Upload folder using huggingface_hub

4/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

4/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": [
+    "k_proj",
+    "q_proj",
+    "v_proj",
+    "o_proj",
+    "gate_proj",
+    "up_proj",
+    "down_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

4/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39b46fb1063515c656f2e2e1fcbf48f25a2939a1c85d0712219561e0683b110e
+size 791751704

4/latest

@@ -0,0 +1 @@
+global_step894

4/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc0b3d34bfe9bc26b9bfdd738949a84e2686603b7d06269f72f54bb61c8dd135
+size 15920

4/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4a92bdcf6befda8c629d7e999989942cfdec5814eacfc510728a95a2b4d0fe9
+size 15920

4/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c20938814ca8218106c84487eef0cad6f0e141f7e368642e08f885064fce959
+size 15920

4/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5ffbb1fa7bdec7500e256de9070616c009b22912f87b8acaa470ee120646694
+size 15920

4/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a48916a39546ca414cff77317973922cae1cbd03f94297d59a32a386b8b0273
+size 15920

4/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca4fbb5047af217cac564917e9b553bf765d9153e27205d237f7418d1587027d
+size 15920

4/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f1a0932f643ca489372b800cde2e91571b1e995ba3231d14a01ee1e3848558bf
+size 15920

4/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df6c60b39ac608b8280eb6f871e91a751db8d3ecd14f5df4e24e222cb127ec52
+size 15920

4/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:93c1f01d54b50b61d90f2f120fbe974a5c1d11242bfd9364bac1bdc4e97b70f0
+size 1064

4/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
+  }
+}

4/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
+}

4/trainer_state.json

@@ -0,0 +1,1280 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 500,
+  "global_step": 894,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011198208286674132,
+      "grad_norm": 1.0978649854660034,
+      "learning_rate": 1.0714285714285714e-06,
+      "loss": 1.2808,
+      "step": 5
+    },
+    {
+      "epoch": 0.022396416573348264,
+      "grad_norm": 0.865485429763794,
+      "learning_rate": 2.410714285714286e-06,
+      "loss": 1.3219,
+      "step": 10
+    },
+    {
+      "epoch": 0.0335946248600224,
+      "grad_norm": 0.7268191576004028,
+      "learning_rate": 3.75e-06,
+      "loss": 1.2607,
+      "step": 15
+    },
+    {
+      "epoch": 0.04479283314669653,
+      "grad_norm": 0.6289479732513428,
+      "learning_rate": 5.0892857142857146e-06,
+      "loss": 1.2549,
+      "step": 20
+    },
+    {
+      "epoch": 0.055991041433370664,
+      "grad_norm": 0.5705351829528809,
+      "learning_rate": 6.428571428571429e-06,
+      "loss": 1.2545,
+      "step": 25
+    },
+    {
+      "epoch": 0.0671892497200448,
+      "grad_norm": 0.8970493078231812,
+      "learning_rate": 7.767857142857144e-06,
+      "loss": 1.2063,
+      "step": 30
+    },
+    {
+      "epoch": 0.07838745800671892,
+      "grad_norm": 0.5323396325111389,
+      "learning_rate": 9.107142857142856e-06,
+      "loss": 1.1831,
+      "step": 35
+    },
+    {
+      "epoch": 0.08958566629339305,
+      "grad_norm": 0.4824765622615814,
+      "learning_rate": 1.044642857142857e-05,
+      "loss": 1.1912,
+      "step": 40
+    },
+    {
+      "epoch": 0.10078387458006718,
+      "grad_norm": 0.5041835308074951,
+      "learning_rate": 1.1785714285714286e-05,
+      "loss": 1.1675,
+      "step": 45
+    },
+    {
+      "epoch": 0.11198208286674133,
+      "grad_norm": 0.4978722631931305,
+      "learning_rate": 1.3125e-05,
+      "loss": 1.1681,
+      "step": 50
+    },
+    {
+      "epoch": 0.12318029115341546,
+      "grad_norm": 0.428523451089859,
+      "learning_rate": 1.4464285714285715e-05,
+      "loss": 1.154,
+      "step": 55
+    },
+    {
+      "epoch": 0.1343784994400896,
+      "grad_norm": 0.4901936650276184,
+      "learning_rate": 1.580357142857143e-05,
+      "loss": 1.1488,
+      "step": 60
+    },
+    {
+      "epoch": 0.1455767077267637,
+      "grad_norm": 0.4443103075027466,
+      "learning_rate": 1.7142857142857142e-05,
+      "loss": 1.2096,
+      "step": 65
+    },
+    {
+      "epoch": 0.15677491601343785,
+      "grad_norm": 0.5249817967414856,
+      "learning_rate": 1.848214285714286e-05,
+      "loss": 1.1517,
+      "step": 70
+    },
+    {
+      "epoch": 0.167973124300112,
+      "grad_norm": 0.4530683755874634,
+      "learning_rate": 1.982142857142857e-05,
+      "loss": 1.1722,
+      "step": 75
+    },
+    {
+      "epoch": 0.1791713325867861,
+      "grad_norm": 0.47545140981674194,
+      "learning_rate": 2.1160714285714287e-05,
+      "loss": 1.1244,
+      "step": 80
+    },
+    {
+      "epoch": 0.19036954087346025,
+      "grad_norm": 0.4221179187297821,
+      "learning_rate": 2.25e-05,
+      "loss": 1.0931,
+      "step": 85
+    },
+    {
+      "epoch": 0.20156774916013437,
+      "grad_norm": 0.4603467285633087,
+      "learning_rate": 2.3839285714285713e-05,
+      "loss": 1.1179,
+      "step": 90
+    },
+    {
+      "epoch": 0.2127659574468085,
+      "grad_norm": 0.48785561323165894,
+      "learning_rate": 2.517857142857143e-05,
+      "loss": 1.1119,
+      "step": 95
+    },
+    {
+      "epoch": 0.22396416573348266,
+      "grad_norm": 0.46568208932876587,
+      "learning_rate": 2.6517857142857143e-05,
+      "loss": 1.1662,
+      "step": 100
+    },
+    {
+      "epoch": 0.23516237402015677,
+      "grad_norm": 0.43084895610809326,
+      "learning_rate": 2.7857142857142858e-05,
+      "loss": 1.1031,
+      "step": 105
+    },
+    {
+      "epoch": 0.24636058230683092,
+      "grad_norm": 0.46002936363220215,
+      "learning_rate": 2.9196428571428573e-05,
+      "loss": 1.1044,
+      "step": 110
+    },
+    {
+      "epoch": 0.25755879059350506,
+      "grad_norm": 0.5120033025741577,
+      "learning_rate": 2.9999934306758047e-05,
+      "loss": 1.0682,
+      "step": 115
+    },
+    {
+      "epoch": 0.2687569988801792,
+      "grad_norm": 0.5539029836654663,
+      "learning_rate": 2.9999195264394326e-05,
+      "loss": 1.0544,
+      "step": 120
+    },
+    {
+      "epoch": 0.2799552071668533,
+      "grad_norm": 0.5363937616348267,
+      "learning_rate": 2.9997635103707554e-05,
+      "loss": 1.0822,
+      "step": 125
+    },
+    {
+      "epoch": 0.2911534154535274,
+      "grad_norm": 0.49808958172798157,
+      "learning_rate": 2.999525391010742e-05,
+      "loss": 1.0351,
+      "step": 130
+    },
+    {
+      "epoch": 0.3023516237402016,
+      "grad_norm": 0.6337416768074036,
+      "learning_rate": 2.9992051813950364e-05,
+      "loss": 1.0445,
+      "step": 135
+    },
+    {
+      "epoch": 0.3135498320268757,
+      "grad_norm": 0.5711157321929932,
+      "learning_rate": 2.998802899053244e-05,
+      "loss": 1.0135,
+      "step": 140
+    },
+    {
+      "epoch": 0.3247480403135498,
+      "grad_norm": 0.6856613755226135,
+      "learning_rate": 2.998318566007973e-05,
+      "loss": 1.013,
+      "step": 145
+    },
+    {
+      "epoch": 0.335946248600224,
+      "grad_norm": 0.5463095307350159,
+      "learning_rate": 2.99775220877363e-05,
+      "loss": 1.0258,
+      "step": 150
+    },
+    {
+      "epoch": 0.3471444568868981,
+      "grad_norm": 0.6269132494926453,
+      "learning_rate": 2.9971038583549633e-05,
+      "loss": 1.0128,
+      "step": 155
+    },
+    {
+      "epoch": 0.3583426651735722,
+      "grad_norm": 0.5507866740226746,
+      "learning_rate": 2.9963735502453715e-05,
+      "loss": 0.9991,
+      "step": 160
+    },
+    {
+      "epoch": 0.36954087346024633,
+      "grad_norm": 0.6331682205200195,
+      "learning_rate": 2.995561324424958e-05,
+      "loss": 0.9721,
+      "step": 165
+    },
+    {
+      "epoch": 0.3807390817469205,
+      "grad_norm": 0.6370624303817749,
+      "learning_rate": 2.9946672253583415e-05,
+      "loss": 0.9839,
+      "step": 170
+    },
+    {
+      "epoch": 0.3919372900335946,
+      "grad_norm": 0.588683545589447,
+      "learning_rate": 2.9936913019922235e-05,
+      "loss": 0.9616,
+      "step": 175
+    },
+    {
+      "epoch": 0.40313549832026874,
+      "grad_norm": 0.6478800773620605,
+      "learning_rate": 2.9926336077527062e-05,
+      "loss": 0.9782,
+      "step": 180
+    },
+    {
+      "epoch": 0.4143337066069429,
+      "grad_norm": 0.6072382926940918,
+      "learning_rate": 2.9914942005423723e-05,
+      "loss": 0.9035,
+      "step": 185
+    },
+    {
+      "epoch": 0.425531914893617,
+      "grad_norm": 0.830245852470398,
+      "learning_rate": 2.9902731427371096e-05,
+      "loss": 0.9616,
+      "step": 190
+    },
+    {
+      "epoch": 0.43673012318029114,
+      "grad_norm": 0.7037926912307739,
+      "learning_rate": 2.9889705011827006e-05,
+      "loss": 0.9044,
+      "step": 195
+    },
+    {
+      "epoch": 0.4479283314669653,
+      "grad_norm": 0.689367949962616,
+      "learning_rate": 2.9875863471911608e-05,
+      "loss": 0.9293,
+      "step": 200
+    },
+    {
+      "epoch": 0.45912653975363943,
+      "grad_norm": 0.7107216715812683,
+      "learning_rate": 2.9861207565368363e-05,
+      "loss": 0.908,
+      "step": 205
+    },
+    {
+      "epoch": 0.47032474804031354,
+      "grad_norm": 0.7383051514625549,
+      "learning_rate": 2.9845738094522533e-05,
+      "loss": 0.9435,
+      "step": 210
+    },
+    {
+      "epoch": 0.48152295632698766,
+      "grad_norm": 0.7229745388031006,
+      "learning_rate": 2.9829455906237287e-05,
+      "loss": 0.8984,
+      "step": 215
+    },
+    {
+      "epoch": 0.49272116461366183,
+      "grad_norm": 0.746848464012146,
+      "learning_rate": 2.9812361891867325e-05,
+      "loss": 0.8979,
+      "step": 220
+    },
+    {
+      "epoch": 0.503919372900336,
+      "grad_norm": 0.723301112651825,
+      "learning_rate": 2.979445698721007e-05,
+      "loss": 0.9337,
+      "step": 225
+    },
+    {
+      "epoch": 0.5151175811870101,
+      "grad_norm": 0.7699582576751709,
+      "learning_rate": 2.9775742172454473e-05,
+      "loss": 0.9443,
+      "step": 230
+    },
+    {
+      "epoch": 0.5263157894736842,
+      "grad_norm": 0.7237251400947571,
+      "learning_rate": 2.9756218472127302e-05,
+      "loss": 0.8489,
+      "step": 235
+    },
+    {
+      "epoch": 0.5375139977603584,
+      "grad_norm": 0.8070515990257263,
+      "learning_rate": 2.9735886955037118e-05,
+      "loss": 0.8639,
+      "step": 240
+    },
+    {
+      "epoch": 0.5487122060470325,
+      "grad_norm": 0.7536729574203491,
+      "learning_rate": 2.9714748734215714e-05,
+      "loss": 0.8083,
+      "step": 245
+    },
+    {
+      "epoch": 0.5599104143337066,
+      "grad_norm": 0.8057552576065063,
+      "learning_rate": 2.969280496685719e-05,
+      "loss": 0.8714,
+      "step": 250
+    },
+    {
+      "epoch": 0.5711086226203808,
+      "grad_norm": 0.8718234896659851,
+      "learning_rate": 2.9670056854254634e-05,
+      "loss": 0.8651,
+      "step": 255
+    },
+    {
+      "epoch": 0.5823068309070548,
+      "grad_norm": 0.8957020044326782,
+      "learning_rate": 2.9646505641734324e-05,
+      "loss": 0.8428,
+      "step": 260
+    },
+    {
+      "epoch": 0.593505039193729,
+      "grad_norm": 0.7812408804893494,
+      "learning_rate": 2.9622152618587576e-05,
+      "loss": 0.8646,
+      "step": 265
+    },
+    {
+      "epoch": 0.6047032474804032,
+      "grad_norm": 0.918906033039093,
+      "learning_rate": 2.9596999118000145e-05,
+      "loss": 0.8891,
+      "step": 270
+    },
+    {
+      "epoch": 0.6159014557670772,
+      "grad_norm": 0.9143264293670654,
+      "learning_rate": 2.9571046516979256e-05,
+      "loss": 0.814,
+      "step": 275
+    },
+    {
+      "epoch": 0.6270996640537514,
+      "grad_norm": 0.9321989417076111,
+      "learning_rate": 2.954429623627821e-05,
+      "loss": 0.7757,
+      "step": 280
+    },
+    {
+      "epoch": 0.6382978723404256,
+      "grad_norm": 0.7887775301933289,
+      "learning_rate": 2.9516749740318623e-05,
+      "loss": 0.8416,
+      "step": 285
+    },
+    {
+      "epoch": 0.6494960806270996,
+      "grad_norm": 0.8512288331985474,
+      "learning_rate": 2.948840853711022e-05,
+      "loss": 0.8131,
+      "step": 290
+    },
+    {
+      "epoch": 0.6606942889137738,
+      "grad_norm": 0.8105642199516296,
+      "learning_rate": 2.9459274178168335e-05,
+      "loss": 0.8115,
+      "step": 295
+    },
+    {
+      "epoch": 0.671892497200448,
+      "grad_norm": 0.7953616976737976,
+      "learning_rate": 2.9429348258428933e-05,
+      "loss": 0.8266,
+      "step": 300
+    },
+    {
+      "epoch": 0.683090705487122,
+      "grad_norm": 2.0142805576324463,
+      "learning_rate": 2.9398632416161298e-05,
+      "loss": 0.7923,
+      "step": 305
+    },
+    {
+      "epoch": 0.6942889137737962,
+      "grad_norm": 0.9255501627922058,
+      "learning_rate": 2.936712833287837e-05,
+      "loss": 0.7464,
+      "step": 310
+    },
+    {
+      "epoch": 0.7054871220604704,
+      "grad_norm": 0.8493944406509399,
+      "learning_rate": 2.9334837733244686e-05,
+      "loss": 0.7682,
+      "step": 315
+    },
+    {
+      "epoch": 0.7166853303471444,
+      "grad_norm": 1.1492375135421753,
+      "learning_rate": 2.9301762384981944e-05,
+      "loss": 0.7543,
+      "step": 320
+    },
+    {
+      "epoch": 0.7278835386338186,
+      "grad_norm": 0.9279423952102661,
+      "learning_rate": 2.926790409877225e-05,
+      "loss": 0.7954,
+      "step": 325
+    },
+    {
+      "epoch": 0.7390817469204927,
+      "grad_norm": 0.8645827770233154,
+      "learning_rate": 2.9233264728158997e-05,
+      "loss": 0.8014,
+      "step": 330
+    },
+    {
+      "epoch": 0.7502799552071668,
+      "grad_norm": 0.8726319670677185,
+      "learning_rate": 2.9197846169445376e-05,
+      "loss": 0.7396,
+      "step": 335
+    },
+    {
+      "epoch": 0.761478163493841,
+      "grad_norm": 0.9269015192985535,
+      "learning_rate": 2.916165036159058e-05,
+      "loss": 0.7109,
+      "step": 340
+    },
+    {
+      "epoch": 0.7726763717805151,
+      "grad_norm": 0.9171246290206909,
+      "learning_rate": 2.912467928610366e-05,
+      "loss": 0.7041,
+      "step": 345
+    },
+    {
+      "epoch": 0.7838745800671892,
+      "grad_norm": 0.8157719373703003,
+      "learning_rate": 2.9086934966935015e-05,
+      "loss": 0.7498,
+      "step": 350
+    },
+    {
+      "epoch": 0.7950727883538634,
+      "grad_norm": 1.0310955047607422,
+      "learning_rate": 2.9048419470365656e-05,
+      "loss": 0.7538,
+      "step": 355
+    },
+    {
+      "epoch": 0.8062709966405375,
+      "grad_norm": 0.9810894131660461,
+      "learning_rate": 2.9009134904894015e-05,
+      "loss": 0.698,
+      "step": 360
+    },
+    {
+      "epoch": 0.8174692049272116,
+      "grad_norm": 0.8990864157676697,
+      "learning_rate": 2.8969083421120587e-05,
+      "loss": 0.6898,
+      "step": 365
+    },
+    {
+      "epoch": 0.8286674132138858,
+      "grad_norm": 0.8737045526504517,
+      "learning_rate": 2.892826721163013e-05,
+      "loss": 0.7175,
+      "step": 370
+    },
+    {
+      "epoch": 0.8398656215005599,
+      "grad_norm": 0.8609009981155396,
+      "learning_rate": 2.8886688510871706e-05,
+      "loss": 0.7466,
+      "step": 375
+    },
+    {
+      "epoch": 0.851063829787234,
+      "grad_norm": 0.9786498546600342,
+      "learning_rate": 2.884434959503628e-05,
+      "loss": 0.7221,
+      "step": 380
+    },
+    {
+      "epoch": 0.8622620380739082,
+      "grad_norm": 0.9019100069999695,
+      "learning_rate": 2.8801252781932194e-05,
+      "loss": 0.7047,
+      "step": 385
+    },
+    {
+      "epoch": 0.8734602463605823,
+      "grad_norm": 0.9129027128219604,
+      "learning_rate": 2.8757400430858193e-05,
+      "loss": 0.7235,
+      "step": 390
+    },
+    {
+      "epoch": 0.8846584546472565,
+      "grad_norm": 0.9829080700874329,
+      "learning_rate": 2.871279494247435e-05,
+      "loss": 0.6777,
+      "step": 395
+    },
+    {
+      "epoch": 0.8958566629339306,
+      "grad_norm": 0.9491150975227356,
+      "learning_rate": 2.8667438758670582e-05,
+      "loss": 0.7271,
+      "step": 400
+    },
+    {
+      "epoch": 0.9070548712206047,
+      "grad_norm": 0.9399238228797913,
+      "learning_rate": 2.8621334362433017e-05,
+      "loss": 0.6362,
+      "step": 405
+    },
+    {
+      "epoch": 0.9182530795072789,
+      "grad_norm": 0.9197304844856262,
+      "learning_rate": 2.857448427770802e-05,
+      "loss": 0.666,
+      "step": 410
+    },
+    {
+      "epoch": 0.9294512877939529,
+      "grad_norm": 0.9620344042778015,
+      "learning_rate": 2.8526891069264058e-05,
+      "loss": 0.677,
+      "step": 415
+    },
+    {
+      "epoch": 0.9406494960806271,
+      "grad_norm": 0.8626346588134766,
+      "learning_rate": 2.847855734255128e-05,
+      "loss": 0.6432,
+      "step": 420
+    },
+    {
+      "epoch": 0.9518477043673013,
+      "grad_norm": 0.9515690803527832,
+      "learning_rate": 2.8429485743558876e-05,
+      "loss": 0.6336,
+      "step": 425
+    },
+    {
+      "epoch": 0.9630459126539753,
+      "grad_norm": 1.0451574325561523,
+      "learning_rate": 2.8379678958670245e-05,
+      "loss": 0.6471,
+      "step": 430
+    },
+    {
+      "epoch": 0.9742441209406495,
+      "grad_norm": 0.9338580369949341,
+      "learning_rate": 2.8329139714515916e-05,
+      "loss": 0.6526,
+      "step": 435
+    },
+    {
+      "epoch": 0.9854423292273237,
+      "grad_norm": 1.034649133682251,
+      "learning_rate": 2.8277870777824278e-05,
+      "loss": 0.6459,
+      "step": 440
+    },
+    {
+      "epoch": 0.9966405375139977,
+      "grad_norm": 1.0309592485427856,
+      "learning_rate": 2.822587495527013e-05,
+      "loss": 0.6495,
+      "step": 445
+    },
+    {
+      "epoch": 1.0067189249720045,
+      "grad_norm": 0.9272093772888184,
+      "learning_rate": 2.817315509332102e-05,
+      "loss": 0.6473,
+      "step": 450
+    },
+    {
+      "epoch": 1.0179171332586787,
+      "grad_norm": 1.0917242765426636,
+      "learning_rate": 2.8119714078081428e-05,
+      "loss": 0.5854,
+      "step": 455
+    },
+    {
+      "epoch": 1.0291153415453527,
+      "grad_norm": 0.9767428040504456,
+      "learning_rate": 2.8065554835134766e-05,
+      "loss": 0.5468,
+      "step": 460
+    },
+    {
+      "epoch": 1.0403135498320268,
+      "grad_norm": 0.9543578028678894,
+      "learning_rate": 2.8010680329383213e-05,
+      "loss": 0.5365,
+      "step": 465
+    },
+    {
+      "epoch": 1.051511758118701,
+      "grad_norm": 0.9958118200302124,
+      "learning_rate": 2.7955093564885412e-05,
+      "loss": 0.5602,
+      "step": 470
+    },
+    {
+      "epoch": 1.0627099664053752,
+      "grad_norm": 0.9233402609825134,
+      "learning_rate": 2.7898797584692003e-05,
+      "loss": 0.5523,
+      "step": 475
+    },
+    {
+      "epoch": 1.0739081746920494,
+      "grad_norm": 1.1012506484985352,
+      "learning_rate": 2.784179547067906e-05,
+      "loss": 0.5821,
+      "step": 480
+    },
+    {
+      "epoch": 1.0851063829787233,
+      "grad_norm": 0.9179466962814331,
+      "learning_rate": 2.7784090343379337e-05,
+      "loss": 0.5227,
+      "step": 485
+    },
+    {
+      "epoch": 1.0963045912653975,
+      "grad_norm": 1.1877224445343018,
+      "learning_rate": 2.7725685361811472e-05,
+      "loss": 0.5462,
+      "step": 490
+    },
+    {
+      "epoch": 1.1075027995520716,
+      "grad_norm": 0.9985042810440063,
+      "learning_rate": 2.7666583723307035e-05,
+      "loss": 0.5437,
+      "step": 495
+    },
+    {
+      "epoch": 1.1187010078387458,
+      "grad_norm": 1.211738109588623,
+      "learning_rate": 2.7606788663335498e-05,
+      "loss": 0.5293,
+      "step": 500
+    },
+    {
+      "epoch": 1.12989921612542,
+      "grad_norm": 1.0043666362762451,
+      "learning_rate": 2.7546303455327093e-05,
+      "loss": 0.5199,
+      "step": 505
+    },
+    {
+      "epoch": 1.1410974244120942,
+      "grad_norm": 1.134454369544983,
+      "learning_rate": 2.7485131410493644e-05,
+      "loss": 0.5272,
+      "step": 510
+    },
+    {
+      "epoch": 1.1522956326987681,
+      "grad_norm": 1.006064534187317,
+      "learning_rate": 2.742327587764726e-05,
+      "loss": 0.5601,
+      "step": 515
+    },
+    {
+      "epoch": 1.1634938409854423,
+      "grad_norm": 1.0664838552474976,
+      "learning_rate": 2.7360740243017042e-05,
+      "loss": 0.5324,
+      "step": 520
+    },
+    {
+      "epoch": 1.1746920492721165,
+      "grad_norm": 0.9521324634552002,
+      "learning_rate": 2.729752793006368e-05,
+      "loss": 0.5258,
+      "step": 525
+    },
+    {
+      "epoch": 1.1858902575587906,
+      "grad_norm": 1.0168689489364624,
+      "learning_rate": 2.723364239929206e-05,
+      "loss": 0.5062,
+      "step": 530
+    },
+    {
+      "epoch": 1.1970884658454648,
+      "grad_norm": 1.09732985496521,
+      "learning_rate": 2.71690871480618e-05,
+      "loss": 0.5249,
+      "step": 535
+    },
+    {
+      "epoch": 1.2082866741321387,
+      "grad_norm": 1.062366008758545,
+      "learning_rate": 2.7103865710395803e-05,
+      "loss": 0.5003,
+      "step": 540
+    },
+    {
+      "epoch": 1.219484882418813,
+      "grad_norm": 1.0284899473190308,
+      "learning_rate": 2.7037981656786802e-05,
+      "loss": 0.4832,
+      "step": 545
+    },
+    {
+      "epoch": 1.230683090705487,
+      "grad_norm": 1.1590033769607544,
+      "learning_rate": 2.6971438594001862e-05,
+      "loss": 0.502,
+      "step": 550
+    },
+    {
+      "epoch": 1.2418812989921613,
+      "grad_norm": 1.2048217058181763,
+      "learning_rate": 2.690424016488496e-05,
+      "loss": 0.5068,
+      "step": 555
+    },
+    {
+      "epoch": 1.2530795072788354,
+      "grad_norm": 1.0057001113891602,
+      "learning_rate": 2.6836390048157555e-05,
+      "loss": 0.501,
+      "step": 560
+    },
+    {
+      "epoch": 1.2642777155655094,
+      "grad_norm": 1.0815175771713257,
+      "learning_rate": 2.676789195821719e-05,
+      "loss": 0.4754,
+      "step": 565
+    },
+    {
+      "epoch": 1.2754759238521838,
+      "grad_norm": 0.9652005434036255,
+      "learning_rate": 2.6698749644934155e-05,
+      "loss": 0.4928,
+      "step": 570
+    },
+    {
+      "epoch": 1.2866741321388577,
+      "grad_norm": 1.0275248289108276,
+      "learning_rate": 2.6628966893446215e-05,
+      "loss": 0.4819,
+      "step": 575
+    },
+    {
+      "epoch": 1.297872340425532,
+      "grad_norm": 1.0320004224777222,
+      "learning_rate": 2.655854752395137e-05,
+      "loss": 0.5305,
+      "step": 580
+    },
+    {
+      "epoch": 1.309070548712206,
+      "grad_norm": 1.03120756149292,
+      "learning_rate": 2.6487495391498757e-05,
+      "loss": 0.4958,
+      "step": 585
+    },
+    {
+      "epoch": 1.3202687569988802,
+      "grad_norm": 0.9786254167556763,
+      "learning_rate": 2.6415814385777565e-05,
+      "loss": 0.4638,
+      "step": 590
+    },
+    {
+      "epoch": 1.3314669652855544,
+      "grad_norm": 1.0948402881622314,
+      "learning_rate": 2.634350843090414e-05,
+      "loss": 0.4689,
+      "step": 595
+    },
+    {
+      "epoch": 1.3426651735722284,
+      "grad_norm": 1.3032407760620117,
+      "learning_rate": 2.6270581485207137e-05,
+      "loss": 0.4767,
+      "step": 600
+    },
+    {
+      "epoch": 1.3538633818589025,
+      "grad_norm": 1.0457333326339722,
+      "learning_rate": 2.6197037541010827e-05,
+      "loss": 0.4747,
+      "step": 605
+    },
+    {
+      "epoch": 1.3650615901455767,
+      "grad_norm": 0.9719673991203308,
+      "learning_rate": 2.6122880624416553e-05,
+      "loss": 0.4693,
+      "step": 610
+    },
+    {
+      "epoch": 1.3762597984322509,
+      "grad_norm": 1.0594805479049683,
+      "learning_rate": 2.604811479508231e-05,
+      "loss": 0.4199,
+      "step": 615
+    },
+    {
+      "epoch": 1.387458006718925,
+      "grad_norm": 1.0848315954208374,
+      "learning_rate": 2.5972744146000504e-05,
+      "loss": 0.4375,
+      "step": 620
+    },
+    {
+      "epoch": 1.398656215005599,
+      "grad_norm": 1.1041808128356934,
+      "learning_rate": 2.5896772803273903e-05,
+      "loss": 0.5127,
+      "step": 625
+    },
+    {
+      "epoch": 1.4098544232922732,
+      "grad_norm": 1.0456072092056274,
+      "learning_rate": 2.582020492588973e-05,
+      "loss": 0.484,
+      "step": 630
+    },
+    {
+      "epoch": 1.4210526315789473,
+      "grad_norm": 0.9886185526847839,
+      "learning_rate": 2.574304470549201e-05,
+      "loss": 0.4442,
+      "step": 635
+    },
+    {
+      "epoch": 1.4322508398656215,
+      "grad_norm": 1.148207426071167,
+      "learning_rate": 2.566529636615207e-05,
+      "loss": 0.4223,
+      "step": 640
+    },
+    {
+      "epoch": 1.4434490481522957,
+      "grad_norm": 0.9983200430870056,
+      "learning_rate": 2.5586964164137325e-05,
+      "loss": 0.4234,
+      "step": 645
+    },
+    {
+      "epoch": 1.4546472564389696,
+      "grad_norm": 1.0275447368621826,
+      "learning_rate": 2.5508052387678258e-05,
+      "loss": 0.4618,
+      "step": 650
+    },
+    {
+      "epoch": 1.465845464725644,
+      "grad_norm": 1.1754274368286133,
+      "learning_rate": 2.5428565356733672e-05,
+      "loss": 0.4258,
+      "step": 655
+    },
+    {
+      "epoch": 1.477043673012318,
+      "grad_norm": 1.3802127838134766,
+      "learning_rate": 2.534850742275418e-05,
+      "loss": 0.4388,
+      "step": 660
+    },
+    {
+      "epoch": 1.4882418812989922,
+      "grad_norm": 1.1148616075515747,
+      "learning_rate": 2.5267882968444017e-05,
+      "loss": 0.4366,
+      "step": 665
+    },
+    {
+      "epoch": 1.4994400895856663,
+      "grad_norm": 1.0787454843521118,
+      "learning_rate": 2.518669640752109e-05,
+      "loss": 0.4121,
+      "step": 670
+    },
+    {
+      "epoch": 1.5106382978723403,
+      "grad_norm": 1.1478720903396606,
+      "learning_rate": 2.5104952184475346e-05,
+      "loss": 0.4201,
+      "step": 675
+    },
+    {
+      "epoch": 1.5218365061590147,
+      "grad_norm": 0.9818703532218933,
+      "learning_rate": 2.5022654774325507e-05,
+      "loss": 0.4343,
+      "step": 680
+    },
+    {
+      "epoch": 1.5330347144456886,
+      "grad_norm": 1.0454514026641846,
+      "learning_rate": 2.4939808682374028e-05,
+      "loss": 0.4594,
+      "step": 685
+    },
+    {
+      "epoch": 1.5442329227323628,
+      "grad_norm": 0.996567964553833,
+      "learning_rate": 2.48564184439605e-05,
+      "loss": 0.4373,
+      "step": 690
+    },
+    {
+      "epoch": 1.555431131019037,
+      "grad_norm": 1.1099140644073486,
+      "learning_rate": 2.4772488624213352e-05,
+      "loss": 0.4076,
+      "step": 695
+    },
+    {
+      "epoch": 1.5666293393057111,
+      "grad_norm": 1.1161383390426636,
+      "learning_rate": 2.4688023817799944e-05,
+      "loss": 0.4096,
+      "step": 700
+    },
+    {
+      "epoch": 1.5778275475923853,
+      "grad_norm": 1.0627449750900269,
+      "learning_rate": 2.460302864867502e-05,
+      "loss": 0.4008,
+      "step": 705
+    },
+    {
+      "epoch": 1.5890257558790593,
+      "grad_norm": 1.0296977758407593,
+      "learning_rate": 2.4517507769827598e-05,
+      "loss": 0.469,
+      "step": 710
+    },
+    {
+      "epoch": 1.6002239641657336,
+      "grad_norm": 1.1758477687835693,
+      "learning_rate": 2.4431465863026223e-05,
+      "loss": 0.3854,
+      "step": 715
+    },
+    {
+      "epoch": 1.6114221724524076,
+      "grad_norm": 1.0715276002883911,
+      "learning_rate": 2.434490763856268e-05,
+      "loss": 0.4006,
+      "step": 720
+    },
+    {
+      "epoch": 1.6226203807390818,
+      "grad_norm": 1.1095898151397705,
+      "learning_rate": 2.4257837834994123e-05,
+      "loss": 0.3902,
+      "step": 725
+    },
+    {
+      "epoch": 1.633818589025756,
+      "grad_norm": 1.100608468055725,
+      "learning_rate": 2.4170261218883686e-05,
+      "loss": 0.4034,
+      "step": 730
+    },
+    {
+      "epoch": 1.64501679731243,
+      "grad_norm": 1.1237363815307617,
+      "learning_rate": 2.4082182584539526e-05,
+      "loss": 0.4103,
+      "step": 735
+    },
+    {
+      "epoch": 1.6562150055991043,
+      "grad_norm": 1.0138239860534668,
+      "learning_rate": 2.3993606753752356e-05,
+      "loss": 0.3801,
+      "step": 740
+    },
+    {
+      "epoch": 1.6674132138857782,
+      "grad_norm": 1.0473990440368652,
+      "learning_rate": 2.39045385755315e-05,
+      "loss": 0.3642,
+      "step": 745
+    },
+    {
+      "epoch": 1.6786114221724524,
+      "grad_norm": 0.9463981986045837,
+      "learning_rate": 2.381498292583942e-05,
+      "loss": 0.4,
+      "step": 750
+    },
+    {
+      "epoch": 1.6898096304591266,
+      "grad_norm": 1.0383639335632324,
+      "learning_rate": 2.37249447073248e-05,
+      "loss": 0.4388,
+      "step": 755
+    },
+    {
+      "epoch": 1.7010078387458005,
+      "grad_norm": 1.0742672681808472,
+      "learning_rate": 2.3634428849054156e-05,
+      "loss": 0.3784,
+      "step": 760
+    },
+    {
+      "epoch": 1.712206047032475,
+      "grad_norm": 0.9601068496704102,
+      "learning_rate": 2.3543440306241965e-05,
+      "loss": 0.4132,
+      "step": 765
+    },
+    {
+      "epoch": 1.7234042553191489,
+      "grad_norm": 1.1224182844161987,
+      "learning_rate": 2.3451984059979444e-05,
+      "loss": 0.3628,
+      "step": 770
+    },
+    {
+      "epoch": 1.734602463605823,
+      "grad_norm": 1.1569995880126953,
+      "learning_rate": 2.336006511696184e-05,
+      "loss": 0.3678,
+      "step": 775
+    },
+    {
+      "epoch": 1.7458006718924972,
+      "grad_norm": 1.1215687990188599,
+      "learning_rate": 2.3267688509214335e-05,
+      "loss": 0.3658,
+      "step": 780
+    },
+    {
+      "epoch": 1.7569988801791714,
+      "grad_norm": 1.0910794734954834,
+      "learning_rate": 2.317485929381658e-05,
+      "loss": 0.3702,
+      "step": 785
+    },
+    {
+      "epoch": 1.7681970884658456,
+      "grad_norm": 1.161943793296814,
+      "learning_rate": 2.3081582552625867e-05,
+      "loss": 0.3851,
+      "step": 790
+    },
+    {
+      "epoch": 1.7793952967525195,
+      "grad_norm": 1.1181591749191284,
+      "learning_rate": 2.29878633919989e-05,
+      "loss": 0.3623,
+      "step": 795
+    },
+    {
+      "epoch": 1.7905935050391937,
+      "grad_norm": 0.9931478500366211,
+      "learning_rate": 2.2893706942512257e-05,
+      "loss": 0.3783,
+      "step": 800
+    },
+    {
+      "epoch": 1.8017917133258678,
+      "grad_norm": 1.125067114830017,
+      "learning_rate": 2.2799118358681535e-05,
+      "loss": 0.3615,
+      "step": 805
+    },
+    {
+      "epoch": 1.812989921612542,
+      "grad_norm": 1.1069613695144653,
+      "learning_rate": 2.2704102818679164e-05,
+      "loss": 0.3572,
+      "step": 810
+    },
+    {
+      "epoch": 1.8241881298992162,
+      "grad_norm": 0.9140533208847046,
+      "learning_rate": 2.2608665524050923e-05,
+      "loss": 0.3553,
+      "step": 815
+    },
+    {
+      "epoch": 1.8353863381858901,
+      "grad_norm": 0.987232506275177,
+      "learning_rate": 2.25128116994312e-05,
+      "loss": 0.3563,
+      "step": 820
+    },
+    {
+      "epoch": 1.8465845464725645,
+      "grad_norm": 1.1312346458435059,
+      "learning_rate": 2.241654659225696e-05,
+      "loss": 0.3563,
+      "step": 825
+    },
+    {
+      "epoch": 1.8577827547592385,
+      "grad_norm": 1.0891382694244385,
+      "learning_rate": 2.231987547248049e-05,
+      "loss": 0.3586,
+      "step": 830
+    },
+    {
+      "epoch": 1.8689809630459127,
+      "grad_norm": 1.134059190750122,
+      "learning_rate": 2.2222803632280894e-05,
+      "loss": 0.3539,
+      "step": 835
+    },
+    {
+      "epoch": 1.8801791713325868,
+      "grad_norm": 1.0670528411865234,
+      "learning_rate": 2.2125336385774385e-05,
+      "loss": 0.3506,
+      "step": 840
+    },
+    {
+      "epoch": 1.8913773796192608,
+      "grad_norm": 1.1700224876403809,
+      "learning_rate": 2.2027479068723345e-05,
+      "loss": 0.344,
+      "step": 845
+    },
+    {
+      "epoch": 1.9025755879059352,
+      "grad_norm": 1.1153310537338257,
+      "learning_rate": 2.1929237038244254e-05,
+      "loss": 0.335,
+      "step": 850
+    },
+    {
+      "epoch": 1.9137737961926091,
+      "grad_norm": 3.077934741973877,
+      "learning_rate": 2.1830615672514404e-05,
+      "loss": 0.3383,
+      "step": 855
+    },
+    {
+      "epoch": 1.9249720044792833,
+      "grad_norm": 0.9678553342819214,
+      "learning_rate": 2.1731620370477468e-05,
+      "loss": 0.337,
+      "step": 860
+    },
+    {
+      "epoch": 1.9361702127659575,
+      "grad_norm": 0.9908361434936523,
+      "learning_rate": 2.1632256551547952e-05,
+      "loss": 0.3673,
+      "step": 865
+    },
+    {
+      "epoch": 1.9473684210526314,
+      "grad_norm": 1.0404701232910156,
+      "learning_rate": 2.1532529655314514e-05,
+      "loss": 0.3276,
+      "step": 870
+    },
+    {
+      "epoch": 1.9585666293393058,
+      "grad_norm": 0.9922332763671875,
+      "learning_rate": 2.1432445141242166e-05,
+      "loss": 0.3408,
+      "step": 875
+    },
+    {
+      "epoch": 1.9697648376259798,
+      "grad_norm": 1.036054253578186,
+      "learning_rate": 2.1332008488373417e-05,
+      "loss": 0.3758,
+      "step": 880
+    },
+    {
+      "epoch": 1.980963045912654,
+      "grad_norm": 1.0212777853012085,
+      "learning_rate": 2.12312251950283e-05,
+      "loss": 0.3191,
+      "step": 885
+    },
+    {
+      "epoch": 1.992161254199328,
+      "grad_norm": 1.0531727075576782,
+      "learning_rate": 2.1130100778503407e-05,
+      "loss": 0.3344,
+      "step": 890
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2235,
+  "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": 1.352177971225428e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

4/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9317fa147bdb1a2d195af8ff71738fd12cc2987a716c383cacdd75619d1cb9e
+size 7736

4/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)