0.08432668447494507,
+      "learning_rate": 4.999197688241076e-06,
+      "loss": 0.0525,
+      "step": 102
+    },
+    {
+      "epoch": 1.7758620689655173,
+      "grad_norm": 0.08552516251802444,
+      "learning_rate": 4.998194919241471e-06,
+      "loss": 0.0528,
+      "step": 103
+    },
+    {
+      "epoch": 1.793103448275862,
+      "grad_norm": 0.10473164916038513,
+      "learning_rate": 4.996791267927632e-06,
+      "loss": 0.0509,
+      "step": 104
+    },
+    {
+      "epoch": 1.8103448275862069,
+      "grad_norm": 0.08742501586675644,
+      "learning_rate": 4.994986959541788e-06,
+      "loss": 0.0527,
+      "step": 105
+    },
+    {
+      "epoch": 1.8275862068965516,
+      "grad_norm": 0.07423796504735947,
+      "learning_rate": 4.9927822836191185e-06,
+      "loss": 0.0532,
+      "step": 106
+    },
+    {
+      "epoch": 1.8448275862068966,
+      "grad_norm": 0.0921708270907402,
+      "learning_rate": 4.990177593941303e-06,
+      "loss": 0.0545,
+      "step": 107
+    },
+    {
+      "epoch": 1.8620689655172413,
+      "grad_norm": 0.08664074540138245,
+      "learning_rate": 4.987173308479738e-06,
+      "loss": 0.0505,
+      "step": 108
+    },
+    {
+      "epoch": 1.8793103448275863,
+      "grad_norm": 0.09824781864881516,
+      "learning_rate": 4.9837699093284765e-06,
+      "loss": 0.053,
+      "step": 109
+    },
+    {
+      "epoch": 1.896551724137931,
+      "grad_norm": 0.0782749354839325,
+      "learning_rate": 4.9799679426268575e-06,
+      "loss": 0.0525,
+      "step": 110
+    },
+    {
+      "epoch": 1.9137931034482758,
+      "grad_norm": 0.09107070416212082,
+      "learning_rate": 4.975768018471877e-06,
+      "loss": 0.0536,
+      "step": 111
+    },
+    {
+      "epoch": 1.9310344827586206,
+      "grad_norm": 0.07815398275852203,
+      "learning_rate": 4.971170810820279e-06,
+      "loss": 0.0486,
+      "step": 112
+    },
+    {
+      "epoch": 1.9482758620689655,
+      "grad_norm": 0.08175615966320038,
+      "learning_rate": 4.966177057380409e-06,
+      "loss": 0.0545,
+      "step": 113
+    },
+    {
+      "epoch": 1.9655172413793105,
+      "grad_norm": 0.08906937390565872,
+      "learning_rate": 4.960787559493836e-06,
+      "loss": 0.0478,
+      "step": 114
+    },
+    {
+      "epoch": 1.9827586206896552,
+      "grad_norm": 0.08419270813465118,
+      "learning_rate": 4.955003182006761e-06,
+      "loss": 0.0528,
+      "step": 115
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.10093575716018677,
+      "learning_rate": 4.948824853131237e-06,
+      "loss": 0.049,
+      "step": 116
+    },
+    {
+      "epoch": 2.0172413793103448,
+      "grad_norm": 0.10379553586244583,
+      "learning_rate": 4.942253564296217e-06,
+      "loss": 0.0489,
+      "step": 117
+    },
+    {
+      "epoch": 2.0344827586206895,
+      "grad_norm": 0.08708551526069641,
+      "learning_rate": 4.935290369988468e-06,
+      "loss": 0.0496,
+      "step": 118
+    },
+    {
+      "epoch": 2.0517241379310347,
+      "grad_norm": 0.0780441164970398,
+      "learning_rate": 4.927936387583348e-06,
+      "loss": 0.0488,
+      "step": 119
+    },
+    {
+      "epoch": 2.0689655172413794,
+      "grad_norm": 0.08206549286842346,
+      "learning_rate": 4.920192797165511e-06,
+      "loss": 0.0493,
+      "step": 120
+    },
+    {
+      "epoch": 2.086206896551724,
+      "grad_norm": 0.0740761011838913,
+      "learning_rate": 4.912060841339536e-06,
+      "loss": 0.0472,
+      "step": 121
+    },
+    {
+      "epoch": 2.103448275862069,
+      "grad_norm": 0.10379961133003235,
+      "learning_rate": 4.9035418250305314e-06,
+      "loss": 0.0512,
+      "step": 122
+    },
+    {
+      "epoch": 2.1206896551724137,
+      "grad_norm": 0.0821395292878151,
+      "learning_rate": 4.894637115274728e-06,
+      "loss": 0.0474,
+      "step": 123
+    },
+    {
+      "epoch": 2.1379310344827585,
+      "grad_norm": 0.09478521347045898,
+      "learning_rate": 4.8853481410001225e-06,
+      "loss": 0.0475,
+      "step": 124
+    },
+    {
+      "epoch": 2.1551724137931036,
+      "grad_norm": 0.0971507877111435,
+      "learning_rate": 4.875676392797169e-06,
+      "loss": 0.0509,
+      "step": 125
+    },
+    {
+      "epoch": 2.1724137931034484,
+      "grad_norm": 0.11455722898244858,
+      "learning_rate": 4.865623422679593e-06,
+      "loss": 0.051,
+      "step": 126
+    },
+    {
+      "epoch": 2.189655172413793,
+      "grad_norm": 0.10524770617485046,
+      "learning_rate": 4.855190843835338e-06,
+      "loss": 0.0486,
+      "step": 127
+    },
+    {
+      "epoch": 2.206896551724138,
+      "grad_norm": 0.11388484388589859,
+      "learning_rate": 4.844380330367701e-06,
+      "loss": 0.0431,
+      "step": 128
+    },
+    {
+      "epoch": 2.2241379310344827,
+      "grad_norm": 0.0915762260556221,
+      "learning_rate": 4.833193617026692e-06,
+      "loss": 0.048,
+      "step": 129
+    },
+    {
+      "epoch": 2.2413793103448274,
+      "grad_norm": 0.10419991612434387,
+      "learning_rate": 4.821632498930656e-06,
+      "loss": 0.0518,
+      "step": 130
+    },
+    {
+      "epoch": 2.2586206896551726,
+      "grad_norm": 0.08066820353269577,
+      "learning_rate": 4.809698831278217e-06,
+      "loss": 0.0488,
+      "step": 131
+    },
+    {
+      "epoch": 2.2758620689655173,
+      "grad_norm": 0.07880278676748276,
+      "learning_rate": 4.797394529050577e-06,
+      "loss": 0.048,
+      "step": 132
+    },
+    {
+      "epoch": 2.293103448275862,
+      "grad_norm": 0.09013557434082031,
+      "learning_rate": 4.784721566704217e-06,
+      "loss": 0.0483,
+      "step": 133
+    },
+    {
+      "epoch": 2.310344827586207,
+      "grad_norm": 0.1121375784277916,
+      "learning_rate": 4.771681977854062e-06,
+      "loss": 0.0517,
+      "step": 134
+    },
+    {
+      "epoch": 2.3275862068965516,
+      "grad_norm": 0.0830090343952179,
+      "learning_rate": 4.75827785494715e-06,
+      "loss": 0.0489,
+      "step": 135
+    },
+    {
+      "epoch": 2.344827586206897,
+      "grad_norm": 0.08320346474647522,
+      "learning_rate": 4.744511348926855e-06,
+      "loss": 0.049,
+      "step": 136
+    },
+    {
+      "epoch": 2.3620689655172415,
+      "grad_norm": 0.08780596405267715,
+      "learning_rate": 4.730384668887731e-06,
+      "loss": 0.0518,
+      "step": 137
+    },
+    {
+      "epoch": 2.3793103448275863,
+      "grad_norm": 0.09375182539224625,
+      "learning_rate": 4.715900081721021e-06,
+      "loss": 0.0497,
+      "step": 138
+    },
+    {
+      "epoch": 2.396551724137931,
+      "grad_norm": 0.09378068894147873,
+      "learning_rate": 4.7010599117508936e-06,
+      "loss": 0.049,
+      "step": 139
+    },
+    {
+      "epoch": 2.413793103448276,
+      "grad_norm": 0.11429137736558914,
+      "learning_rate": 4.685866540361456e-06,
+      "loss": 0.0464,
+      "step": 140
+    },
+    {
+      "epoch": 2.4310344827586206,
+      "grad_norm": 0.08398573845624924,
+      "learning_rate": 4.670322405614621e-06,
+      "loss": 0.0487,
+      "step": 141
+    },
+    {
+      "epoch": 2.4482758620689653,
+      "grad_norm": 0.08914072066545486,
+      "learning_rate": 4.654430001858874e-06,
+      "loss": 0.0483,
+      "step": 142
+    },
+    {
+      "epoch": 2.4655172413793105,
+      "grad_norm": 0.09420543164014816,
+      "learning_rate": 4.638191879329005e-06,
+      "loss": 0.0496,
+      "step": 143
+    },
+    {
+      "epoch": 2.4827586206896552,
+      "grad_norm": 0.13024932146072388,
+      "learning_rate": 4.621610643736878e-06,
+      "loss": 0.0437,
+      "step": 144
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.1074167937040329,
+      "learning_rate": 4.6046889558532925e-06,
+      "loss": 0.0488,
+      "step": 145
+    },
+    {
+      "epoch": 2.5172413793103448,
+      "grad_norm": 0.11312957108020782,
+      "learning_rate": 4.587429531081019e-06,
+      "loss": 0.0494,
+      "step": 146
+    },
+    {
+      "epoch": 2.5344827586206895,
+      "grad_norm": 0.09910023957490921,
+      "learning_rate": 4.569835139019054e-06,
+      "loss": 0.0488,
+      "step": 147
+    },
+    {
+      "epoch": 2.5517241379310347,
+      "grad_norm": 0.09021929651498795,
+      "learning_rate": 4.551908603018191e-06,
+      "loss": 0.0449,
+      "step": 148
+    },
+    {
+      "epoch": 2.5689655172413794,
+      "grad_norm": 0.08959867060184479,
+      "learning_rate": 4.53365279972796e-06,
+      "loss": 0.0519,
+      "step": 149
+    },
+    {
+      "epoch": 2.586206896551724,
+      "grad_norm": 0.1165613979101181,
+      "learning_rate": 4.515070658635013e-06,
+      "loss": 0.0478,
+      "step": 150
+    },
+    {
+      "epoch": 2.603448275862069,
+      "grad_norm": 0.10728327184915543,
+      "learning_rate": 4.4961651615930344e-06,
+      "loss": 0.0493,
+      "step": 151
+    },
+    {
+      "epoch": 2.6206896551724137,
+      "grad_norm": 0.09753444045782089,
+      "learning_rate": 4.476939342344246e-06,
+      "loss": 0.0507,
+      "step": 152
+    },
+    {
+      "epoch": 2.637931034482759,
+      "grad_norm": 0.08872995525598526,
+      "learning_rate": 4.457396286032589e-06,
+      "loss": 0.046,
+      "step": 153
+    },
+    {
+      "epoch": 2.655172413793103,
+      "grad_norm": 0.08858656883239746,
+      "learning_rate": 4.437539128708647e-06,
+      "loss": 0.0478,
+      "step": 154
+    },
+    {
+      "epoch": 2.6724137931034484,
+      "grad_norm": 0.0953850969672203,
+      "learning_rate": 4.417371056826417e-06,
+      "loss": 0.0461,
+      "step": 155
+    },
+    {
+      "epoch": 2.689655172413793,
+      "grad_norm": 0.09668698161840439,
+      "learning_rate": 4.396895306731978e-06,
+      "loss": 0.0476,
+      "step": 156
+    },
+    {
+      "epoch": 2.706896551724138,
+      "grad_norm": 0.11896238476037979,
+      "learning_rate": 4.376115164144157e-06,
+      "loss": 0.0488,
+      "step": 157
+    },
+    {
+      "epoch": 2.7241379310344827,
+      "grad_norm": 0.08256583660840988,
+      "learning_rate": 4.355033963627277e-06,
+      "loss": 0.0448,
+      "step": 158
+    },
+    {
+      "epoch": 2.7413793103448274,
+      "grad_norm": 0.10837860405445099,
+      "learning_rate": 4.333655088056065e-06,
+      "loss": 0.0499,
+      "step": 159
+    },
+    {
+      "epoch": 2.7586206896551726,
+      "grad_norm": 0.09428149461746216,
+      "learning_rate": 4.3119819680728e-06,
+      "loss": 0.0483,
+      "step": 160
+    },
+    {
+      "epoch": 2.7758620689655173,
+      "grad_norm": 0.08462841808795929,
+      "learning_rate": 4.290018081536807e-06,
+      "loss": 0.0474,
+      "step": 161
+    },
+    {
+      "epoch": 2.793103448275862,
+      "grad_norm": 0.0951421931385994,
+      "learning_rate": 4.267766952966369e-06,
+      "loss": 0.0471,
+      "step": 162
+    },
+    {
+      "epoch": 2.810344827586207,
+      "grad_norm": 0.08592601120471954,
+      "learning_rate": 4.245232152973148e-06,
+      "loss": 0.0452,
+      "step": 163
+    },
+    {
+      "epoch": 2.8275862068965516,
+      "grad_norm": 0.09917694330215454,
+      "learning_rate": 4.222417297689217e-06,
+      "loss": 0.0489,
+      "step": 164
+    },
+    {
+      "epoch": 2.844827586206897,
+      "grad_norm": 0.08759764581918716,
+      "learning_rate": 4.199326048186783e-06,
+      "loss": 0.0457,
+      "step": 165
+    },
+    {
+      "epoch": 2.862068965517241,
+      "grad_norm": 0.08396806567907333,
+      "learning_rate": 4.175962109890697e-06,
+      "loss": 0.0452,
+      "step": 166
+    },
+    {
+      "epoch": 2.8793103448275863,
+      "grad_norm": 0.10327853262424469,
+      "learning_rate": 4.152329231983852e-06,
+      "loss": 0.0469,
+      "step": 167
+    },
+    {
+      "epoch": 2.896551724137931,
+      "grad_norm": 0.11525449901819229,
+      "learning_rate": 4.128431206805556e-06,
+      "loss": 0.0486,
+      "step": 168
+    },
+    {
+      "epoch": 2.913793103448276,
+      "grad_norm": 0.09090360999107361,
+      "learning_rate": 4.104271869242975e-06,
+      "loss": 0.043,
+      "step": 169
+    },
+    {
+      "epoch": 2.9310344827586206,
+      "grad_norm": 0.09151678532361984,
+      "learning_rate": 4.07985509611576e-06,
+      "loss": 0.0453,
+      "step": 170
+    },
+    {
+      "epoch": 2.9482758620689653,
+      "grad_norm": 0.0855179950594902,
+      "learning_rate": 4.0551848055539345e-06,
+      "loss": 0.044,
+      "step": 171
+    },
+    {
+      "epoch": 2.9655172413793105,
+      "grad_norm": 0.09055305272340775,
+      "learning_rate": 4.030264956369158e-06,
+      "loss": 0.0418,
+      "step": 172
+    },
+    {
+      "epoch": 2.9827586206896552,
+      "grad_norm": 0.11615358293056488,
+      "learning_rate": 4.005099547419458e-06,
+      "loss": 0.0456,
+      "step": 173
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.19266903400421143,
+      "learning_rate": 3.979692616967543e-06,
+      "loss": 0.0505,
+      "step": 174
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 348,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 6,
+  "save_steps": 58,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 6.894933011423298e+18,
+  "train_batch_size": 4,
+  "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:e4a8bf738cfba89eb19d24178060f165bf0f68185db34cebf3b311b7c61eebc5
+size 7992

checkpoint-174/zero_to_fp32.py

@@ -0,0 +1,604 @@
+#!/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, exclude_frozen_parameters):
+    """
+    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,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+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,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _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,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _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, exclude_frozen_parameters=False):
+    """
+    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``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    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, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    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``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    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("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    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,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)

checkpoint-232/README.md

@@ -0,0 +1,202 @@
+---
+base_model: Qwen/Qwen2.5-32B-Instruct
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.14.0