Upload folder using huggingface_hub

30/README.md

@@ -0,0 +1,202 @@
+---
+base_model: ibm-granite/granite-3.3-8b-base
+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 0.15.2

30/adapter_config.json

@@ -0,0 +1,39 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": null,
+  "base_model_name_or_path": "ibm-granite/granite-3.3-8b-base",
+  "bias": "none",
+  "corda_config": null,
+  "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": 256,
+  "lora_bias": false,
+  "lora_dropout": 0.05,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 128,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "gate_proj",
+    "q_proj",
+    "up_proj",
+    "k_proj",
+    "o_proj",
+    "down_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

30/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6bc1319eaca922acdd8ef73d77e7c919e74f8cd4071930e0d445ccbc2726614f
+size 791751704

30/latest

@@ -0,0 +1 @@
+global_step854

30/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d88a0a4bcbeb191c9d1cd9f518f83a0401b853e028c6414109c1dd86c2e69bfb
+size 15920

30/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dac0c2b6756257d92d596e76985f2d77ab290c11c274035d706419280efc6839
+size 15920

30/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf8f6fada1ae59ca8e0e6532c88166bfe80da726a597701f5cb98e10dd8357c4
+size 15920

30/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d072c3b38dd87b985770d8cc2a76ad2d5b0d83fcb7773949116fb1019f424c0d
+size 15920

30/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e72ea7dae0be5512ac981c38aec63c3c08927b838a2720b49e6ed4f4ca68b08
+size 15920

30/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1b1686261ab898fa2e038ccce1658ab3bffc6c5e45047b031123c723fac3060
+size 15920

30/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da30eefc1e87ac1f1d84f6f4074b1a5f8143281fe1a22eabe7179a5a9a46a7c3
+size 15920

30/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8431196fb458b6f1da19fb0bc733d93b4ace658d2f7b10f3836cec19da14a528
+size 15920

30/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:adbdd1ef41f9684bebce724bf18a5ee6dbfeda6ae42b959c381f8ca329ea5117
+size 1064

30/special_tokens_map.json

@@ -0,0 +1,51 @@
+{
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<fim_prefix>",
+    "<fim_middle>",
+    "<fim_suffix>",
+    "<fim_pad>",
+    "<filename>",
+    "<gh_stars>",
+    "<issue_start>",
+    "<issue_comment>",
+    "<issue_closed>",
+    "<jupyter_start>",
+    "<jupyter_text>",
+    "<jupyter_code>",
+    "<jupyter_output>",
+    "<empty_output>",
+    "<commit_before>",
+    "<commit_msg>",
+    "<commit_after>",
+    "<reponame>"
+  ],
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "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": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

30/tokenizer_config.json

@@ -0,0 +1,188 @@
+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "<|endoftext|>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "<fim_prefix>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "<fim_middle>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "<fim_suffix>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "<fim_pad>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "5": {
+      "content": "<filename>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "6": {
+      "content": "<gh_stars>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "7": {
+      "content": "<issue_start>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "8": {
+      "content": "<issue_comment>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "9": {
+      "content": "<issue_closed>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "10": {
+      "content": "<jupyter_start>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "11": {
+      "content": "<jupyter_text>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "12": {
+      "content": "<jupyter_code>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "13": {
+      "content": "<jupyter_output>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "14": {
+      "content": "<empty_output>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "15": {
+      "content": "<commit_before>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "16": {
+      "content": "<commit_msg>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "17": {
+      "content": "<commit_after>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "18": {
+      "content": "<reponame>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<fim_prefix>",
+    "<fim_middle>",
+    "<fim_suffix>",
+    "<fim_pad>",
+    "<filename>",
+    "<gh_stars>",
+    "<issue_start>",
+    "<issue_comment>",
+    "<issue_closed>",
+    "<jupyter_start>",
+    "<jupyter_text>",
+    "<jupyter_code>",
+    "<jupyter_output>",
+    "<empty_output>",
+    "<commit_before>",
+    "<commit_msg>",
+    "<commit_after>",
+    "<reponame>"
+  ],
+  "bos_token": "<|endoftext|>",
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "<|endoftext|>",
+  "extra_special_tokens": {},
+  "model_max_length": 8192,
+  "pad_token": "<|endoftext|>",
+  "padding_side": "left",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>",
+  "vocab_size": 49152
+}

30/trainer_state.json

@@ -0,0 +1,1224 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 854,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011723329425556858,
+      "grad_norm": 1.0167624950408936,
+      "learning_rate": 1.1214953271028036e-06,
+      "loss": 1.2603,
+      "step": 5
+    },
+    {
+      "epoch": 0.023446658851113716,
+      "grad_norm": 1.003950595855713,
+      "learning_rate": 2.5233644859813085e-06,
+      "loss": 1.2102,
+      "step": 10
+    },
+    {
+      "epoch": 0.035169988276670575,
+      "grad_norm": 0.7265106439590454,
+      "learning_rate": 3.925233644859813e-06,
+      "loss": 1.2177,
+      "step": 15
+    },
+    {
+      "epoch": 0.04689331770222743,
+      "grad_norm": 0.49719780683517456,
+      "learning_rate": 5.3271028037383174e-06,
+      "loss": 1.2017,
+      "step": 20
+    },
+    {
+      "epoch": 0.05861664712778429,
+      "grad_norm": 0.5746001601219177,
+      "learning_rate": 6.728971962616822e-06,
+      "loss": 1.2264,
+      "step": 25
+    },
+    {
+      "epoch": 0.07033997655334115,
+      "grad_norm": 0.4860682189464569,
+      "learning_rate": 8.130841121495327e-06,
+      "loss": 1.1947,
+      "step": 30
+    },
+    {
+      "epoch": 0.08206330597889801,
+      "grad_norm": 0.5203776359558105,
+      "learning_rate": 9.532710280373831e-06,
+      "loss": 1.227,
+      "step": 35
+    },
+    {
+      "epoch": 0.09378663540445487,
+      "grad_norm": 0.4837038218975067,
+      "learning_rate": 1.0934579439252336e-05,
+      "loss": 1.1814,
+      "step": 40
+    },
+    {
+      "epoch": 0.10550996483001172,
+      "grad_norm": 0.4256620705127716,
+      "learning_rate": 1.233644859813084e-05,
+      "loss": 1.1335,
+      "step": 45
+    },
+    {
+      "epoch": 0.11723329425556858,
+      "grad_norm": 0.4722168445587158,
+      "learning_rate": 1.3738317757009345e-05,
+      "loss": 1.1764,
+      "step": 50
+    },
+    {
+      "epoch": 0.12895662368112543,
+      "grad_norm": 0.6030006408691406,
+      "learning_rate": 1.5140186915887848e-05,
+      "loss": 1.1571,
+      "step": 55
+    },
+    {
+      "epoch": 0.1406799531066823,
+      "grad_norm": 0.44915902614593506,
+      "learning_rate": 1.6542056074766353e-05,
+      "loss": 1.1779,
+      "step": 60
+    },
+    {
+      "epoch": 0.15240328253223914,
+      "grad_norm": 0.4629223942756653,
+      "learning_rate": 1.7943925233644858e-05,
+      "loss": 1.1396,
+      "step": 65
+    },
+    {
+      "epoch": 0.16412661195779601,
+      "grad_norm": 0.3984740674495697,
+      "learning_rate": 1.9345794392523363e-05,
+      "loss": 1.0817,
+      "step": 70
+    },
+    {
+      "epoch": 0.17584994138335286,
+      "grad_norm": 0.4412227272987366,
+      "learning_rate": 2.0747663551401867e-05,
+      "loss": 1.1101,
+      "step": 75
+    },
+    {
+      "epoch": 0.18757327080890973,
+      "grad_norm": 0.44576549530029297,
+      "learning_rate": 2.2149532710280372e-05,
+      "loss": 1.1586,
+      "step": 80
+    },
+    {
+      "epoch": 0.19929660023446658,
+      "grad_norm": 0.5096731185913086,
+      "learning_rate": 2.355140186915888e-05,
+      "loss": 1.1397,
+      "step": 85
+    },
+    {
+      "epoch": 0.21101992966002345,
+      "grad_norm": 0.7355549335479736,
+      "learning_rate": 2.4953271028037385e-05,
+      "loss": 1.104,
+      "step": 90
+    },
+    {
+      "epoch": 0.2227432590855803,
+      "grad_norm": 0.4670790135860443,
+      "learning_rate": 2.635514018691589e-05,
+      "loss": 1.0484,
+      "step": 95
+    },
+    {
+      "epoch": 0.23446658851113716,
+      "grad_norm": 0.4539933204650879,
+      "learning_rate": 2.7757009345794394e-05,
+      "loss": 1.0425,
+      "step": 100
+    },
+    {
+      "epoch": 0.246189917936694,
+      "grad_norm": 0.43292170763015747,
+      "learning_rate": 2.91588785046729e-05,
+      "loss": 1.0667,
+      "step": 105
+    },
+    {
+      "epoch": 0.25791324736225085,
+      "grad_norm": 0.4727552831172943,
+      "learning_rate": 2.9999928007915035e-05,
+      "loss": 1.0815,
+      "step": 110
+    },
+    {
+      "epoch": 0.2696365767878077,
+      "grad_norm": 0.4670151174068451,
+      "learning_rate": 2.9999118104895362e-05,
+      "loss": 1.0226,
+      "step": 115
+    },
+    {
+      "epoch": 0.2813599062133646,
+      "grad_norm": 0.6354214549064636,
+      "learning_rate": 2.999740835750041e-05,
+      "loss": 1.0837,
+      "step": 120
+    },
+    {
+      "epoch": 0.29308323563892147,
+      "grad_norm": 0.5098438262939453,
+      "learning_rate": 2.999479886830331e-05,
+      "loss": 1.0596,
+      "step": 125
+    },
+    {
+      "epoch": 0.3048065650644783,
+      "grad_norm": 0.50492262840271,
+      "learning_rate": 2.9991289793855547e-05,
+      "loss": 1.0148,
+      "step": 130
+    },
+    {
+      "epoch": 0.31652989449003516,
+      "grad_norm": 0.6335922479629517,
+      "learning_rate": 2.998688134467756e-05,
+      "loss": 1.0116,
+      "step": 135
+    },
+    {
+      "epoch": 0.32825322391559203,
+      "grad_norm": 0.5228518843650818,
+      "learning_rate": 2.998157378524611e-05,
+      "loss": 1.0064,
+      "step": 140
+    },
+    {
+      "epoch": 0.3399765533411489,
+      "grad_norm": 0.506489634513855,
+      "learning_rate": 2.9975367433978412e-05,
+      "loss": 0.9833,
+      "step": 145
+    },
+    {
+      "epoch": 0.3516998827667057,
+      "grad_norm": 0.5143541097640991,
+      "learning_rate": 2.996826266321305e-05,
+      "loss": 1.0063,
+      "step": 150
+    },
+    {
+      "epoch": 0.3634232121922626,
+      "grad_norm": 0.520057737827301,
+      "learning_rate": 2.9960259899187615e-05,
+      "loss": 0.9258,
+      "step": 155
+    },
+    {
+      "epoch": 0.37514654161781946,
+      "grad_norm": 0.6407086849212646,
+      "learning_rate": 2.995135962201315e-05,
+      "loss": 0.9591,
+      "step": 160
+    },
+    {
+      "epoch": 0.38686987104337633,
+      "grad_norm": 0.6441727876663208,
+      "learning_rate": 2.9941562365645334e-05,
+      "loss": 0.9368,
+      "step": 165
+    },
+    {
+      "epoch": 0.39859320046893315,
+      "grad_norm": 0.7262651920318604,
+      "learning_rate": 2.9930868717852458e-05,
+      "loss": 0.92,
+      "step": 170
+    },
+    {
+      "epoch": 0.41031652989449,
+      "grad_norm": 0.6420608758926392,
+      "learning_rate": 2.9919279320180168e-05,
+      "loss": 0.9794,
+      "step": 175
+    },
+    {
+      "epoch": 0.4220398593200469,
+      "grad_norm": 0.722493588924408,
+      "learning_rate": 2.9906794867912953e-05,
+      "loss": 0.9152,
+      "step": 180
+    },
+    {
+      "epoch": 0.43376318874560377,
+      "grad_norm": 0.6337221264839172,
+      "learning_rate": 2.989341611003247e-05,
+      "loss": 0.9169,
+      "step": 185
+    },
+    {
+      "epoch": 0.4454865181711606,
+      "grad_norm": 0.6387234330177307,
+      "learning_rate": 2.9879143849172567e-05,
+      "loss": 0.9026,
+      "step": 190
+    },
+    {
+      "epoch": 0.45720984759671746,
+      "grad_norm": 0.6389036774635315,
+      "learning_rate": 2.9863978941571166e-05,
+      "loss": 0.9135,
+      "step": 195
+    },
+    {
+      "epoch": 0.46893317702227433,
+      "grad_norm": 0.6367496848106384,
+      "learning_rate": 2.9847922297018875e-05,
+      "loss": 0.8923,
+      "step": 200
+    },
+    {
+      "epoch": 0.4806565064478312,
+      "grad_norm": 0.6974584460258484,
+      "learning_rate": 2.9830974878804416e-05,
+      "loss": 0.9135,
+      "step": 205
+    },
+    {
+      "epoch": 0.492379835873388,
+      "grad_norm": 0.624444842338562,
+      "learning_rate": 2.9813137703656828e-05,
+      "loss": 0.9045,
+      "step": 210
+    },
+    {
+      "epoch": 0.5041031652989449,
+      "grad_norm": 0.800079882144928,
+      "learning_rate": 2.979441184168448e-05,
+      "loss": 0.8796,
+      "step": 215
+    },
+    {
+      "epoch": 0.5158264947245017,
+      "grad_norm": 0.6745848059654236,
+      "learning_rate": 2.977479841631085e-05,
+      "loss": 0.9248,
+      "step": 220
+    },
+    {
+      "epoch": 0.5275498241500586,
+      "grad_norm": 0.7025408744812012,
+      "learning_rate": 2.9754298604207157e-05,
+      "loss": 0.8907,
+      "step": 225
+    },
+    {
+      "epoch": 0.5392731535756154,
+      "grad_norm": 0.6598261594772339,
+      "learning_rate": 2.9732913635221744e-05,
+      "loss": 0.8116,
+      "step": 230
+    },
+    {
+      "epoch": 0.5509964830011723,
+      "grad_norm": 0.7040488719940186,
+      "learning_rate": 2.9710644792306318e-05,
+      "loss": 0.835,
+      "step": 235
+    },
+    {
+      "epoch": 0.5627198124267292,
+      "grad_norm": 0.7311633229255676,
+      "learning_rate": 2.9687493411438954e-05,
+      "loss": 0.8546,
+      "step": 240
+    },
+    {
+      "epoch": 0.5744431418522861,
+      "grad_norm": 0.6836857795715332,
+      "learning_rate": 2.9663460881543972e-05,
+      "loss": 0.8712,
+      "step": 245
+    },
+    {
+      "epoch": 0.5861664712778429,
+      "grad_norm": 0.779832661151886,
+      "learning_rate": 2.9638548644408598e-05,
+      "loss": 0.8439,
+      "step": 250
+    },
+    {
+      "epoch": 0.5978898007033998,
+      "grad_norm": 0.7419824004173279,
+      "learning_rate": 2.9612758194596455e-05,
+      "loss": 0.7932,
+      "step": 255
+    },
+    {
+      "epoch": 0.6096131301289566,
+      "grad_norm": 0.755138635635376,
+      "learning_rate": 2.958609107935794e-05,
+      "loss": 0.8141,
+      "step": 260
+    },
+    {
+      "epoch": 0.6213364595545134,
+      "grad_norm": 0.7689526677131653,
+      "learning_rate": 2.9558548898537344e-05,
+      "loss": 0.8432,
+      "step": 265
+    },
+    {
+      "epoch": 0.6330597889800703,
+      "grad_norm": 0.8379843831062317,
+      "learning_rate": 2.9530133304476917e-05,
+      "loss": 0.8148,
+      "step": 270
+    },
+    {
+      "epoch": 0.6447831184056272,
+      "grad_norm": 0.7174904346466064,
+      "learning_rate": 2.9500846001917716e-05,
+      "loss": 0.8016,
+      "step": 275
+    },
+    {
+      "epoch": 0.6565064478311841,
+      "grad_norm": 0.7718838453292847,
+      "learning_rate": 2.9470688747897342e-05,
+      "loss": 0.8087,
+      "step": 280
+    },
+    {
+      "epoch": 0.6682297772567409,
+      "grad_norm": 0.7535898089408875,
+      "learning_rate": 2.9439663351644523e-05,
+      "loss": 0.897,
+      "step": 285
+    },
+    {
+      "epoch": 0.6799531066822978,
+      "grad_norm": 0.8394252061843872,
+      "learning_rate": 2.9407771674470585e-05,
+      "loss": 0.784,
+      "step": 290
+    },
+    {
+      "epoch": 0.6916764361078547,
+      "grad_norm": 0.9917625784873962,
+      "learning_rate": 2.9375015629657764e-05,
+      "loss": 0.7353,
+      "step": 295
+    },
+    {
+      "epoch": 0.7033997655334114,
+      "grad_norm": 0.9116941094398499,
+      "learning_rate": 2.9341397182344447e-05,
+      "loss": 0.7743,
+      "step": 300
+    },
+    {
+      "epoch": 0.7151230949589683,
+      "grad_norm": 0.7784647345542908,
+      "learning_rate": 2.9306918349407255e-05,
+      "loss": 0.8243,
+      "step": 305
+    },
+    {
+      "epoch": 0.7268464243845252,
+      "grad_norm": 0.8619483709335327,
+      "learning_rate": 2.9271581199340067e-05,
+      "loss": 0.7863,
+      "step": 310
+    },
+    {
+      "epoch": 0.738569753810082,
+      "grad_norm": 0.9895829558372498,
+      "learning_rate": 2.92353878521299e-05,
+      "loss": 0.7315,
+      "step": 315
+    },
+    {
+      "epoch": 0.7502930832356389,
+      "grad_norm": 0.8617073893547058,
+      "learning_rate": 2.9198340479129737e-05,
+      "loss": 0.7513,
+      "step": 320
+    },
+    {
+      "epoch": 0.7620164126611958,
+      "grad_norm": 0.8190014362335205,
+      "learning_rate": 2.9160441302928274e-05,
+      "loss": 0.6954,
+      "step": 325
+    },
+    {
+      "epoch": 0.7737397420867527,
+      "grad_norm": 1.011149287223816,
+      "learning_rate": 2.912169259721655e-05,
+      "loss": 0.7307,
+      "step": 330
+    },
+    {
+      "epoch": 0.7854630715123095,
+      "grad_norm": 0.9251821637153625,
+      "learning_rate": 2.908209668665157e-05,
+      "loss": 0.6944,
+      "step": 335
+    },
+    {
+      "epoch": 0.7971864009378663,
+      "grad_norm": 0.924638569355011,
+      "learning_rate": 2.9041655946716823e-05,
+      "loss": 0.6899,
+      "step": 340
+    },
+    {
+      "epoch": 0.8089097303634232,
+      "grad_norm": 0.8665716052055359,
+      "learning_rate": 2.900037280357977e-05,
+      "loss": 0.7343,
+      "step": 345
+    },
+    {
+      "epoch": 0.82063305978898,
+      "grad_norm": 0.8739374279975891,
+      "learning_rate": 2.8958249733946297e-05,
+      "loss": 0.6976,
+      "step": 350
+    },
+    {
+      "epoch": 0.8323563892145369,
+      "grad_norm": 0.7861272096633911,
+      "learning_rate": 2.8915289264912143e-05,
+      "loss": 0.7163,
+      "step": 355
+    },
+    {
+      "epoch": 0.8440797186400938,
+      "grad_norm": 0.9159606099128723,
+      "learning_rate": 2.887149397381126e-05,
+      "loss": 0.7008,
+      "step": 360
+    },
+    {
+      "epoch": 0.8558030480656507,
+      "grad_norm": 0.8858708143234253,
+      "learning_rate": 2.8826866488061208e-05,
+      "loss": 0.7147,
+      "step": 365
+    },
+    {
+      "epoch": 0.8675263774912075,
+      "grad_norm": 0.8299254179000854,
+      "learning_rate": 2.878140948500554e-05,
+      "loss": 0.6445,
+      "step": 370
+    },
+    {
+      "epoch": 0.8792497069167644,
+      "grad_norm": 0.8464540243148804,
+      "learning_rate": 2.8735125691753153e-05,
+      "loss": 0.6563,
+      "step": 375
+    },
+    {
+      "epoch": 0.8909730363423212,
+      "grad_norm": 0.908107578754425,
+      "learning_rate": 2.8688017885014698e-05,
+      "loss": 0.675,
+      "step": 380
+    },
+    {
+      "epoch": 0.902696365767878,
+      "grad_norm": 0.8772581219673157,
+      "learning_rate": 2.8640088890936e-05,
+      "loss": 0.6839,
+      "step": 385
+    },
+    {
+      "epoch": 0.9144196951934349,
+      "grad_norm": 0.8744145035743713,
+      "learning_rate": 2.8591341584928492e-05,
+      "loss": 0.6885,
+      "step": 390
+    },
+    {
+      "epoch": 0.9261430246189918,
+      "grad_norm": 0.9394176006317139,
+      "learning_rate": 2.854177889149672e-05,
+      "loss": 0.6802,
+      "step": 395
+    },
+    {
+      "epoch": 0.9378663540445487,
+      "grad_norm": 0.9551927447319031,
+      "learning_rate": 2.8491403784062903e-05,
+      "loss": 0.683,
+      "step": 400
+    },
+    {
+      "epoch": 0.9495896834701055,
+      "grad_norm": 0.85567706823349,
+      "learning_rate": 2.844021928478852e-05,
+      "loss": 0.6881,
+      "step": 405
+    },
+    {
+      "epoch": 0.9613130128956624,
+      "grad_norm": 0.8525532484054565,
+      "learning_rate": 2.838822846439303e-05,
+      "loss": 0.6391,
+      "step": 410
+    },
+    {
+      "epoch": 0.9730363423212193,
+      "grad_norm": 0.9406780004501343,
+      "learning_rate": 2.833543444196963e-05,
+      "loss": 0.6776,
+      "step": 415
+    },
+    {
+      "epoch": 0.984759671746776,
+      "grad_norm": 0.9766943454742432,
+      "learning_rate": 2.8281840384798147e-05,
+      "loss": 0.6413,
+      "step": 420
+    },
+    {
+      "epoch": 0.9964830011723329,
+      "grad_norm": 0.9217318296432495,
+      "learning_rate": 2.8227449508155012e-05,
+      "loss": 0.6083,
+      "step": 425
+    },
+    {
+      "epoch": 1.0070339976553342,
+      "grad_norm": 0.9758092761039734,
+      "learning_rate": 2.8172265075120356e-05,
+      "loss": 0.6333,
+      "step": 430
+    },
+    {
+      "epoch": 1.018757327080891,
+      "grad_norm": 0.9279267191886902,
+      "learning_rate": 2.8116290396382274e-05,
+      "loss": 0.5677,
+      "step": 435
+    },
+    {
+      "epoch": 1.030480656506448,
+      "grad_norm": 1.0296062231063843,
+      "learning_rate": 2.8059528830038188e-05,
+      "loss": 0.5534,
+      "step": 440
+    },
+    {
+      "epoch": 1.0422039859320047,
+      "grad_norm": 0.9819368720054626,
+      "learning_rate": 2.8001983781393387e-05,
+      "loss": 0.5744,
+      "step": 445
+    },
+    {
+      "epoch": 1.0539273153575615,
+      "grad_norm": 0.8835839629173279,
+      "learning_rate": 2.7943658702756728e-05,
+      "loss": 0.5895,
+      "step": 450
+    },
+    {
+      "epoch": 1.0656506447831184,
+      "grad_norm": 1.0302790403366089,
+      "learning_rate": 2.788455709323354e-05,
+      "loss": 0.5508,
+      "step": 455
+    },
+    {
+      "epoch": 1.0773739742086752,
+      "grad_norm": 0.9213964939117432,
+      "learning_rate": 2.7824682498515663e-05,
+      "loss": 0.5488,
+      "step": 460
+    },
+    {
+      "epoch": 1.0890973036342322,
+      "grad_norm": 0.822827160358429,
+      "learning_rate": 2.7764038510668786e-05,
+      "loss": 0.4908,
+      "step": 465
+    },
+    {
+      "epoch": 1.100820633059789,
+      "grad_norm": 0.8995427489280701,
+      "learning_rate": 2.770262876791689e-05,
+      "loss": 0.581,
+      "step": 470
+    },
+    {
+      "epoch": 1.112543962485346,
+      "grad_norm": 1.0242862701416016,
+      "learning_rate": 2.7640456954424027e-05,
+      "loss": 0.5482,
+      "step": 475
+    },
+    {
+      "epoch": 1.1242672919109027,
+      "grad_norm": 0.8724642992019653,
+      "learning_rate": 2.7577526800073257e-05,
+      "loss": 0.5637,
+      "step": 480
+    },
+    {
+      "epoch": 1.1359906213364597,
+      "grad_norm": 0.8943935632705688,
+      "learning_rate": 2.751384208024292e-05,
+      "loss": 0.5149,
+      "step": 485
+    },
+    {
+      "epoch": 1.1477139507620164,
+      "grad_norm": 0.9048977494239807,
+      "learning_rate": 2.7449406615580095e-05,
+      "loss": 0.5346,
+      "step": 490
+    },
+    {
+      "epoch": 1.1594372801875732,
+      "grad_norm": 1.0172462463378906,
+      "learning_rate": 2.7384224271771426e-05,
+      "loss": 0.5525,
+      "step": 495
+    },
+    {
+      "epoch": 1.1711606096131302,
+      "grad_norm": 0.8447465896606445,
+      "learning_rate": 2.7318298959311183e-05,
+      "loss": 0.5121,
+      "step": 500
+    },
+    {
+      "epoch": 1.182883939038687,
+      "grad_norm": 0.8748401999473572,
+      "learning_rate": 2.7251634633266668e-05,
+      "loss": 0.5696,
+      "step": 505
+    },
+    {
+      "epoch": 1.194607268464244,
+      "grad_norm": 0.8940523266792297,
+      "learning_rate": 2.718423529304095e-05,
+      "loss": 0.5077,
+      "step": 510
+    },
+    {
+      "epoch": 1.2063305978898007,
+      "grad_norm": 1.0012867450714111,
+      "learning_rate": 2.711610498213289e-05,
+      "loss": 0.492,
+      "step": 515
+    },
+    {
+      "epoch": 1.2180539273153577,
+      "grad_norm": 1.0748014450073242,
+      "learning_rate": 2.704724778789461e-05,
+      "loss": 0.538,
+      "step": 520
+    },
+    {
+      "epoch": 1.2297772567409144,
+      "grad_norm": 1.0266234874725342,
+      "learning_rate": 2.697766784128624e-05,
+      "loss": 0.5203,
+      "step": 525
+    },
+    {
+      "epoch": 1.2415005861664712,
+      "grad_norm": 0.9066937565803528,
+      "learning_rate": 2.6907369316628103e-05,
+      "loss": 0.4937,
+      "step": 530
+    },
+    {
+      "epoch": 1.2532239155920282,
+      "grad_norm": 0.9238837361335754,
+      "learning_rate": 2.6836356431350298e-05,
+      "loss": 0.5185,
+      "step": 535
+    },
+    {
+      "epoch": 1.264947245017585,
+      "grad_norm": 1.0094565153121948,
+      "learning_rate": 2.676463344573965e-05,
+      "loss": 0.4904,
+      "step": 540
+    },
+    {
+      "epoch": 1.276670574443142,
+      "grad_norm": 1.3660484552383423,
+      "learning_rate": 2.6692204662684154e-05,
+      "loss": 0.4977,
+      "step": 545
+    },
+    {
+      "epoch": 1.2883939038686987,
+      "grad_norm": 1.0245323181152344,
+      "learning_rate": 2.6619074427414817e-05,
+      "loss": 0.4955,
+      "step": 550
+    },
+    {
+      "epoch": 1.3001172332942557,
+      "grad_norm": 0.8795239329338074,
+      "learning_rate": 2.6545247127244975e-05,
+      "loss": 0.5172,
+      "step": 555
+    },
+    {
+      "epoch": 1.3118405627198124,
+      "grad_norm": 1.1408592462539673,
+      "learning_rate": 2.647072719130708e-05,
+      "loss": 0.5139,
+      "step": 560
+    },
+    {
+      "epoch": 1.3235638921453692,
+      "grad_norm": 0.9821284413337708,
+      "learning_rate": 2.639551909028699e-05,
+      "loss": 0.4778,
+      "step": 565
+    },
+    {
+      "epoch": 1.3352872215709262,
+      "grad_norm": 0.9982385635375977,
+      "learning_rate": 2.6319627336155757e-05,
+      "loss": 0.5432,
+      "step": 570
+    },
+    {
+      "epoch": 1.347010550996483,
+      "grad_norm": 0.9764225482940674,
+      "learning_rate": 2.6243056481898937e-05,
+      "loss": 0.4723,
+      "step": 575
+    },
+    {
+      "epoch": 1.35873388042204,
+      "grad_norm": 0.9747533202171326,
+      "learning_rate": 2.6165811121243442e-05,
+      "loss": 0.4889,
+      "step": 580
+    },
+    {
+      "epoch": 1.3704572098475967,
+      "grad_norm": 0.9515569806098938,
+      "learning_rate": 2.6087895888381953e-05,
+      "loss": 0.4758,
+      "step": 585
+    },
+    {
+      "epoch": 1.3821805392731537,
+      "grad_norm": 1.0449175834655762,
+      "learning_rate": 2.6009315457694893e-05,
+      "loss": 0.4579,
+      "step": 590
+    },
+    {
+      "epoch": 1.3939038686987104,
+      "grad_norm": 1.0853800773620605,
+      "learning_rate": 2.5930074543469996e-05,
+      "loss": 0.4906,
+      "step": 595
+    },
+    {
+      "epoch": 1.4056271981242672,
+      "grad_norm": 1.2414027452468872,
+      "learning_rate": 2.5850177899619493e-05,
+      "loss": 0.4848,
+      "step": 600
+    },
+    {
+      "epoch": 1.4173505275498242,
+      "grad_norm": 0.9966293573379517,
+      "learning_rate": 2.5769630319394898e-05,
+      "loss": 0.4616,
+      "step": 605
+    },
+    {
+      "epoch": 1.4290738569753811,
+      "grad_norm": 1.1271042823791504,
+      "learning_rate": 2.5688436635099454e-05,
+      "loss": 0.4419,
+      "step": 610
+    },
+    {
+      "epoch": 1.440797186400938,
+      "grad_norm": 0.9376438856124878,
+      "learning_rate": 2.5606601717798212e-05,
+      "loss": 0.4418,
+      "step": 615
+    },
+    {
+      "epoch": 1.4525205158264947,
+      "grad_norm": 0.9620692729949951,
+      "learning_rate": 2.5524130477025832e-05,
+      "loss": 0.4359,
+      "step": 620
+    },
+    {
+      "epoch": 1.4642438452520516,
+      "grad_norm": 1.0992698669433594,
+      "learning_rate": 2.5441027860491996e-05,
+      "loss": 0.4916,
+      "step": 625
+    },
+    {
+      "epoch": 1.4759671746776084,
+      "grad_norm": 1.1124409437179565,
+      "learning_rate": 2.5357298853784647e-05,
+      "loss": 0.4339,
+      "step": 630
+    },
+    {
+      "epoch": 1.4876905041031652,
+      "grad_norm": 0.9310411214828491,
+      "learning_rate": 2.527294848007081e-05,
+      "loss": 0.4252,
+      "step": 635
+    },
+    {
+      "epoch": 1.4994138335287222,
+      "grad_norm": 1.0951443910598755,
+      "learning_rate": 2.5187981799795304e-05,
+      "loss": 0.4427,
+      "step": 640
+    },
+    {
+      "epoch": 1.5111371629542791,
+      "grad_norm": 0.9256319403648376,
+      "learning_rate": 2.5102403910377104e-05,
+      "loss": 0.4199,
+      "step": 645
+    },
+    {
+      "epoch": 1.522860492379836,
+      "grad_norm": 1.00556480884552,
+      "learning_rate": 2.501621994590356e-05,
+      "loss": 0.4612,
+      "step": 650
+    },
+    {
+      "epoch": 1.5345838218053927,
+      "grad_norm": 1.0466468334197998,
+      "learning_rate": 2.4929435076822366e-05,
+      "loss": 0.424,
+      "step": 655
+    },
+    {
+      "epoch": 1.5463071512309496,
+      "grad_norm": 1.121028184890747,
+      "learning_rate": 2.484205450963138e-05,
+      "loss": 0.4242,
+      "step": 660
+    },
+    {
+      "epoch": 1.5580304806565064,
+      "grad_norm": 1.1106789112091064,
+      "learning_rate": 2.4754083486566273e-05,
+      "loss": 0.4158,
+      "step": 665
+    },
+    {
+      "epoch": 1.5697538100820632,
+      "grad_norm": 1.1398696899414062,
+      "learning_rate": 2.4665527285286016e-05,
+      "loss": 0.4184,
+      "step": 670
+    },
+    {
+      "epoch": 1.5814771395076201,
+      "grad_norm": 1.0724526643753052,
+      "learning_rate": 2.4576391218556272e-05,
+      "loss": 0.4171,
+      "step": 675
+    },
+    {
+      "epoch": 1.5932004689331771,
+      "grad_norm": 1.181626558303833,
+      "learning_rate": 2.4486680633930658e-05,
+      "loss": 0.4238,
+      "step": 680
+    },
+    {
+      "epoch": 1.6049237983587339,
+      "grad_norm": 0.981561541557312,
+      "learning_rate": 2.4396400913429935e-05,
+      "loss": 0.385,
+      "step": 685
+    },
+    {
+      "epoch": 1.6166471277842906,
+      "grad_norm": 1.1553294658660889,
+      "learning_rate": 2.430555747321911e-05,
+      "loss": 0.3756,
+      "step": 690
+    },
+    {
+      "epoch": 1.6283704572098476,
+      "grad_norm": 1.1467562913894653,
+      "learning_rate": 2.4214155763282524e-05,
+      "loss": 0.4236,
+      "step": 695
+    },
+    {
+      "epoch": 1.6400937866354046,
+      "grad_norm": 1.1394572257995605,
+      "learning_rate": 2.4122201267096865e-05,
+      "loss": 0.4253,
+      "step": 700
+    },
+    {
+      "epoch": 1.6518171160609612,
+      "grad_norm": 1.0924723148345947,
+      "learning_rate": 2.402969950130222e-05,
+      "loss": 0.405,
+      "step": 705
+    },
+    {
+      "epoch": 1.6635404454865181,
+      "grad_norm": 0.964448869228363,
+      "learning_rate": 2.39366560153711e-05,
+      "loss": 0.3834,
+      "step": 710
+    },
+    {
+      "epoch": 1.6752637749120751,
+      "grad_norm": 1.0330153703689575,
+      "learning_rate": 2.384307639127553e-05,
+      "loss": 0.4008,
+      "step": 715
+    },
+    {
+      "epoch": 1.6869871043376319,
+      "grad_norm": 1.1188576221466064,
+      "learning_rate": 2.3748966243152127e-05,
+      "loss": 0.3738,
+      "step": 720
+    },
+    {
+      "epoch": 1.6987104337631886,
+      "grad_norm": 1.020409345626831,
+      "learning_rate": 2.3654331216965334e-05,
+      "loss": 0.4174,
+      "step": 725
+    },
+    {
+      "epoch": 1.7104337631887456,
+      "grad_norm": 1.2046691179275513,
+      "learning_rate": 2.3559176990168685e-05,
+      "loss": 0.399,
+      "step": 730
+    },
+    {
+      "epoch": 1.7221570926143026,
+      "grad_norm": 1.0990289449691772,
+      "learning_rate": 2.3463509271364184e-05,
+      "loss": 0.3899,
+      "step": 735
+    },
+    {
+      "epoch": 1.7338804220398594,
+      "grad_norm": 1.1505696773529053,
+      "learning_rate": 2.3367333799959853e-05,
+      "loss": 0.4048,
+      "step": 740
+    },
+    {
+      "epoch": 1.7456037514654161,
+      "grad_norm": 1.0134183168411255,
+      "learning_rate": 2.327065634582538e-05,
+      "loss": 0.3504,
+      "step": 745
+    },
+    {
+      "epoch": 1.7573270808909731,
+      "grad_norm": 0.9383838772773743,
+      "learning_rate": 2.3173482708945982e-05,
+      "loss": 0.3706,
+      "step": 750
+    },
+    {
+      "epoch": 1.7690504103165299,
+      "grad_norm": 0.9833553433418274,
+      "learning_rate": 2.3075818719074454e-05,
+      "loss": 0.3587,
+      "step": 755
+    },
+    {
+      "epoch": 1.7807737397420866,
+      "grad_norm": 1.1251225471496582,
+      "learning_rate": 2.2977670235381396e-05,
+      "loss": 0.4207,
+      "step": 760
+    },
+    {
+      "epoch": 1.7924970691676436,
+      "grad_norm": 0.9914352893829346,
+      "learning_rate": 2.2879043146103734e-05,
+      "loss": 0.3935,
+      "step": 765
+    },
+    {
+      "epoch": 1.8042203985932006,
+      "grad_norm": 0.9891310334205627,
+      "learning_rate": 2.2779943368191446e-05,
+      "loss": 0.3655,
+      "step": 770
+    },
+    {
+      "epoch": 1.8159437280187574,
+      "grad_norm": 1.0876110792160034,
+      "learning_rate": 2.268037684695259e-05,
+      "loss": 0.3887,
+      "step": 775
+    },
+    {
+      "epoch": 1.8276670574443141,
+      "grad_norm": 0.9549176096916199,
+      "learning_rate": 2.2580349555696624e-05,
+      "loss": 0.3585,
+      "step": 780
+    },
+    {
+      "epoch": 1.839390386869871,
+      "grad_norm": 1.1817312240600586,
+      "learning_rate": 2.2479867495376063e-05,
+      "loss": 0.3485,
+      "step": 785
+    },
+    {
+      "epoch": 1.8511137162954279,
+      "grad_norm": 0.9575828909873962,
+      "learning_rate": 2.237893669422644e-05,
+      "loss": 0.3891,
+      "step": 790
+    },
+    {
+      "epoch": 1.8628370457209846,
+      "grad_norm": 1.0091192722320557,
+      "learning_rate": 2.227756320740467e-05,
+      "loss": 0.3583,
+      "step": 795
+    },
+    {
+      "epoch": 1.8745603751465416,
+      "grad_norm": 1.1005250215530396,
+      "learning_rate": 2.2175753116625778e-05,
+      "loss": 0.3398,
+      "step": 800
+    },
+    {
+      "epoch": 1.8862837045720986,
+      "grad_norm": 1.1460726261138916,
+      "learning_rate": 2.2073512529798035e-05,
+      "loss": 0.3326,
+      "step": 805
+    },
+    {
+      "epoch": 1.8980070339976554,
+      "grad_norm": 1.0507593154907227,
+      "learning_rate": 2.197084758065653e-05,
+      "loss": 0.3276,
+      "step": 810
+    },
+    {
+      "epoch": 1.9097303634232121,
+      "grad_norm": 1.0259215831756592,
+      "learning_rate": 2.1867764428395176e-05,
+      "loss": 0.35,
+      "step": 815
+    },
+    {
+      "epoch": 1.921453692848769,
+      "grad_norm": 0.8954200744628906,
+      "learning_rate": 2.176426925729722e-05,
+      "loss": 0.3545,
+      "step": 820
+    },
+    {
+      "epoch": 1.9331770222743259,
+      "grad_norm": 0.9455352425575256,
+      "learning_rate": 2.166036827636421e-05,
+      "loss": 0.3505,
+      "step": 825
+    },
+    {
+      "epoch": 1.9449003516998826,
+      "grad_norm": 1.0381656885147095,
+      "learning_rate": 2.155606771894352e-05,
+      "loss": 0.3541,
+      "step": 830
+    },
+    {
+      "epoch": 1.9566236811254396,
+      "grad_norm": 1.0482879877090454,
+      "learning_rate": 2.1451373842354343e-05,
+      "loss": 0.3463,
+      "step": 835
+    },
+    {
+      "epoch": 1.9683470105509966,
+      "grad_norm": 1.111364722251892,
+      "learning_rate": 2.134629292751237e-05,
+      "loss": 0.3135,
+      "step": 840
+    },
+    {
+      "epoch": 1.9800703399765534,
+      "grad_norm": 1.089702844619751,
+      "learning_rate": 2.1240831278552914e-05,
+      "loss": 0.3445,
+      "step": 845
+    },
+    {
+      "epoch": 1.9917936694021101,
+      "grad_norm": 1.0103806257247925,
+      "learning_rate": 2.113499522245273e-05,
+      "loss": 0.3493,
+      "step": 850
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2135,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 2000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.3768850250724803e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

30/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:208407f94a322b1426d8bfa264aa21e0ace5d78b0e8c1d012b2bc19a05d298b4
+size 7736

30/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)