Training in progress, step 350, checkpoint

checkpoint-350/README.md

@@ -0,0 +1,202 @@
+---
+base_model: microsoft/Phi-3-mini-4k-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-350/adapter_config.json

@@ -0,0 +1,34 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "microsoft/Phi-3-mini-4k-instruct",
+  "bias": "none",
+  "eva_config": null,
+  "exclude_modules": null,
+  "fan_in_fan_out": false,
+  "inference_mode": true,
+  "init_lora_weights": true,
+  "layer_replication": null,
+  "layers_pattern": null,
+  "layers_to_transform": null,
+  "loftq_config": {},
+  "lora_alpha": 16,
+  "lora_bias": false,
+  "lora_dropout": 0.0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_up_proj",
+    "o_proj",
+    "down_proj",
+    "qkv_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "use_dora": false,
+  "use_rslora": false
+}

checkpoint-350/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b6e388bd00b4d5c980d869fcf003825ecb3b7824bd719850855eae9bc278ae11
+size 25200088

checkpoint-350/global_step350/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b8e2bf577d1c0471b93cc8ea6869e7417704c598e698e0d250f34d72fcee851
+size 18881328

checkpoint-350/global_step350/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a41c9202e7b073f717403ef0959e673ff405e0f64fa62d31a33ab3d99954f45
+size 18881328

checkpoint-350/global_step350/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:d67db605a439726e404c5663d2309746a8feecd2d984d6dbc1772db29eb4cc83
+size 18881328

checkpoint-350/global_step350/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1767e5eee2601ef39a9931cc4f2ebeea3bf9ab9a0a68d8876662d1e7618d0508
+size 18881392

checkpoint-350/global_step350/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:900fca07ca7ce3d2362a5b3cf3a52c9b6774ff8a9035a4a137a921645d24ecf4
+size 18881392

checkpoint-350/global_step350/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e32c845ba741f1dbda3d176e06aceba5457e72628e70a0c08e1147d1bdac95a3
+size 18881392

checkpoint-350/global_step350/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cdf898f385fc680566f7f89b152c0007111ab7238a48a0a93445c0e938619499
+size 18881392

checkpoint-350/global_step350/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:572d4141eb0df53ce1da244d4cb3a3ff474e4ebd44fd1f2ac8ee65ff9a96aed6
+size 18881392

checkpoint-350/global_step350/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7411b83b77a6b64df84140c56d54a2a3e48304f7137b077db05d6bf0d123e9e7
+size 25379244

checkpoint-350/latest

@@ -0,0 +1 @@
+global_step350

checkpoint-350/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04cb5208648fd09a2e0403d51973f74ffbfd93cbd5da59e1e99c8df03769a86c
+size 15984

checkpoint-350/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7034685b36b93a4dd3a50697b0b1c314b249b2189ec2cb96b757312b1514a579
+size 15984

checkpoint-350/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e851fe1c1de0057f4eecefed6a131fa9021334eb43f6e7e65fdb270a25ac864
+size 15984

checkpoint-350/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:978379030048e432baa510ec4fc9514faa08fe564ab964b3a4d05e8f60306495
+size 15984

checkpoint-350/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdbc75d90af112615b53d15931e8157a80e37bcd110aac9a3089f5f6f5344171
+size 15984

checkpoint-350/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c8a310f6ca2ca89570eb2cc68544656b30224f00b2d6d96eeda6e0cb8be50ab
+size 15984

checkpoint-350/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c5b8110fcf6e044b6860c6305be969cfe03129549b92dc6fc2394448e9265d6
+size 15984

checkpoint-350/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6f936acaf5a2d5fe8c38d945450417facbf1577584c216908a396d3cc20bec88
+size 15984

checkpoint-350/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a14cd0c4c28f82898909e19382d2a0bea7882767db0ae279df13535097f3228
+size 1064

checkpoint-350/special_tokens_map.json

@@ -0,0 +1,30 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|end|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-350/tokenizer_config.json

@@ -0,0 +1,133 @@
+{
+  "add_bos_token": false,
+  "add_eos_token": false,
+  "add_prefix_space": null,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<unk>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": false
+    },
+    "32000": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32001": {
+      "content": "<|assistant|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32002": {
+      "content": "<|placeholder1|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32003": {
+      "content": "<|placeholder2|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32004": {
+      "content": "<|placeholder3|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32005": {
+      "content": "<|placeholder4|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32006": {
+      "content": "<|system|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32007": {
+      "content": "<|end|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "32008": {
+      "content": "<|placeholder5|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32009": {
+      "content": "<|placeholder6|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    },
+    "32010": {
+      "content": "<|user|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": true,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "chat_template": "{% set system_message = 'You are a helpful AI assistant.' %}{% if messages[0]['role'] == 'system' %}{% set system_message = messages[0]['content'] %}{% endif %}{% if system_message is defined %}{{ '<s>' + '<|system|>\n' + system_message + '<|end|>\n' }}{% endif %}{% for message in messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '<|user|>\n' + content + '<|end|>\n<|assistant|>\n' }}{% elif message['role'] == 'assistant' %}{{ content + '<|end|>' + '\n' }}{% endif %}{% endfor %}",
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|end|>",
+  "extra_special_tokens": {},
+  "legacy": false,
+  "model_max_length": 4096,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "right",
+  "sp_model_kwargs": {},
+  "split_special_tokens": false,
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": false
+}

checkpoint-350/trainer_state.json

@@ -0,0 +1,670 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.28254288597376387,
+  "eval_steps": 50,
+  "global_step": 350,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.008072653884964682,
+      "grad_norm": 0.04926518723368645,
+      "learning_rate": 4.999451708687114e-06,
+      "logits/chosen": 14.542106628417969,
+      "logits/rejected": 14.864250183105469,
+      "logps/chosen": -0.2809702754020691,
+      "logps/rejected": -0.3013763725757599,
+      "loss": 0.9221,
+      "rewards/accuracies": 0.42500001192092896,
+      "rewards/chosen": -0.421455442905426,
+      "rewards/margins": 0.03060910664498806,
+      "rewards/rejected": -0.452064573764801,
+      "step": 10
+    },
+    {
+      "epoch": 0.016145307769929364,
+      "grad_norm": 0.056249432265758514,
+      "learning_rate": 4.997807075247147e-06,
+      "logits/chosen": 14.614748001098633,
+      "logits/rejected": 15.259109497070312,
+      "logps/chosen": -0.2828002870082855,
+      "logps/rejected": -0.3477819561958313,
+      "loss": 0.916,
+      "rewards/accuracies": 0.550000011920929,
+      "rewards/chosen": -0.4242004454135895,
+      "rewards/margins": 0.09747247397899628,
+      "rewards/rejected": -0.5216729044914246,
+      "step": 20
+    },
+    {
+      "epoch": 0.024217961654894045,
+      "grad_norm": 0.0664869099855423,
+      "learning_rate": 4.9950668210706795e-06,
+      "logits/chosen": 14.391204833984375,
+      "logits/rejected": 14.82734203338623,
+      "logps/chosen": -0.287629634141922,
+      "logps/rejected": -0.3329126834869385,
+      "loss": 0.914,
+      "rewards/accuracies": 0.48750001192092896,
+      "rewards/chosen": -0.4314444661140442,
+      "rewards/margins": 0.06792456656694412,
+      "rewards/rejected": -0.4993689954280853,
+      "step": 30
+    },
+    {
+      "epoch": 0.03229061553985873,
+      "grad_norm": 0.055584829300642014,
+      "learning_rate": 4.9912321481237616e-06,
+      "logits/chosen": 13.93278980255127,
+      "logits/rejected": 14.886846542358398,
+      "logps/chosen": -0.28155946731567383,
+      "logps/rejected": -0.3678051829338074,
+      "loss": 0.9273,
+      "rewards/accuracies": 0.6000000238418579,
+      "rewards/chosen": -0.42233920097351074,
+      "rewards/margins": 0.12936851382255554,
+      "rewards/rejected": -0.5517078042030334,
+      "step": 40
+    },
+    {
+      "epoch": 0.04036326942482341,
+      "grad_norm": 0.07997103035449982,
+      "learning_rate": 4.986304738420684e-06,
+      "logits/chosen": 14.172693252563477,
+      "logits/rejected": 14.742494583129883,
+      "logps/chosen": -0.28225988149642944,
+      "logps/rejected": -0.34329456090927124,
+      "loss": 0.9081,
+      "rewards/accuracies": 0.48750001192092896,
+      "rewards/chosen": -0.42338982224464417,
+      "rewards/margins": 0.0915520042181015,
+      "rewards/rejected": -0.5149418115615845,
+      "step": 50
+    },
+    {
+      "epoch": 0.04036326942482341,
+      "eval_logits/chosen": 14.200166702270508,
+      "eval_logits/rejected": 14.817726135253906,
+      "eval_logps/chosen": -0.2625390887260437,
+      "eval_logps/rejected": -0.3458769917488098,
+      "eval_loss": 0.9080610275268555,
+      "eval_rewards/accuracies": 0.5544554591178894,
+      "eval_rewards/chosen": -0.39380866289138794,
+      "eval_rewards/margins": 0.1250067949295044,
+      "eval_rewards/rejected": -0.5188154578208923,
+      "eval_runtime": 29.8098,
+      "eval_samples_per_second": 26.87,
+      "eval_steps_per_second": 3.388,
+      "step": 50
+    },
+    {
+      "epoch": 0.04843592330978809,
+      "grad_norm": 0.06322095543146133,
+      "learning_rate": 4.980286753286196e-06,
+      "logits/chosen": 14.008010864257812,
+      "logits/rejected": 14.975939750671387,
+      "logps/chosen": -0.2697351574897766,
+      "logps/rejected": -0.3445274233818054,
+      "loss": 0.9195,
+      "rewards/accuracies": 0.5,
+      "rewards/chosen": -0.40460270643234253,
+      "rewards/margins": 0.1121884360909462,
+      "rewards/rejected": -0.5167912244796753,
+      "step": 60
+    },
+    {
+      "epoch": 0.056508577194752774,
+      "grad_norm": 0.12975476682186127,
+      "learning_rate": 4.973180832407471e-06,
+      "logits/chosen": 14.213134765625,
+      "logits/rejected": 14.889978408813477,
+      "logps/chosen": -0.2839818596839905,
+      "logps/rejected": -0.3382417559623718,
+      "loss": 0.9241,
+      "rewards/accuracies": 0.5375000238418579,
+      "rewards/chosen": -0.4259727895259857,
+      "rewards/margins": 0.08138985931873322,
+      "rewards/rejected": -0.5073626637458801,
+      "step": 70
+    },
+    {
+      "epoch": 0.06458123107971746,
+      "grad_norm": 0.08352109789848328,
+      "learning_rate": 4.964990092676263e-06,
+      "logits/chosen": 14.395146369934082,
+      "logits/rejected": 14.900177001953125,
+      "logps/chosen": -0.27067264914512634,
+      "logps/rejected": -0.3516673743724823,
+      "loss": 0.9251,
+      "rewards/accuracies": 0.637499988079071,
+      "rewards/chosen": -0.4060089588165283,
+      "rewards/margins": 0.12149210274219513,
+      "rewards/rejected": -0.5275009870529175,
+      "step": 80
+    },
+    {
+      "epoch": 0.07265388496468214,
+      "grad_norm": 0.0793827548623085,
+      "learning_rate": 4.9557181268217225e-06,
+      "logits/chosen": 13.905145645141602,
+      "logits/rejected": 14.897878646850586,
+      "logps/chosen": -0.26814645528793335,
+      "logps/rejected": -0.38061466813087463,
+      "loss": 0.9055,
+      "rewards/accuracies": 0.6000000238418579,
+      "rewards/chosen": -0.4022197127342224,
+      "rewards/margins": 0.16870227456092834,
+      "rewards/rejected": -0.5709219574928284,
+      "step": 90
+    },
+    {
+      "epoch": 0.08072653884964683,
+      "grad_norm": 0.09212008118629456,
+      "learning_rate": 4.9453690018345144e-06,
+      "logits/chosen": 14.11627197265625,
+      "logits/rejected": 14.571977615356445,
+      "logps/chosen": -0.3130624294281006,
+      "logps/rejected": -0.35073375701904297,
+      "loss": 0.911,
+      "rewards/accuracies": 0.48750001192092896,
+      "rewards/chosen": -0.46959367394447327,
+      "rewards/margins": 0.05650699883699417,
+      "rewards/rejected": -0.5261006951332092,
+      "step": 100
+    },
+    {
+      "epoch": 0.08072653884964683,
+      "eval_logits/chosen": 13.845876693725586,
+      "eval_logits/rejected": 14.490789413452148,
+      "eval_logps/chosen": -0.2531408667564392,
+      "eval_logps/rejected": -0.3464036285877228,
+      "eval_loss": 0.8986235857009888,
+      "eval_rewards/accuracies": 0.5544554591178894,
+      "eval_rewards/chosen": -0.3797112703323364,
+      "eval_rewards/margins": 0.13989417254924774,
+      "eval_rewards/rejected": -0.5196054577827454,
+      "eval_runtime": 29.0914,
+      "eval_samples_per_second": 27.534,
+      "eval_steps_per_second": 3.472,
+      "step": 100
+    },
+    {
+      "epoch": 0.08879919273461151,
+      "grad_norm": 0.08554862439632416,
+      "learning_rate": 4.933947257182901e-06,
+      "logits/chosen": 13.416229248046875,
+      "logits/rejected": 14.582674026489258,
+      "logps/chosen": -0.2523443102836609,
+      "logps/rejected": -0.38751405477523804,
+      "loss": 0.8993,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -0.3785164952278137,
+      "rewards/margins": 0.20275457203388214,
+      "rewards/rejected": -0.5812710523605347,
+      "step": 110
+    },
+    {
+      "epoch": 0.09687184661957618,
+      "grad_norm": 0.29209578037261963,
+      "learning_rate": 4.921457902821578e-06,
+      "logits/chosen": 13.874654769897461,
+      "logits/rejected": 14.423624992370605,
+      "logps/chosen": -0.27300480008125305,
+      "logps/rejected": -0.3526575267314911,
+      "loss": 0.8942,
+      "rewards/accuracies": 0.6000000238418579,
+      "rewards/chosen": -0.40950721502304077,
+      "rewards/margins": 0.11947910487651825,
+      "rewards/rejected": -0.5289863348007202,
+      "step": 120
+    },
+    {
+      "epoch": 0.10494450050454086,
+      "grad_norm": 0.1028478816151619,
+      "learning_rate": 4.907906416994146e-06,
+      "logits/chosen": 14.030723571777344,
+      "logits/rejected": 14.711235046386719,
+      "logps/chosen": -0.27410784363746643,
+      "logps/rejected": -0.3665519058704376,
+      "loss": 0.8922,
+      "rewards/accuracies": 0.550000011920929,
+      "rewards/chosen": -0.41116175055503845,
+      "rewards/margins": 0.1386660784482956,
+      "rewards/rejected": -0.5498278737068176,
+      "step": 130
+    },
+    {
+      "epoch": 0.11301715438950555,
+      "grad_norm": 0.08459590375423431,
+      "learning_rate": 4.893298743830168e-06,
+      "logits/chosen": 13.477182388305664,
+      "logits/rejected": 13.919464111328125,
+      "logps/chosen": -0.2495063841342926,
+      "logps/rejected": -0.3594816029071808,
+      "loss": 0.8957,
+      "rewards/accuracies": 0.612500011920929,
+      "rewards/chosen": -0.3742595613002777,
+      "rewards/margins": 0.16496284306049347,
+      "rewards/rejected": -0.5392224192619324,
+      "step": 140
+    },
+    {
+      "epoch": 0.12108980827447023,
+      "grad_norm": 0.12467797845602036,
+      "learning_rate": 4.8776412907378845e-06,
+      "logits/chosen": 12.801656723022461,
+      "logits/rejected": 13.564155578613281,
+      "logps/chosen": -0.2779986262321472,
+      "logps/rejected": -0.33566445112228394,
+      "loss": 0.892,
+      "rewards/accuracies": 0.4375,
+      "rewards/chosen": -0.4169979691505432,
+      "rewards/margins": 0.08649872243404388,
+      "rewards/rejected": -0.5034967064857483,
+      "step": 150
+    },
+    {
+      "epoch": 0.12108980827447023,
+      "eval_logits/chosen": 12.169118881225586,
+      "eval_logits/rejected": 13.019683837890625,
+      "eval_logps/chosen": -0.25547870993614197,
+      "eval_logps/rejected": -0.37277930974960327,
+      "eval_loss": 0.8789658546447754,
+      "eval_rewards/accuracies": 0.5742574334144592,
+      "eval_rewards/chosen": -0.38321802020072937,
+      "eval_rewards/margins": 0.17595094442367554,
+      "eval_rewards/rejected": -0.5591689944267273,
+      "eval_runtime": 29.0961,
+      "eval_samples_per_second": 27.529,
+      "eval_steps_per_second": 3.471,
+      "step": 150
+    },
+    {
+      "epoch": 0.12916246215943492,
+      "grad_norm": 0.1712370663881302,
+      "learning_rate": 4.860940925593703e-06,
+      "logits/chosen": 11.972528457641602,
+      "logits/rejected": 12.684088706970215,
+      "logps/chosen": -0.28340521454811096,
+      "logps/rejected": -0.3864063024520874,
+      "loss": 0.8789,
+      "rewards/accuracies": 0.574999988079071,
+      "rewards/chosen": -0.42510780692100525,
+      "rewards/margins": 0.15450166165828705,
+      "rewards/rejected": -0.5796095132827759,
+      "step": 160
+    },
+    {
+      "epoch": 0.13723511604439959,
+      "grad_norm": 0.21750673651695251,
+      "learning_rate": 4.84320497372973e-06,
+      "logits/chosen": 11.93881607055664,
+      "logits/rejected": 12.458230972290039,
+      "logps/chosen": -0.274599552154541,
+      "logps/rejected": -0.3882916271686554,
+      "loss": 0.8857,
+      "rewards/accuracies": 0.675000011920929,
+      "rewards/chosen": -0.4118993282318115,
+      "rewards/margins": 0.1705380380153656,
+      "rewards/rejected": -0.5824374556541443,
+      "step": 170
+    },
+    {
+      "epoch": 0.14530776992936428,
+      "grad_norm": 0.16000741720199585,
+      "learning_rate": 4.824441214720629e-06,
+      "logits/chosen": 9.631464958190918,
+      "logits/rejected": 10.772969245910645,
+      "logps/chosen": -0.287865549325943,
+      "logps/rejected": -0.4817379415035248,
+      "loss": 0.8699,
+      "rewards/accuracies": 0.6625000238418579,
+      "rewards/chosen": -0.4317983090877533,
+      "rewards/margins": 0.29080861806869507,
+      "rewards/rejected": -0.722606897354126,
+      "step": 180
+    },
+    {
+      "epoch": 0.15338042381432895,
+      "grad_norm": 0.46293890476226807,
+      "learning_rate": 4.804657878971252e-06,
+      "logits/chosen": 9.081937789916992,
+      "logits/rejected": 10.024572372436523,
+      "logps/chosen": -0.2941994071006775,
+      "logps/rejected": -0.4772109389305115,
+      "loss": 0.8565,
+      "rewards/accuracies": 0.675000011920929,
+      "rewards/chosen": -0.44129911065101624,
+      "rewards/margins": 0.2745172679424286,
+      "rewards/rejected": -0.7158163785934448,
+      "step": 190
+    },
+    {
+      "epoch": 0.16145307769929365,
+      "grad_norm": 0.24658434092998505,
+      "learning_rate": 4.783863644106502e-06,
+      "logits/chosen": 7.767125606536865,
+      "logits/rejected": 8.009145736694336,
+      "logps/chosen": -0.32877305150032043,
+      "logps/rejected": -0.47733697295188904,
+      "loss": 0.8482,
+      "rewards/accuracies": 0.5874999761581421,
+      "rewards/chosen": -0.49315959215164185,
+      "rewards/margins": 0.2228458821773529,
+      "rewards/rejected": -0.7160054445266724,
+      "step": 200
+    },
+    {
+      "epoch": 0.16145307769929365,
+      "eval_logits/chosen": 6.716187953948975,
+      "eval_logits/rejected": 7.248146057128906,
+      "eval_logps/chosen": -0.30652713775634766,
+      "eval_logps/rejected": -0.5056277513504028,
+      "eval_loss": 0.8243693113327026,
+      "eval_rewards/accuracies": 0.603960394859314,
+      "eval_rewards/chosen": -0.4597907066345215,
+      "eval_rewards/margins": 0.2986510097980499,
+      "eval_rewards/rejected": -0.7584417462348938,
+      "eval_runtime": 29.0989,
+      "eval_samples_per_second": 27.527,
+      "eval_steps_per_second": 3.471,
+      "step": 200
+    },
+    {
+      "epoch": 0.16952573158425832,
+      "grad_norm": 0.40796443819999695,
+      "learning_rate": 4.762067631165049e-06,
+      "logits/chosen": 5.932644844055176,
+      "logits/rejected": 6.521953582763672,
+      "logps/chosen": -0.32976508140563965,
+      "logps/rejected": -0.5628186464309692,
+      "loss": 0.7988,
+      "rewards/accuracies": 0.612500011920929,
+      "rewards/chosen": -0.4946475923061371,
+      "rewards/margins": 0.3495803475379944,
+      "rewards/rejected": -0.8442279696464539,
+      "step": 210
+    },
+    {
+      "epoch": 0.17759838546922302,
+      "grad_norm": 0.42468318343162537,
+      "learning_rate": 4.7392794005985324e-06,
+      "logits/chosen": 6.069305896759033,
+      "logits/rejected": 5.8950395584106445,
+      "logps/chosen": -0.37205421924591064,
+      "logps/rejected": -0.6190425753593445,
+      "loss": 0.796,
+      "rewards/accuracies": 0.6000000238418579,
+      "rewards/chosen": -0.5580812692642212,
+      "rewards/margins": 0.3704826235771179,
+      "rewards/rejected": -0.9285639524459839,
+      "step": 220
+    },
+    {
+      "epoch": 0.1856710393541877,
+      "grad_norm": 0.4138280153274536,
+      "learning_rate": 4.715508948078037e-06,
+      "logits/chosen": 3.5271706581115723,
+      "logits/rejected": 3.363534927368164,
+      "logps/chosen": -0.41677650809288025,
+      "logps/rejected": -0.7121980786323547,
+      "loss": 0.7457,
+      "rewards/accuracies": 0.637499988079071,
+      "rewards/chosen": -0.6251648664474487,
+      "rewards/margins": 0.4431324005126953,
+      "rewards/rejected": -1.0682971477508545,
+      "step": 230
+    },
+    {
+      "epoch": 0.19374369323915236,
+      "grad_norm": 1.4327284097671509,
+      "learning_rate": 4.690766700109659e-06,
+      "logits/chosen": 2.296924114227295,
+      "logits/rejected": 1.6135867834091187,
+      "logps/chosen": -0.506227970123291,
+      "logps/rejected": -0.8080593943595886,
+      "loss": 0.7453,
+      "rewards/accuracies": 0.550000011920929,
+      "rewards/chosen": -0.7593418955802917,
+      "rewards/margins": 0.4527471661567688,
+      "rewards/rejected": -1.21208918094635,
+      "step": 240
+    },
+    {
+      "epoch": 0.20181634712411706,
+      "grad_norm": 0.4413074553012848,
+      "learning_rate": 4.665063509461098e-06,
+      "logits/chosen": 1.4450469017028809,
+      "logits/rejected": 0.40727120637893677,
+      "logps/chosen": -0.4749869406223297,
+      "logps/rejected": -0.9967275857925415,
+      "loss": 0.704,
+      "rewards/accuracies": 0.6625000238418579,
+      "rewards/chosen": -0.712480366230011,
+      "rewards/margins": 0.7826108932495117,
+      "rewards/rejected": -1.4950913190841675,
+      "step": 250
+    },
+    {
+      "epoch": 0.20181634712411706,
+      "eval_logits/chosen": 1.4614256620407104,
+      "eval_logits/rejected": 0.6895493865013123,
+      "eval_logps/chosen": -0.5038881301879883,
+      "eval_logps/rejected": -0.973581850528717,
+      "eval_loss": 0.6982013583183289,
+      "eval_rewards/accuracies": 0.6138613820075989,
+      "eval_rewards/chosen": -0.7558321952819824,
+      "eval_rewards/margins": 0.7045406699180603,
+      "eval_rewards/rejected": -1.460372805595398,
+      "eval_runtime": 29.0993,
+      "eval_samples_per_second": 27.526,
+      "eval_steps_per_second": 3.471,
+      "step": 250
+    },
+    {
+      "epoch": 0.20988900100908173,
+      "grad_norm": 0.734251856803894,
+      "learning_rate": 4.638410650401267e-06,
+      "logits/chosen": 2.4274039268493652,
+      "logits/rejected": 1.498230218887329,
+      "logps/chosen": -0.5549699068069458,
+      "logps/rejected": -0.9348627328872681,
+      "loss": 0.7149,
+      "rewards/accuracies": 0.5249999761581421,
+      "rewards/chosen": -0.8324548602104187,
+      "rewards/margins": 0.5698392987251282,
+      "rewards/rejected": -1.4022941589355469,
+      "step": 260
+    },
+    {
+      "epoch": 0.21796165489404642,
+      "grad_norm": 0.4445085823535919,
+      "learning_rate": 4.610819813755038e-06,
+      "logits/chosen": 2.042858600616455,
+      "logits/rejected": 1.1695036888122559,
+      "logps/chosen": -0.5573975443840027,
+      "logps/rejected": -1.15065598487854,
+      "loss": 0.6811,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.8360962867736816,
+      "rewards/margins": 0.889887809753418,
+      "rewards/rejected": -1.72598397731781,
+      "step": 270
+    },
+    {
+      "epoch": 0.2260343087790111,
+      "grad_norm": 0.5875476598739624,
+      "learning_rate": 4.582303101775249e-06,
+      "logits/chosen": 1.3029518127441406,
+      "logits/rejected": 0.3808741271495819,
+      "logps/chosen": -0.6656169891357422,
+      "logps/rejected": -1.5305824279785156,
+      "loss": 0.637,
+      "rewards/accuracies": 0.625,
+      "rewards/chosen": -0.9984253644943237,
+      "rewards/margins": 1.2974482774734497,
+      "rewards/rejected": -2.2958736419677734,
+      "step": 280
+    },
+    {
+      "epoch": 0.2341069626639758,
+      "grad_norm": 0.3814420998096466,
+      "learning_rate": 4.55287302283426e-06,
+      "logits/chosen": 2.0200212001800537,
+      "logits/rejected": 1.2681838274002075,
+      "logps/chosen": -0.6776013970375061,
+      "logps/rejected": -1.3369777202606201,
+      "loss": 0.63,
+      "rewards/accuracies": 0.5874999761581421,
+      "rewards/chosen": -1.016402006149292,
+      "rewards/margins": 0.9890643358230591,
+      "rewards/rejected": -2.0054664611816406,
+      "step": 290
+    },
+    {
+      "epoch": 0.24217961654894046,
+      "grad_norm": 0.6442322731018066,
+      "learning_rate": 4.522542485937369e-06,
+      "logits/chosen": 1.6268196105957031,
+      "logits/rejected": 0.4954712390899658,
+      "logps/chosen": -0.8117648363113403,
+      "logps/rejected": -1.7705228328704834,
+      "loss": 0.6104,
+      "rewards/accuracies": 0.7749999761581421,
+      "rewards/chosen": -1.2176473140716553,
+      "rewards/margins": 1.4381370544433594,
+      "rewards/rejected": -2.6557843685150146,
+      "step": 300
+    },
+    {
+      "epoch": 0.24217961654894046,
+      "eval_logits/chosen": 1.4003607034683228,
+      "eval_logits/rejected": 0.46628010272979736,
+      "eval_logps/chosen": -0.7612115740776062,
+      "eval_logps/rejected": -1.6895866394042969,
+      "eval_loss": 0.5781419277191162,
+      "eval_rewards/accuracies": 0.6534653306007385,
+      "eval_rewards/chosen": -1.141817331314087,
+      "eval_rewards/margins": 1.392562747001648,
+      "eval_rewards/rejected": -2.5343799591064453,
+      "eval_runtime": 29.102,
+      "eval_samples_per_second": 27.524,
+      "eval_steps_per_second": 3.471,
+      "step": 300
+    },
+    {
+      "epoch": 0.25025227043390513,
+      "grad_norm": 0.7889758944511414,
+      "learning_rate": 4.491324795060491e-06,
+      "logits/chosen": 1.1702089309692383,
+      "logits/rejected": 0.354276180267334,
+      "logps/chosen": -0.8091352581977844,
+      "logps/rejected": -1.944819450378418,
+      "loss": 0.5807,
+      "rewards/accuracies": 0.6875,
+      "rewards/chosen": -1.2137027978897095,
+      "rewards/margins": 1.7035261392593384,
+      "rewards/rejected": -2.9172286987304688,
+      "step": 310
+    },
+    {
+      "epoch": 0.25832492431886983,
+      "grad_norm": 1.410145878791809,
+      "learning_rate": 4.4592336433146e-06,
+      "logits/chosen": 2.0981459617614746,
+      "logits/rejected": 1.2289329767227173,
+      "logps/chosen": -0.8151634931564331,
+      "logps/rejected": -1.977259874343872,
+      "loss": 0.5007,
+      "rewards/accuracies": 0.6625000238418579,
+      "rewards/chosen": -1.222745418548584,
+      "rewards/margins": 1.7431443929672241,
+      "rewards/rejected": -2.9658896923065186,
+      "step": 320
+    },
+    {
+      "epoch": 0.26639757820383453,
+      "grad_norm": 0.5564689040184021,
+      "learning_rate": 4.426283106939474e-06,
+      "logits/chosen": 1.7921768426895142,
+      "logits/rejected": 0.7705962061882019,
+      "logps/chosen": -0.9244564771652222,
+      "logps/rejected": -2.5274672508239746,
+      "loss": 0.4757,
+      "rewards/accuracies": 0.75,
+      "rewards/chosen": -1.3866846561431885,
+      "rewards/margins": 2.4045166969299316,
+      "rewards/rejected": -3.791201114654541,
+      "step": 330
+    },
+    {
+      "epoch": 0.27447023208879917,
+      "grad_norm": 0.7554243803024292,
+      "learning_rate": 4.3924876391293915e-06,
+      "logits/chosen": 1.947997808456421,
+      "logits/rejected": 0.9661592245101929,
+      "logps/chosen": -0.9154227375984192,
+      "logps/rejected": -3.0766491889953613,
+      "loss": 0.4581,
+      "rewards/accuracies": 0.7749999761581421,
+      "rewards/chosen": -1.3731342554092407,
+      "rewards/margins": 3.24183988571167,
+      "rewards/rejected": -4.614974021911621,
+      "step": 340
+    },
+    {
+      "epoch": 0.28254288597376387,
+      "grad_norm": 2.936426877975464,
+      "learning_rate": 4.357862063693486e-06,
+      "logits/chosen": 1.2846364974975586,
+      "logits/rejected": 0.9968118667602539,
+      "logps/chosen": -1.0272481441497803,
+      "logps/rejected": -2.6980624198913574,
+      "loss": 0.4989,
+      "rewards/accuracies": 0.7124999761581421,
+      "rewards/chosen": -1.54087233543396,
+      "rewards/margins": 2.506221294403076,
+      "rewards/rejected": -4.047093868255615,
+      "step": 350
+    },
+    {
+      "epoch": 0.28254288597376387,
+      "eval_logits/chosen": 1.841333031654358,
+      "eval_logits/rejected": 1.1651691198349,
+      "eval_logps/chosen": -0.9644113183021545,
+      "eval_logps/rejected": -2.631535053253174,
+      "eval_loss": 0.44154587388038635,
+      "eval_rewards/accuracies": 0.6732673048973083,
+      "eval_rewards/chosen": -1.4466170072555542,
+      "eval_rewards/margins": 2.500684976577759,
+      "eval_rewards/rejected": -3.9473025798797607,
+      "eval_runtime": 29.1062,
+      "eval_samples_per_second": 27.52,
+      "eval_steps_per_second": 3.47,
+      "step": 350
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 1500,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8.159619413432074e+17,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-350/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:448b13b23e63cb2ff5ca633312a3af2a8ddf68feacd12942d98b8afe864faf6a
+size 7224

checkpoint-350/zero_to_fp32.py

@@ -0,0 +1,674 @@
+#!/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 . output_dir/
+#   or
+#   python zero_to_fp32.py . output_dir/ --safe_serialization
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+import json
+from tqdm import tqdm
+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 tqdm(param_shapes.items(), desc='Gathering Sharded Weights'):
+        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_dir,
+                                               max_shard_size="5GB",
+                                               safe_serialization=False,
+                                               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_dir``: directory to the pytorch fp32 state_dict output files
+        - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB
+        - ``safe_serialization``:  whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).
+        - ``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
+    """
+    # Dependency pre-check
+    if safe_serialization:
+        try:
+            from safetensors.torch import save_file
+        except ImportError:
+            print('If you want to use `safe_serialization`, please `pip install safetensors`')
+            raise
+    if max_shard_size is not None:
+        try:
+            from huggingface_hub import split_torch_state_dict_into_shards
+        except ImportError:
+            print('If you want to use `max_shard_size`, please `pip install huggingface_hub`')
+            raise
+
+    # Convert zero checkpoint to state_dict
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+
+    # Shard the model if it is too big.
+    weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin"
+    if max_shard_size is not None:
+        filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors")
+        state_dict_split = split_torch_state_dict_into_shards(state_dict,
+                                                              filename_pattern=filename_pattern,
+                                                              max_shard_size=max_shard_size)
+    else:
+        from collections import namedtuple
+        StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"])
+        state_dict_split = StateDictSplit(is_sharded=False,
+                                          filename_to_tensors={weights_name: list(state_dict.keys())})
+
+    # Save the model
+    filename_to_tensors = state_dict_split.filename_to_tensors.items()
+    for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"):
+        shard = {tensor: state_dict[tensor].contiguous() for tensor in tensors}
+        output_path = os.path.join(output_dir, shard_file)
+        if safe_serialization:
+            save_file(shard, output_path, metadata={"format": "pt"})
+        else:
+            torch.save(shard, output_path)
+
+    # Save index if sharded
+    if state_dict_split.is_sharded:
+        index = {
+            "metadata": state_dict_split.metadata,
+            "weight_map": state_dict_split.tensor_to_filename,
+        }
+        save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json"
+        save_index_file = os.path.join(output_dir, save_index_file)
+        with open(save_index_file, "w", encoding="utf-8") as f:
+            content = json.dumps(index, indent=2, sort_keys=True) + "\n"
+            f.write(content)
+
+
+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_dir",
+                        type=str,
+                        help="directory to the pytorch fp32 state_dict output files"
+                        "(e.g. path/checkpoint-12-output/)")
+    parser.add_argument(
+        "--max_shard_size",
+        type=str,
+        default="5GB",
+        help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size"
+        "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`"
+        "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances"
+        "without CPU OOM issues.")
+    parser.add_argument(
+        "--safe_serialization",
+        default=False,
+        action='store_true',
+        help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).")
+    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_dir,
+                                               max_shard_size=args.max_shard_size,
+                                               safe_serialization=args.safe_serialization,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)