Upload folder using huggingface_hub

48_128_e5_3e-5/checkpoint-1176/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

48_128_e5_3e-5/checkpoint-1176/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": [
+    "o_proj",
+    "down_proj",
+    "v_proj",
+    "gate_proj",
+    "k_proj",
+    "q_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

48_128_e5_3e-5/checkpoint-1176/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cd1a063d8572056c6bd928e822e76607c80c1822197fc4264d187023665d89c4
+size 791751704

48_128_e5_3e-5/checkpoint-1176/latest

@@ -0,0 +1 @@
+global_step1176

48_128_e5_3e-5/checkpoint-1176/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1f803ed5f7b663eb62d108434eefb0ce0c1557aacef6f3905885e4b7599ce83
+size 15920

48_128_e5_3e-5/checkpoint-1176/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85e2c939e321d2efde2f698132d734c6a7b1dc8d862f6def20e0555b6f360c42
+size 15920

48_128_e5_3e-5/checkpoint-1176/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ed527aed147d3cfdb818f435599f291c1aaa92f2cd15eeb9ff28f3b6b526677
+size 15920

48_128_e5_3e-5/checkpoint-1176/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c86604a6f272d214acc2acfbdf922a4761e01c8bf94346d70b51ea695c1d1185
+size 15920

48_128_e5_3e-5/checkpoint-1176/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7fa3fb547354595959f3c78bb39d1227cee63247ac0694e9663adcd55d5b379
+size 15920

48_128_e5_3e-5/checkpoint-1176/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb668d04264bedf6cd0412c84a3236c681fb7100855645af5cfaea759163440b
+size 15920

48_128_e5_3e-5/checkpoint-1176/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0524ddc0ce8edd98f96c42b58d0f1425137877aa79444ac05b8a9d0740dd4b00
+size 15920

48_128_e5_3e-5/checkpoint-1176/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e7bb24ab461b5837eb85274af82454e07476aa50b0dc9b6ad90d55f8cb275a8
+size 15920

48_128_e5_3e-5/checkpoint-1176/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6959b7b09c0e2b04531ae5f256a4553a02e25b851702fdabfb77a6c1987824fd
+size 1064

48_128_e5_3e-5/checkpoint-1176/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
+  }
+}

48_128_e5_3e-5/checkpoint-1176/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
+}

48_128_e5_3e-5/checkpoint-1176/trainer_state.json

@@ -0,0 +1,1679 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1176,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.017006802721088437,
+      "grad_norm": 1.1845334768295288,
+      "learning_rate": 1.6216216216216219e-06,
+      "loss": 1.3293,
+      "step": 5
+    },
+    {
+      "epoch": 0.034013605442176874,
+      "grad_norm": 1.0181540250778198,
+      "learning_rate": 3.648648648648649e-06,
+      "loss": 1.2783,
+      "step": 10
+    },
+    {
+      "epoch": 0.05102040816326531,
+      "grad_norm": 0.6886288523674011,
+      "learning_rate": 5.675675675675676e-06,
+      "loss": 1.2906,
+      "step": 15
+    },
+    {
+      "epoch": 0.06802721088435375,
+      "grad_norm": 0.6959572434425354,
+      "learning_rate": 7.702702702702703e-06,
+      "loss": 1.2818,
+      "step": 20
+    },
+    {
+      "epoch": 0.08503401360544217,
+      "grad_norm": 0.6249067187309265,
+      "learning_rate": 9.72972972972973e-06,
+      "loss": 1.2667,
+      "step": 25
+    },
+    {
+      "epoch": 0.10204081632653061,
+      "grad_norm": 0.5868192911148071,
+      "learning_rate": 1.1756756756756757e-05,
+      "loss": 1.2477,
+      "step": 30
+    },
+    {
+      "epoch": 0.11904761904761904,
+      "grad_norm": 0.5181043148040771,
+      "learning_rate": 1.3783783783783784e-05,
+      "loss": 1.2207,
+      "step": 35
+    },
+    {
+      "epoch": 0.1360544217687075,
+      "grad_norm": 0.4317700266838074,
+      "learning_rate": 1.5810810810810808e-05,
+      "loss": 1.21,
+      "step": 40
+    },
+    {
+      "epoch": 0.15306122448979592,
+      "grad_norm": 0.4549165964126587,
+      "learning_rate": 1.783783783783784e-05,
+      "loss": 1.1368,
+      "step": 45
+    },
+    {
+      "epoch": 0.17006802721088435,
+      "grad_norm": 0.5308559536933899,
+      "learning_rate": 1.9864864864864866e-05,
+      "loss": 1.1875,
+      "step": 50
+    },
+    {
+      "epoch": 0.1870748299319728,
+      "grad_norm": 0.5514539480209351,
+      "learning_rate": 2.1891891891891892e-05,
+      "loss": 1.1593,
+      "step": 55
+    },
+    {
+      "epoch": 0.20408163265306123,
+      "grad_norm": 0.46228182315826416,
+      "learning_rate": 2.3918918918918917e-05,
+      "loss": 1.1345,
+      "step": 60
+    },
+    {
+      "epoch": 0.22108843537414966,
+      "grad_norm": 0.5649062395095825,
+      "learning_rate": 2.594594594594595e-05,
+      "loss": 1.1065,
+      "step": 65
+    },
+    {
+      "epoch": 0.23809523809523808,
+      "grad_norm": 0.49446532130241394,
+      "learning_rate": 2.7972972972972975e-05,
+      "loss": 1.1328,
+      "step": 70
+    },
+    {
+      "epoch": 0.25510204081632654,
+      "grad_norm": 0.5560067296028137,
+      "learning_rate": 3e-05,
+      "loss": 1.1697,
+      "step": 75
+    },
+    {
+      "epoch": 0.272108843537415,
+      "grad_norm": 0.4814499616622925,
+      "learning_rate": 2.999905043303196e-05,
+      "loss": 1.1659,
+      "step": 80
+    },
+    {
+      "epoch": 0.2891156462585034,
+      "grad_norm": 0.5074186325073242,
+      "learning_rate": 2.999620185235149e-05,
+      "loss": 1.0924,
+      "step": 85
+    },
+    {
+      "epoch": 0.30612244897959184,
+      "grad_norm": 0.592327892780304,
+      "learning_rate": 2.9991454618614338e-05,
+      "loss": 1.0793,
+      "step": 90
+    },
+    {
+      "epoch": 0.3231292517006803,
+      "grad_norm": 0.5456128120422363,
+      "learning_rate": 2.998480933286269e-05,
+      "loss": 1.0836,
+      "step": 95
+    },
+    {
+      "epoch": 0.3401360544217687,
+      "grad_norm": 0.6163952946662903,
+      "learning_rate": 2.9976266836449057e-05,
+      "loss": 1.0816,
+      "step": 100
+    },
+    {
+      "epoch": 0.35714285714285715,
+      "grad_norm": 0.5601843595504761,
+      "learning_rate": 2.9965828210929758e-05,
+      "loss": 1.091,
+      "step": 105
+    },
+    {
+      "epoch": 0.3741496598639456,
+      "grad_norm": 0.5945920944213867,
+      "learning_rate": 2.9953494777927995e-05,
+      "loss": 1.0349,
+      "step": 110
+    },
+    {
+      "epoch": 0.391156462585034,
+      "grad_norm": 0.5780587792396545,
+      "learning_rate": 2.993926809896651e-05,
+      "loss": 1.0389,
+      "step": 115
+    },
+    {
+      "epoch": 0.40816326530612246,
+      "grad_norm": 0.6240988969802856,
+      "learning_rate": 2.9923149975269885e-05,
+      "loss": 1.0521,
+      "step": 120
+    },
+    {
+      "epoch": 0.42517006802721086,
+      "grad_norm": 0.6076016426086426,
+      "learning_rate": 2.990514244753651e-05,
+      "loss": 0.9944,
+      "step": 125
+    },
+    {
+      "epoch": 0.4421768707482993,
+      "grad_norm": 0.6623475551605225,
+      "learning_rate": 2.988524779568018e-05,
+      "loss": 0.935,
+      "step": 130
+    },
+    {
+      "epoch": 0.45918367346938777,
+      "grad_norm": 0.6462083458900452,
+      "learning_rate": 2.9863468538541466e-05,
+      "loss": 1.0017,
+      "step": 135
+    },
+    {
+      "epoch": 0.47619047619047616,
+      "grad_norm": 0.6660473346710205,
+      "learning_rate": 2.9839807433568787e-05,
+      "loss": 0.8931,
+      "step": 140
+    },
+    {
+      "epoch": 0.4931972789115646,
+      "grad_norm": 0.680072546005249,
+      "learning_rate": 2.9814267476469304e-05,
+      "loss": 0.9786,
+      "step": 145
+    },
+    {
+      "epoch": 0.5102040816326531,
+      "grad_norm": 0.6932852268218994,
+      "learning_rate": 2.9786851900829633e-05,
+      "loss": 0.9335,
+      "step": 150
+    },
+    {
+      "epoch": 0.5272108843537415,
+      "grad_norm": 0.7095908522605896,
+      "learning_rate": 2.9757564177706448e-05,
+      "loss": 0.9278,
+      "step": 155
+    },
+    {
+      "epoch": 0.54421768707483,
+      "grad_norm": 0.7165976762771606,
+      "learning_rate": 2.972640801518701e-05,
+      "loss": 0.9029,
+      "step": 160
+    },
+    {
+      "epoch": 0.5612244897959183,
+      "grad_norm": 0.745604395866394,
+      "learning_rate": 2.969338735791968e-05,
+      "loss": 0.8675,
+      "step": 165
+    },
+    {
+      "epoch": 0.5782312925170068,
+      "grad_norm": 0.6811717748641968,
+      "learning_rate": 2.9658506386614525e-05,
+      "loss": 0.9125,
+      "step": 170
+    },
+    {
+      "epoch": 0.5952380952380952,
+      "grad_norm": 0.7211567163467407,
+      "learning_rate": 2.962176951751396e-05,
+      "loss": 0.8937,
+      "step": 175
+    },
+    {
+      "epoch": 0.6122448979591837,
+      "grad_norm": 0.7217546701431274,
+      "learning_rate": 2.958318140183364e-05,
+      "loss": 0.9474,
+      "step": 180
+    },
+    {
+      "epoch": 0.6292517006802721,
+      "grad_norm": 0.7711513042449951,
+      "learning_rate": 2.9542746925173566e-05,
+      "loss": 0.882,
+      "step": 185
+    },
+    {
+      "epoch": 0.6462585034013606,
+      "grad_norm": 0.8753324747085571,
+      "learning_rate": 2.9500471206899528e-05,
+      "loss": 0.874,
+      "step": 190
+    },
+    {
+      "epoch": 0.6632653061224489,
+      "grad_norm": 0.873315691947937,
+      "learning_rate": 2.945635959949494e-05,
+      "loss": 0.8252,
+      "step": 195
+    },
+    {
+      "epoch": 0.6802721088435374,
+      "grad_norm": 1.0285780429840088,
+      "learning_rate": 2.9410417687883173e-05,
+      "loss": 0.885,
+      "step": 200
+    },
+    {
+      "epoch": 0.6972789115646258,
+      "grad_norm": 0.8351210951805115,
+      "learning_rate": 2.936265128872046e-05,
+      "loss": 0.8756,
+      "step": 205
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 0.8613854646682739,
+      "learning_rate": 2.931306644965944e-05,
+      "loss": 0.8489,
+      "step": 210
+    },
+    {
+      "epoch": 0.7312925170068028,
+      "grad_norm": 0.9348616003990173,
+      "learning_rate": 2.9261669448583492e-05,
+      "loss": 0.8657,
+      "step": 215
+    },
+    {
+      "epoch": 0.7482993197278912,
+      "grad_norm": 1.0146417617797852,
+      "learning_rate": 2.9208466792811875e-05,
+      "loss": 0.815,
+      "step": 220
+    },
+    {
+      "epoch": 0.7653061224489796,
+      "grad_norm": 0.9285673499107361,
+      "learning_rate": 2.915346521827586e-05,
+      "loss": 0.8164,
+      "step": 225
+    },
+    {
+      "epoch": 0.782312925170068,
+      "grad_norm": 0.951054573059082,
+      "learning_rate": 2.9096671688665893e-05,
+      "loss": 0.8515,
+      "step": 230
+    },
+    {
+      "epoch": 0.7993197278911565,
+      "grad_norm": 0.9739980697631836,
+      "learning_rate": 2.9038093394549946e-05,
+      "loss": 0.7521,
+      "step": 235
+    },
+    {
+      "epoch": 0.8163265306122449,
+      "grad_norm": 1.0542875528335571,
+      "learning_rate": 2.8977737752463094e-05,
+      "loss": 0.8105,
+      "step": 240
+    },
+    {
+      "epoch": 0.8333333333333334,
+      "grad_norm": 0.9150990843772888,
+      "learning_rate": 2.891561240396855e-05,
+      "loss": 0.8032,
+      "step": 245
+    },
+    {
+      "epoch": 0.8503401360544217,
+      "grad_norm": 1.0160540342330933,
+      "learning_rate": 2.8851725214690155e-05,
+      "loss": 0.7096,
+      "step": 250
+    },
+    {
+      "epoch": 0.8673469387755102,
+      "grad_norm": 0.9452092051506042,
+      "learning_rate": 2.8786084273316524e-05,
+      "loss": 0.6958,
+      "step": 255
+    },
+    {
+      "epoch": 0.8843537414965986,
+      "grad_norm": 0.8861579895019531,
+      "learning_rate": 2.8718697890576944e-05,
+      "loss": 0.7085,
+      "step": 260
+    },
+    {
+      "epoch": 0.9013605442176871,
+      "grad_norm": 1.0629994869232178,
+      "learning_rate": 2.864957459818918e-05,
+      "loss": 0.6941,
+      "step": 265
+    },
+    {
+      "epoch": 0.9183673469387755,
+      "grad_norm": 0.9347429275512695,
+      "learning_rate": 2.8578723147779237e-05,
+      "loss": 0.689,
+      "step": 270
+    },
+    {
+      "epoch": 0.935374149659864,
+      "grad_norm": 1.0432939529418945,
+      "learning_rate": 2.850615250977339e-05,
+      "loss": 0.7042,
+      "step": 275
+    },
+    {
+      "epoch": 0.9523809523809523,
+      "grad_norm": 0.948538601398468,
+      "learning_rate": 2.843187187226239e-05,
+      "loss": 0.7113,
+      "step": 280
+    },
+    {
+      "epoch": 0.9693877551020408,
+      "grad_norm": 1.053505778312683,
+      "learning_rate": 2.835589063983821e-05,
+      "loss": 0.7173,
+      "step": 285
+    },
+    {
+      "epoch": 0.9863945578231292,
+      "grad_norm": 0.9371920824050903,
+      "learning_rate": 2.827821843240331e-05,
+      "loss": 0.6272,
+      "step": 290
+    },
+    {
+      "epoch": 1.0034013605442176,
+      "grad_norm": 1.1553608179092407,
+      "learning_rate": 2.8198865083952694e-05,
+      "loss": 0.6671,
+      "step": 295
+    },
+    {
+      "epoch": 1.0204081632653061,
+      "grad_norm": 1.0597134828567505,
+      "learning_rate": 2.811784064132883e-05,
+      "loss": 0.6007,
+      "step": 300
+    },
+    {
+      "epoch": 1.0374149659863945,
+      "grad_norm": 1.1430370807647705,
+      "learning_rate": 2.803515536294963e-05,
+      "loss": 0.6015,
+      "step": 305
+    },
+    {
+      "epoch": 1.054421768707483,
+      "grad_norm": 1.0705440044403076,
+      "learning_rate": 2.795081971750963e-05,
+      "loss": 0.5987,
+      "step": 310
+    },
+    {
+      "epoch": 1.0714285714285714,
+      "grad_norm": 1.1100271940231323,
+      "learning_rate": 2.786484438265459e-05,
+      "loss": 0.6058,
+      "step": 315
+    },
+    {
+      "epoch": 1.08843537414966,
+      "grad_norm": 1.1145716905593872,
+      "learning_rate": 2.7777240243629578e-05,
+      "loss": 0.5264,
+      "step": 320
+    },
+    {
+      "epoch": 1.1054421768707483,
+      "grad_norm": 1.0852850675582886,
+      "learning_rate": 2.7688018391900826e-05,
+      "loss": 0.549,
+      "step": 325
+    },
+    {
+      "epoch": 1.1224489795918366,
+      "grad_norm": 1.1767306327819824,
+      "learning_rate": 2.7597190123751422e-05,
+      "loss": 0.5997,
+      "step": 330
+    },
+    {
+      "epoch": 1.1394557823129252,
+      "grad_norm": 1.2128592729568481,
+      "learning_rate": 2.750476693885113e-05,
+      "loss": 0.5681,
+      "step": 335
+    },
+    {
+      "epoch": 1.1564625850340136,
+      "grad_norm": 1.313896894454956,
+      "learning_rate": 2.7410760538800408e-05,
+      "loss": 0.5698,
+      "step": 340
+    },
+    {
+      "epoch": 1.1734693877551021,
+      "grad_norm": 1.1703464984893799,
+      "learning_rate": 2.7315182825648895e-05,
+      "loss": 0.569,
+      "step": 345
+    },
+    {
+      "epoch": 1.1904761904761905,
+      "grad_norm": 1.1567600965499878,
+      "learning_rate": 2.7218045900388504e-05,
+      "loss": 0.5139,
+      "step": 350
+    },
+    {
+      "epoch": 1.2074829931972788,
+      "grad_norm": 1.097703456878662,
+      "learning_rate": 2.7119362061421303e-05,
+      "loss": 0.5093,
+      "step": 355
+    },
+    {
+      "epoch": 1.2244897959183674,
+      "grad_norm": 1.0932475328445435,
+      "learning_rate": 2.7019143803002465e-05,
+      "loss": 0.5731,
+      "step": 360
+    },
+    {
+      "epoch": 1.2414965986394557,
+      "grad_norm": 1.2174867391586304,
+      "learning_rate": 2.6917403813658364e-05,
+      "loss": 0.5077,
+      "step": 365
+    },
+    {
+      "epoch": 1.2585034013605443,
+      "grad_norm": 1.0711535215377808,
+      "learning_rate": 2.6814154974580092e-05,
+      "loss": 0.5205,
+      "step": 370
+    },
+    {
+      "epoch": 1.2755102040816326,
+      "grad_norm": 1.1222914457321167,
+      "learning_rate": 2.67094103579926e-05,
+      "loss": 0.5386,
+      "step": 375
+    },
+    {
+      "epoch": 1.2925170068027212,
+      "grad_norm": 1.1327167749404907,
+      "learning_rate": 2.6603183225499608e-05,
+      "loss": 0.5446,
+      "step": 380
+    },
+    {
+      "epoch": 1.3095238095238095,
+      "grad_norm": 1.2905243635177612,
+      "learning_rate": 2.6495487026404607e-05,
+      "loss": 0.5009,
+      "step": 385
+    },
+    {
+      "epoch": 1.3265306122448979,
+      "grad_norm": 1.059859037399292,
+      "learning_rate": 2.6386335396008033e-05,
+      "loss": 0.4772,
+      "step": 390
+    },
+    {
+      "epoch": 1.3435374149659864,
+      "grad_norm": 1.329107403755188,
+      "learning_rate": 2.6275742153880907e-05,
+      "loss": 0.5027,
+      "step": 395
+    },
+    {
+      "epoch": 1.3605442176870748,
+      "grad_norm": 1.1252979040145874,
+      "learning_rate": 2.6163721302115184e-05,
+      "loss": 0.4714,
+      "step": 400
+    },
+    {
+      "epoch": 1.3775510204081631,
+      "grad_norm": 1.1349155902862549,
+      "learning_rate": 2.6050287023550936e-05,
+      "loss": 0.4851,
+      "step": 405
+    },
+    {
+      "epoch": 1.3945578231292517,
+      "grad_norm": 1.0614500045776367,
+      "learning_rate": 2.59354536799807e-05,
+      "loss": 0.4893,
+      "step": 410
+    },
+    {
+      "epoch": 1.4115646258503403,
+      "grad_norm": 1.2005356550216675,
+      "learning_rate": 2.5819235810331115e-05,
+      "loss": 0.5233,
+      "step": 415
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 1.0654675960540771,
+      "learning_rate": 2.5701648128822205e-05,
+      "loss": 0.4876,
+      "step": 420
+    },
+    {
+      "epoch": 1.445578231292517,
+      "grad_norm": 1.3152744770050049,
+      "learning_rate": 2.55827055231044e-05,
+      "loss": 0.5113,
+      "step": 425
+    },
+    {
+      "epoch": 1.4625850340136055,
+      "grad_norm": 1.0852030515670776,
+      "learning_rate": 2.5462423052373628e-05,
+      "loss": 0.4794,
+      "step": 430
+    },
+    {
+      "epoch": 1.4795918367346939,
+      "grad_norm": 1.1766146421432495,
+      "learning_rate": 2.534081594546469e-05,
+      "loss": 0.4488,
+      "step": 435
+    },
+    {
+      "epoch": 1.4965986394557822,
+      "grad_norm": 1.2981992959976196,
+      "learning_rate": 2.5217899598923162e-05,
+      "loss": 0.4567,
+      "step": 440
+    },
+    {
+      "epoch": 1.5136054421768708,
+      "grad_norm": 1.0443862676620483,
+      "learning_rate": 2.5093689575056045e-05,
+      "loss": 0.4448,
+      "step": 445
+    },
+    {
+      "epoch": 1.5306122448979593,
+      "grad_norm": 1.130210280418396,
+      "learning_rate": 2.4968201599961445e-05,
+      "loss": 0.439,
+      "step": 450
+    },
+    {
+      "epoch": 1.5476190476190477,
+      "grad_norm": 1.3261239528656006,
+      "learning_rate": 2.4841451561537496e-05,
+      "loss": 0.4577,
+      "step": 455
+    },
+    {
+      "epoch": 1.564625850340136,
+      "grad_norm": 1.1579346656799316,
+      "learning_rate": 2.471345550747082e-05,
+      "loss": 0.4915,
+      "step": 460
+    },
+    {
+      "epoch": 1.5816326530612246,
+      "grad_norm": 1.1824404001235962,
+      "learning_rate": 2.4584229643204755e-05,
+      "loss": 0.4404,
+      "step": 465
+    },
+    {
+      "epoch": 1.598639455782313,
+      "grad_norm": 1.175586223602295,
+      "learning_rate": 2.4453790329887578e-05,
+      "loss": 0.4649,
+      "step": 470
+    },
+    {
+      "epoch": 1.6156462585034013,
+      "grad_norm": 1.5084452629089355,
+      "learning_rate": 2.4322154082301065e-05,
+      "loss": 0.4425,
+      "step": 475
+    },
+    {
+      "epoch": 1.6326530612244898,
+      "grad_norm": 1.2946665287017822,
+      "learning_rate": 2.4189337566769545e-05,
+      "loss": 0.47,
+      "step": 480
+    },
+    {
+      "epoch": 1.6496598639455784,
+      "grad_norm": 1.0975319147109985,
+      "learning_rate": 2.4055357599049807e-05,
+      "loss": 0.4747,
+      "step": 485
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 1.0334219932556152,
+      "learning_rate": 2.392023114220209e-05,
+      "loss": 0.4076,
+      "step": 490
+    },
+    {
+      "epoch": 1.683673469387755,
+      "grad_norm": 1.2870817184448242,
+      "learning_rate": 2.378397530444238e-05,
+      "loss": 0.4198,
+      "step": 495
+    },
+    {
+      "epoch": 1.7006802721088436,
+      "grad_norm": 1.0723446607589722,
+      "learning_rate": 2.3646607336976375e-05,
+      "loss": 0.4536,
+      "step": 500
+    },
+    {
+      "epoch": 1.717687074829932,
+      "grad_norm": 1.3169423341751099,
+      "learning_rate": 2.3508144631815326e-05,
+      "loss": 0.4213,
+      "step": 505
+    },
+    {
+      "epoch": 1.7346938775510203,
+      "grad_norm": 1.2261208295822144,
+      "learning_rate": 2.3368604719574055e-05,
+      "loss": 0.4414,
+      "step": 510
+    },
+    {
+      "epoch": 1.751700680272109,
+      "grad_norm": 1.2695997953414917,
+      "learning_rate": 2.322800526725141e-05,
+      "loss": 0.3781,
+      "step": 515
+    },
+    {
+      "epoch": 1.7687074829931972,
+      "grad_norm": 1.421828269958496,
+      "learning_rate": 2.308636407599347e-05,
+      "loss": 0.4473,
+      "step": 520
+    },
+    {
+      "epoch": 1.7857142857142856,
+      "grad_norm": 1.2838345766067505,
+      "learning_rate": 2.2943699078839783e-05,
+      "loss": 0.4126,
+      "step": 525
+    },
+    {
+      "epoch": 1.8027210884353742,
+      "grad_norm": 1.290840983390808,
+      "learning_rate": 2.2800028338452853e-05,
+      "loss": 0.4035,
+      "step": 530
+    },
+    {
+      "epoch": 1.8197278911564627,
+      "grad_norm": 1.3847336769104004,
+      "learning_rate": 2.2655370044831253e-05,
+      "loss": 0.378,
+      "step": 535
+    },
+    {
+      "epoch": 1.836734693877551,
+      "grad_norm": 1.149859070777893,
+      "learning_rate": 2.2509742513006633e-05,
+      "loss": 0.3737,
+      "step": 540
+    },
+    {
+      "epoch": 1.8537414965986394,
+      "grad_norm": 1.0945329666137695,
+      "learning_rate": 2.2363164180724828e-05,
+      "loss": 0.4016,
+      "step": 545
+    },
+    {
+      "epoch": 1.870748299319728,
+      "grad_norm": 1.2481576204299927,
+      "learning_rate": 2.2215653606111515e-05,
+      "loss": 0.4054,
+      "step": 550
+    },
+    {
+      "epoch": 1.8877551020408163,
+      "grad_norm": 1.065503478050232,
+      "learning_rate": 2.2067229465322578e-05,
+      "loss": 0.3944,
+      "step": 555
+    },
+    {
+      "epoch": 1.9047619047619047,
+      "grad_norm": 1.2339270114898682,
+      "learning_rate": 2.1917910550179527e-05,
+      "loss": 0.3505,
+      "step": 560
+    },
+    {
+      "epoch": 1.9217687074829932,
+      "grad_norm": 1.1712156534194946,
+      "learning_rate": 2.1767715765790303e-05,
+      "loss": 0.3497,
+      "step": 565
+    },
+    {
+      "epoch": 1.9387755102040818,
+      "grad_norm": 1.4303723573684692,
+      "learning_rate": 2.16166641281557e-05,
+      "loss": 0.3772,
+      "step": 570
+    },
+    {
+      "epoch": 1.95578231292517,
+      "grad_norm": 1.123719573020935,
+      "learning_rate": 2.1464774761761805e-05,
+      "loss": 0.3558,
+      "step": 575
+    },
+    {
+      "epoch": 1.9727891156462585,
+      "grad_norm": 1.1126965284347534,
+      "learning_rate": 2.131206689715863e-05,
+      "loss": 0.3763,
+      "step": 580
+    },
+    {
+      "epoch": 1.989795918367347,
+      "grad_norm": 1.0419843196868896,
+      "learning_rate": 2.1158559868525374e-05,
+      "loss": 0.3507,
+      "step": 585
+    },
+    {
+      "epoch": 2.006802721088435,
+      "grad_norm": 1.15863835811615,
+      "learning_rate": 2.1004273111222532e-05,
+      "loss": 0.3066,
+      "step": 590
+    },
+    {
+      "epoch": 2.0238095238095237,
+      "grad_norm": 1.1892127990722656,
+      "learning_rate": 2.0849226159331222e-05,
+      "loss": 0.2585,
+      "step": 595
+    },
+    {
+      "epoch": 2.0408163265306123,
+      "grad_norm": 1.0521714687347412,
+      "learning_rate": 2.069343864317998e-05,
+      "loss": 0.2707,
+      "step": 600
+    },
+    {
+      "epoch": 2.057823129251701,
+      "grad_norm": 1.334870457649231,
+      "learning_rate": 2.053693028685938e-05,
+      "loss": 0.2645,
+      "step": 605
+    },
+    {
+      "epoch": 2.074829931972789,
+      "grad_norm": 1.16738760471344,
+      "learning_rate": 2.0379720905724818e-05,
+      "loss": 0.2664,
+      "step": 610
+    },
+    {
+      "epoch": 2.0918367346938775,
+      "grad_norm": 1.2727965116500854,
+      "learning_rate": 2.022183040388767e-05,
+      "loss": 0.2688,
+      "step": 615
+    },
+    {
+      "epoch": 2.108843537414966,
+      "grad_norm": 1.0208539962768555,
+      "learning_rate": 2.006327877169529e-05,
+      "loss": 0.2724,
+      "step": 620
+    },
+    {
+      "epoch": 2.1258503401360542,
+      "grad_norm": 1.7100193500518799,
+      "learning_rate": 1.9904086083200035e-05,
+      "loss": 0.2936,
+      "step": 625
+    },
+    {
+      "epoch": 2.142857142857143,
+      "grad_norm": 1.143485426902771,
+      "learning_rate": 1.9744272493617703e-05,
+      "loss": 0.2851,
+      "step": 630
+    },
+    {
+      "epoch": 2.1598639455782314,
+      "grad_norm": 1.1511744260787964,
+      "learning_rate": 1.9583858236775733e-05,
+      "loss": 0.2602,
+      "step": 635
+    },
+    {
+      "epoch": 2.17687074829932,
+      "grad_norm": 1.3126564025878906,
+      "learning_rate": 1.9422863622551376e-05,
+      "loss": 0.2915,
+      "step": 640
+    },
+    {
+      "epoch": 2.193877551020408,
+      "grad_norm": 1.1826177835464478,
+      "learning_rate": 1.926130903430034e-05,
+      "loss": 0.2847,
+      "step": 645
+    },
+    {
+      "epoch": 2.2108843537414966,
+      "grad_norm": 1.19844651222229,
+      "learning_rate": 1.9099214926276024e-05,
+      "loss": 0.3337,
+      "step": 650
+    },
+    {
+      "epoch": 2.227891156462585,
+      "grad_norm": 1.5924816131591797,
+      "learning_rate": 1.8936601821039858e-05,
+      "loss": 0.252,
+      "step": 655
+    },
+    {
+      "epoch": 2.2448979591836733,
+      "grad_norm": 1.3154438734054565,
+      "learning_rate": 1.8773490306862945e-05,
+      "loss": 0.2506,
+      "step": 660
+    },
+    {
+      "epoch": 2.261904761904762,
+      "grad_norm": 1.09384024143219,
+      "learning_rate": 1.860990103511941e-05,
+      "loss": 0.2947,
+      "step": 665
+    },
+    {
+      "epoch": 2.2789115646258504,
+      "grad_norm": 1.2651005983352661,
+      "learning_rate": 1.8445854717671768e-05,
+      "loss": 0.2845,
+      "step": 670
+    },
+    {
+      "epoch": 2.295918367346939,
+      "grad_norm": 1.198618769645691,
+      "learning_rate": 1.828137212424858e-05,
+      "loss": 0.2675,
+      "step": 675
+    },
+    {
+      "epoch": 2.312925170068027,
+      "grad_norm": 1.225216269493103,
+      "learning_rate": 1.8116474079814855e-05,
+      "loss": 0.255,
+      "step": 680
+    },
+    {
+      "epoch": 2.3299319727891157,
+      "grad_norm": 1.2420880794525146,
+      "learning_rate": 1.79511814619354e-05,
+      "loss": 0.2276,
+      "step": 685
+    },
+    {
+      "epoch": 2.3469387755102042,
+      "grad_norm": 1.2797255516052246,
+      "learning_rate": 1.7785515198131556e-05,
+      "loss": 0.2761,
+      "step": 690
+    },
+    {
+      "epoch": 2.3639455782312924,
+      "grad_norm": 1.1373989582061768,
+      "learning_rate": 1.7619496263231557e-05,
+      "loss": 0.2634,
+      "step": 695
+    },
+    {
+      "epoch": 2.380952380952381,
+      "grad_norm": 1.0364551544189453,
+      "learning_rate": 1.745314567671496e-05,
+      "loss": 0.278,
+      "step": 700
+    },
+    {
+      "epoch": 2.3979591836734695,
+      "grad_norm": 1.3783962726593018,
+      "learning_rate": 1.7286484500051383e-05,
+      "loss": 0.2569,
+      "step": 705
+    },
+    {
+      "epoch": 2.4149659863945576,
+      "grad_norm": 1.2774587869644165,
+      "learning_rate": 1.7119533834033907e-05,
+      "loss": 0.2515,
+      "step": 710
+    },
+    {
+      "epoch": 2.431972789115646,
+      "grad_norm": 1.2328591346740723,
+      "learning_rate": 1.6952314816107576e-05,
+      "loss": 0.2344,
+      "step": 715
+    },
+    {
+      "epoch": 2.4489795918367347,
+      "grad_norm": 1.1517112255096436,
+      "learning_rate": 1.678484861769316e-05,
+      "loss": 0.2649,
+      "step": 720
+    },
+    {
+      "epoch": 2.4659863945578233,
+      "grad_norm": 1.1749553680419922,
+      "learning_rate": 1.661715644150671e-05,
+      "loss": 0.2284,
+      "step": 725
+    },
+    {
+      "epoch": 2.4829931972789114,
+      "grad_norm": 1.298263669013977,
+      "learning_rate": 1.644925951887505e-05,
+      "loss": 0.2178,
+      "step": 730
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 1.0253585577011108,
+      "learning_rate": 1.6281179107047765e-05,
+      "loss": 0.2464,
+      "step": 735
+    },
+    {
+      "epoch": 2.5170068027210886,
+      "grad_norm": 1.2435274124145508,
+      "learning_rate": 1.6112936486505785e-05,
+      "loss": 0.2129,
+      "step": 740
+    },
+    {
+      "epoch": 2.534013605442177,
+      "grad_norm": 1.1542493104934692,
+      "learning_rate": 1.5944552958267118e-05,
+      "loss": 0.2152,
+      "step": 745
+    },
+    {
+      "epoch": 2.5510204081632653,
+      "grad_norm": 1.0975438356399536,
+      "learning_rate": 1.5776049841189958e-05,
+      "loss": 0.2312,
+      "step": 750
+    },
+    {
+      "epoch": 2.568027210884354,
+      "grad_norm": 1.0609310865402222,
+      "learning_rate": 1.5607448469273495e-05,
+      "loss": 0.2426,
+      "step": 755
+    },
+    {
+      "epoch": 2.5850340136054424,
+      "grad_norm": 1.144705057144165,
+      "learning_rate": 1.543877018895687e-05,
+      "loss": 0.2047,
+      "step": 760
+    },
+    {
+      "epoch": 2.6020408163265305,
+      "grad_norm": 1.1797125339508057,
+      "learning_rate": 1.5270036356416515e-05,
+      "loss": 0.2153,
+      "step": 765
+    },
+    {
+      "epoch": 2.619047619047619,
+      "grad_norm": 1.1653424501419067,
+      "learning_rate": 1.5101268334862263e-05,
+      "loss": 0.2314,
+      "step": 770
+    },
+    {
+      "epoch": 2.6360544217687076,
+      "grad_norm": 1.0752824544906616,
+      "learning_rate": 1.4932487491832584e-05,
+      "loss": 0.2224,
+      "step": 775
+    },
+    {
+      "epoch": 2.6530612244897958,
+      "grad_norm": 1.1791971921920776,
+      "learning_rate": 1.4763715196489263e-05,
+      "loss": 0.2004,
+      "step": 780
+    },
+    {
+      "epoch": 2.6700680272108843,
+      "grad_norm": 1.1845297813415527,
+      "learning_rate": 1.4594972816911873e-05,
+      "loss": 0.2555,
+      "step": 785
+    },
+    {
+      "epoch": 2.687074829931973,
+      "grad_norm": 1.2175406217575073,
+      "learning_rate": 1.4426281717392377e-05,
+      "loss": 0.235,
+      "step": 790
+    },
+    {
+      "epoch": 2.704081632653061,
+      "grad_norm": 1.1495968103408813,
+      "learning_rate": 1.4257663255730234e-05,
+      "loss": 0.2297,
+      "step": 795
+    },
+    {
+      "epoch": 2.7210884353741496,
+      "grad_norm": 1.2075597047805786,
+      "learning_rate": 1.4089138780528287e-05,
+      "loss": 0.1798,
+      "step": 800
+    },
+    {
+      "epoch": 2.738095238095238,
+      "grad_norm": 1.117581844329834,
+      "learning_rate": 1.392072962848988e-05,
+      "loss": 0.2476,
+      "step": 805
+    },
+    {
+      "epoch": 2.7551020408163263,
+      "grad_norm": 1.2183245420455933,
+      "learning_rate": 1.3752457121717383e-05,
+      "loss": 0.2281,
+      "step": 810
+    },
+    {
+      "epoch": 2.772108843537415,
+      "grad_norm": 1.198248267173767,
+      "learning_rate": 1.3584342565012677e-05,
+      "loss": 0.2031,
+      "step": 815
+    },
+    {
+      "epoch": 2.7891156462585034,
+      "grad_norm": 1.257199764251709,
+      "learning_rate": 1.341640724317975e-05,
+      "loss": 0.2295,
+      "step": 820
+    },
+    {
+      "epoch": 2.806122448979592,
+      "grad_norm": 1.1676485538482666,
+      "learning_rate": 1.324867241832985e-05,
+      "loss": 0.1996,
+      "step": 825
+    },
+    {
+      "epoch": 2.8231292517006805,
+      "grad_norm": 0.9781966209411621,
+      "learning_rate": 1.3081159327189524e-05,
+      "loss": 0.2001,
+      "step": 830
+    },
+    {
+      "epoch": 2.8401360544217686,
+      "grad_norm": 1.2090445756912231,
+      "learning_rate": 1.2913889178411837e-05,
+      "loss": 0.2193,
+      "step": 835
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 1.130637764930725,
+      "learning_rate": 1.2746883149891203e-05,
+      "loss": 0.1832,
+      "step": 840
+    },
+    {
+      "epoch": 2.8741496598639458,
+      "grad_norm": 1.2970364093780518,
+      "learning_rate": 1.2580162386082028e-05,
+      "loss": 0.2086,
+      "step": 845
+    },
+    {
+      "epoch": 2.891156462585034,
+      "grad_norm": 1.0856074094772339,
+      "learning_rate": 1.2413747995321692e-05,
+      "loss": 0.1941,
+      "step": 850
+    },
+    {
+      "epoch": 2.9081632653061225,
+      "grad_norm": 1.1722519397735596,
+      "learning_rate": 1.2247661047157968e-05,
+      "loss": 0.1959,
+      "step": 855
+    },
+    {
+      "epoch": 2.925170068027211,
+      "grad_norm": 1.1799602508544922,
+      "learning_rate": 1.208192256968151e-05,
+      "loss": 0.1968,
+      "step": 860
+    },
+    {
+      "epoch": 2.942176870748299,
+      "grad_norm": 1.1089502573013306,
+      "learning_rate": 1.191655354686346e-05,
+      "loss": 0.2048,
+      "step": 865
+    },
+    {
+      "epoch": 2.9591836734693877,
+      "grad_norm": 1.1572333574295044,
+      "learning_rate": 1.175157491589869e-05,
+      "loss": 0.2141,
+      "step": 870
+    },
+    {
+      "epoch": 2.9761904761904763,
+      "grad_norm": 1.2635926008224487,
+      "learning_rate": 1.1587007564554991e-05,
+      "loss": 0.1973,
+      "step": 875
+    },
+    {
+      "epoch": 2.9931972789115644,
+      "grad_norm": 1.1943570375442505,
+      "learning_rate": 1.1422872328528473e-05,
+      "loss": 0.2053,
+      "step": 880
+    },
+    {
+      "epoch": 3.010204081632653,
+      "grad_norm": 1.0038658380508423,
+      "learning_rate": 1.1259189988805599e-05,
+      "loss": 0.1666,
+      "step": 885
+    },
+    {
+      "epoch": 3.0272108843537415,
+      "grad_norm": 1.3621819019317627,
+      "learning_rate": 1.1095981269032097e-05,
+      "loss": 0.1502,
+      "step": 890
+    },
+    {
+      "epoch": 3.04421768707483,
+      "grad_norm": 1.0059318542480469,
+      "learning_rate": 1.0933266832889206e-05,
+      "loss": 0.1665,
+      "step": 895
+    },
+    {
+      "epoch": 3.061224489795918,
+      "grad_norm": 1.1601946353912354,
+      "learning_rate": 1.077106728147743e-05,
+      "loss": 0.1801,
+      "step": 900
+    },
+    {
+      "epoch": 3.078231292517007,
+      "grad_norm": 1.3520936965942383,
+      "learning_rate": 1.0609403150708261e-05,
+      "loss": 0.1305,
+      "step": 905
+    },
+    {
+      "epoch": 3.0952380952380953,
+      "grad_norm": 1.0658321380615234,
+      "learning_rate": 1.0448294908704173e-05,
+      "loss": 0.127,
+      "step": 910
+    },
+    {
+      "epoch": 3.1122448979591835,
+      "grad_norm": 1.2596958875656128,
+      "learning_rate": 1.028776295320714e-05,
+      "loss": 0.1592,
+      "step": 915
+    },
+    {
+      "epoch": 3.129251700680272,
+      "grad_norm": 0.9525283575057983,
+      "learning_rate": 1.0127827608996144e-05,
+      "loss": 0.1419,
+      "step": 920
+    },
+    {
+      "epoch": 3.1462585034013606,
+      "grad_norm": 1.0077766180038452,
+      "learning_rate": 9.968509125313823e-06,
+      "loss": 0.1314,
+      "step": 925
+    },
+    {
+      "epoch": 3.163265306122449,
+      "grad_norm": 1.0083332061767578,
+      "learning_rate": 9.80982767330278e-06,
+      "loss": 0.1479,
+      "step": 930
+    },
+    {
+      "epoch": 3.1802721088435373,
+      "grad_norm": 0.9684851169586182,
+      "learning_rate": 9.651803343451729e-06,
+      "loss": 0.1385,
+      "step": 935
+    },
+    {
+      "epoch": 3.197278911564626,
+      "grad_norm": 1.190239429473877,
+      "learning_rate": 9.494456143051827e-06,
+      "loss": 0.1361,
+      "step": 940
+    },
+    {
+      "epoch": 3.2142857142857144,
+      "grad_norm": 0.989656925201416,
+      "learning_rate": 9.337805993663618e-06,
+      "loss": 0.1486,
+      "step": 945
+    },
+    {
+      "epoch": 3.2312925170068025,
+      "grad_norm": 1.1670643091201782,
+      "learning_rate": 9.181872728594747e-06,
+      "loss": 0.1476,
+      "step": 950
+    },
+    {
+      "epoch": 3.248299319727891,
+      "grad_norm": 1.0360350608825684,
+      "learning_rate": 9.02667609038892e-06,
+      "loss": 0.1355,
+      "step": 955
+    },
+    {
+      "epoch": 3.2653061224489797,
+      "grad_norm": 1.2193771600723267,
+      "learning_rate": 8.872235728326284e-06,
+      "loss": 0.1634,
+      "step": 960
+    },
+    {
+      "epoch": 3.282312925170068,
+      "grad_norm": 0.9404177665710449,
+      "learning_rate": 8.718571195935696e-06,
+      "loss": 0.144,
+      "step": 965
+    },
+    {
+      "epoch": 3.2993197278911564,
+      "grad_norm": 0.910885751247406,
+      "learning_rate": 8.565701948519034e-06,
+      "loss": 0.1308,
+      "step": 970
+    },
+    {
+      "epoch": 3.316326530612245,
+      "grad_norm": 1.0395634174346924,
+      "learning_rate": 8.413647340688e-06,
+      "loss": 0.1292,
+      "step": 975
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 0.9528372287750244,
+      "learning_rate": 8.262426623913663e-06,
+      "loss": 0.1329,
+      "step": 980
+    },
+    {
+      "epoch": 3.3503401360544216,
+      "grad_norm": 1.2525593042373657,
+      "learning_rate": 8.112058944089003e-06,
+      "loss": 0.1458,
+      "step": 985
+    },
+    {
+      "epoch": 3.36734693877551,
+      "grad_norm": 1.2458698749542236,
+      "learning_rate": 7.96256333910496e-06,
+      "loss": 0.1381,
+      "step": 990
+    },
+    {
+      "epoch": 3.3843537414965987,
+      "grad_norm": 1.031166434288025,
+      "learning_rate": 7.81395873643996e-06,
+      "loss": 0.1378,
+      "step": 995
+    },
+    {
+      "epoch": 3.4013605442176873,
+      "grad_norm": 1.1444342136383057,
+      "learning_rate": 7.666263950763607e-06,
+      "loss": 0.1256,
+      "step": 1000
+    },
+    {
+      "epoch": 3.4183673469387754,
+      "grad_norm": 1.0373079776763916,
+      "learning_rate": 7.519497681554551e-06,
+      "loss": 0.1287,
+      "step": 1005
+    },
+    {
+      "epoch": 3.435374149659864,
+      "grad_norm": 1.1225919723510742,
+      "learning_rate": 7.373678510732955e-06,
+      "loss": 0.1363,
+      "step": 1010
+    },
+    {
+      "epoch": 3.4523809523809526,
+      "grad_norm": 1.0068609714508057,
+      "learning_rate": 7.228824900307881e-06,
+      "loss": 0.1227,
+      "step": 1015
+    },
+    {
+      "epoch": 3.4693877551020407,
+      "grad_norm": 1.1868841648101807,
+      "learning_rate": 7.0849551900398065e-06,
+      "loss": 0.1162,
+      "step": 1020
+    },
+    {
+      "epoch": 3.4863945578231292,
+      "grad_norm": 0.9918032288551331,
+      "learning_rate": 6.94208759511868e-06,
+      "loss": 0.1289,
+      "step": 1025
+    },
+    {
+      "epoch": 3.503401360544218,
+      "grad_norm": 1.0155752897262573,
+      "learning_rate": 6.800240203857696e-06,
+      "loss": 0.138,
+      "step": 1030
+    },
+    {
+      "epoch": 3.520408163265306,
+      "grad_norm": 0.9242071509361267,
+      "learning_rate": 6.659430975403156e-06,
+      "loss": 0.1375,
+      "step": 1035
+    },
+    {
+      "epoch": 3.5374149659863945,
+      "grad_norm": 1.0767987966537476,
+      "learning_rate": 6.519677737460664e-06,
+      "loss": 0.1146,
+      "step": 1040
+    },
+    {
+      "epoch": 3.554421768707483,
+      "grad_norm": 0.9681980013847351,
+      "learning_rate": 6.380998184038025e-06,
+      "loss": 0.1365,
+      "step": 1045
+    },
+    {
+      "epoch": 3.571428571428571,
+      "grad_norm": 0.9982250332832336,
+      "learning_rate": 6.243409873204984e-06,
+      "loss": 0.1307,
+      "step": 1050
+    },
+    {
+      "epoch": 3.5884353741496597,
+      "grad_norm": 0.9535261392593384,
+      "learning_rate": 6.106930224870213e-06,
+      "loss": 0.1305,
+      "step": 1055
+    },
+    {
+      "epoch": 3.6054421768707483,
+      "grad_norm": 0.9636424779891968,
+      "learning_rate": 5.9715765185758395e-06,
+      "loss": 0.1205,
+      "step": 1060
+    },
+    {
+      "epoch": 3.622448979591837,
+      "grad_norm": 1.060290813446045,
+      "learning_rate": 5.8373658913096586e-06,
+      "loss": 0.1068,
+      "step": 1065
+    },
+    {
+      "epoch": 3.6394557823129254,
+      "grad_norm": 0.9654757380485535,
+      "learning_rate": 5.704315335335468e-06,
+      "loss": 0.142,
+      "step": 1070
+    },
+    {
+      "epoch": 3.6564625850340136,
+      "grad_norm": 0.9803591966629028,
+      "learning_rate": 5.572441696041653e-06,
+      "loss": 0.121,
+      "step": 1075
+    },
+    {
+      "epoch": 3.673469387755102,
+      "grad_norm": 0.9896225333213806,
+      "learning_rate": 5.4417616698084785e-06,
+      "loss": 0.1346,
+      "step": 1080
+    },
+    {
+      "epoch": 3.6904761904761907,
+      "grad_norm": 0.9817864298820496,
+      "learning_rate": 5.3122918018941044e-06,
+      "loss": 0.1276,
+      "step": 1085
+    },
+    {
+      "epoch": 3.707482993197279,
+      "grad_norm": 0.9709451794624329,
+      "learning_rate": 5.184048484339855e-06,
+      "loss": 0.1121,
+      "step": 1090
+    },
+    {
+      "epoch": 3.7244897959183674,
+      "grad_norm": 0.9238762259483337,
+      "learning_rate": 5.057047953894831e-06,
+      "loss": 0.1198,
+      "step": 1095
+    },
+    {
+      "epoch": 3.741496598639456,
+      "grad_norm": 0.9318645000457764,
+      "learning_rate": 4.931306289960181e-06,
+      "loss": 0.1254,
+      "step": 1100
+    },
+    {
+      "epoch": 3.758503401360544,
+      "grad_norm": 0.9292995929718018,
+      "learning_rate": 4.806839412553321e-06,
+      "loss": 0.1129,
+      "step": 1105
+    },
+    {
+      "epoch": 3.7755102040816326,
+      "grad_norm": 0.9565867185592651,
+      "learning_rate": 4.6836630802922955e-06,
+      "loss": 0.1088,
+      "step": 1110
+    },
+    {
+      "epoch": 3.792517006802721,
+      "grad_norm": 1.0380818843841553,
+      "learning_rate": 4.5617928884006234e-06,
+      "loss": 0.1002,
+      "step": 1115
+    },
+    {
+      "epoch": 3.8095238095238093,
+      "grad_norm": 1.0153783559799194,
+      "learning_rate": 4.441244266732786e-06,
+      "loss": 0.1002,
+      "step": 1120
+    },
+    {
+      "epoch": 3.826530612244898,
+      "grad_norm": 0.9576224088668823,
+      "learning_rate": 4.3220324778206735e-06,
+      "loss": 0.1272,
+      "step": 1125
+    },
+    {
+      "epoch": 3.8435374149659864,
+      "grad_norm": 0.9821186661720276,
+      "learning_rate": 4.204172614941214e-06,
+      "loss": 0.1186,
+      "step": 1130
+    },
+    {
+      "epoch": 3.8605442176870746,
+      "grad_norm": 0.9095916152000427,
+      "learning_rate": 4.087679600205425e-06,
+      "loss": 0.1312,
+      "step": 1135
+    },
+    {
+      "epoch": 3.877551020408163,
+      "grad_norm": 0.8640242218971252,
+      "learning_rate": 3.9725681826691525e-06,
+      "loss": 0.1124,
+      "step": 1140
+    },
+    {
+      "epoch": 3.8945578231292517,
+      "grad_norm": 0.9114555716514587,
+      "learning_rate": 3.858852936465688e-06,
+      "loss": 0.1027,
+      "step": 1145
+    },
+    {
+      "epoch": 3.9115646258503403,
+      "grad_norm": 1.0309470891952515,
+      "learning_rate": 3.7465482589605713e-06,
+      "loss": 0.1207,
+      "step": 1150
+    },
+    {
+      "epoch": 3.928571428571429,
+      "grad_norm": 0.7389628887176514,
+      "learning_rate": 3.6356683689287566e-06,
+      "loss": 0.1037,
+      "step": 1155
+    },
+    {
+      "epoch": 3.945578231292517,
+      "grad_norm": 0.9068615436553955,
+      "learning_rate": 3.5262273047543787e-06,
+      "loss": 0.1144,
+      "step": 1160
+    },
+    {
+      "epoch": 3.9625850340136055,
+      "grad_norm": 0.9940776824951172,
+      "learning_rate": 3.418238922653357e-06,
+      "loss": 0.1152,
+      "step": 1165
+    },
+    {
+      "epoch": 3.979591836734694,
+      "grad_norm": 0.9465267062187195,
+      "learning_rate": 3.3117168949191134e-06,
+      "loss": 0.1172,
+      "step": 1170
+    },
+    {
+      "epoch": 3.996598639455782,
+      "grad_norm": 0.8871757388114929,
+      "learning_rate": 3.206674708191502e-06,
+      "loss": 0.1064,
+      "step": 1175
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1470,
+  "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": 1.718430489737429e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

48_128_e5_3e-5/checkpoint-1176/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1c73a3d3ae42383cf0a6ea1e73607dc0924ec4b32199fdb2961f1af8855e261
+size 7736

48_128_e5_3e-5/checkpoint-1176/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)

48_128_e5_3e-5/checkpoint-1470/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

48_128_e5_3e-5/checkpoint-1470/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": [
+    "o_proj",
+    "down_proj",
+    "v_proj",
+    "gate_proj",
+    "k_proj",
+    "q_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

48_128_e5_3e-5/checkpoint-1470/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:796b8cc6e6c62062c9b671dfb2cd54702be657b2b1c8cfe115f6527017bf46ef
+size 791751704

48_128_e5_3e-5/checkpoint-1470/latest

@@ -0,0 +1 @@
+global_step1470

48_128_e5_3e-5/checkpoint-1470/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c05e28b48fcaebf1ae9721f67453b549ed8608d73a5e856910168365004c41a3
+size 15920

48_128_e5_3e-5/checkpoint-1470/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e429c943f05a69b02dff731ab7df99f41d5bf6c289545893fe7eff4fb44c9cce
+size 15920

48_128_e5_3e-5/checkpoint-1470/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1cbfc8386d1acf1cde13f4af18c591cb8cf526b491c9a56591a48d1445122ff
+size 15920

48_128_e5_3e-5/checkpoint-1470/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c25dad44d46d5e9c5d52ffb2a6e1d10ddb19ec4ed82f90db2705d7fe6b9f3a9
+size 15920

48_128_e5_3e-5/checkpoint-1470/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea2b13e3c9665525141ebb6a924dd4ff50666037fccc9a544a43b13f905a5ce1
+size 15920

48_128_e5_3e-5/checkpoint-1470/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c48795b6fefc092e9b4c3d0d3d519cbd7b93037f6bd7218974ecd6023fd064b
+size 15920

48_128_e5_3e-5/checkpoint-1470/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a087ec440e71a2bfd2b349da08f2f6655bad7355230af7944a18509d200b2e57
+size 15920

48_128_e5_3e-5/checkpoint-1470/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:497c80aa54510f304ddb0847abc6727b50b8fd1f7903de37c37d712ada6b60d5
+size 15920

48_128_e5_3e-5/checkpoint-1470/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de44feea6e568ad59993ff201f9e2451e1dd56c35ed5c4153b16684ccf64a626
+size 1064

48_128_e5_3e-5/checkpoint-1470/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
+  }
+}

48_128_e5_3e-5/checkpoint-1470/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
+}

48_128_e5_3e-5/checkpoint-1470/trainer_state.json

@@ -0,0 +1,2092 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 1470,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.017006802721088437,
+      "grad_norm": 1.1845334768295288,
+      "learning_rate": 1.6216216216216219e-06,
+      "loss": 1.3293,
+      "step": 5
+    },
+    {
+      "epoch": 0.034013605442176874,
+      "grad_norm": 1.0181540250778198,
+      "learning_rate": 3.648648648648649e-06,
+      "loss": 1.2783,
+      "step": 10
+    },
+    {
+      "epoch": 0.05102040816326531,
+      "grad_norm": 0.6886288523674011,
+      "learning_rate": 5.675675675675676e-06,
+      "loss": 1.2906,
+      "step": 15
+    },
+    {
+      "epoch": 0.06802721088435375,
+      "grad_norm": 0.6959572434425354,
+      "learning_rate": 7.702702702702703e-06,
+      "loss": 1.2818,
+      "step": 20
+    },
+    {
+      "epoch": 0.08503401360544217,
+      "grad_norm": 0.6249067187309265,
+      "learning_rate": 9.72972972972973e-06,
+      "loss": 1.2667,
+      "step": 25
+    },
+    {
+      "epoch": 0.10204081632653061,
+      "grad_norm": 0.5868192911148071,
+      "learning_rate": 1.1756756756756757e-05,
+      "loss": 1.2477,
+      "step": 30
+    },
+    {
+      "epoch": 0.11904761904761904,
+      "grad_norm": 0.5181043148040771,
+      "learning_rate": 1.3783783783783784e-05,
+      "loss": 1.2207,
+      "step": 35
+    },
+    {
+      "epoch": 0.1360544217687075,
+      "grad_norm": 0.4317700266838074,
+      "learning_rate": 1.5810810810810808e-05,
+      "loss": 1.21,
+      "step": 40
+    },
+    {
+      "epoch": 0.15306122448979592,
+      "grad_norm": 0.4549165964126587,
+      "learning_rate": 1.783783783783784e-05,
+      "loss": 1.1368,
+      "step": 45
+    },
+    {
+      "epoch": 0.17006802721088435,
+      "grad_norm": 0.5308559536933899,
+      "learning_rate": 1.9864864864864866e-05,
+      "loss": 1.1875,
+      "step": 50
+    },
+    {
+      "epoch": 0.1870748299319728,
+      "grad_norm": 0.5514539480209351,
+      "learning_rate": 2.1891891891891892e-05,
+      "loss": 1.1593,
+      "step": 55
+    },
+    {
+      "epoch": 0.20408163265306123,
+      "grad_norm": 0.46228182315826416,
+      "learning_rate": 2.3918918918918917e-05,
+      "loss": 1.1345,
+      "step": 60
+    },
+    {
+      "epoch": 0.22108843537414966,
+      "grad_norm": 0.5649062395095825,
+      "learning_rate": 2.594594594594595e-05,
+      "loss": 1.1065,
+      "step": 65
+    },
+    {
+      "epoch": 0.23809523809523808,
+      "grad_norm": 0.49446532130241394,
+      "learning_rate": 2.7972972972972975e-05,
+      "loss": 1.1328,
+      "step": 70
+    },
+    {
+      "epoch": 0.25510204081632654,
+      "grad_norm": 0.5560067296028137,
+      "learning_rate": 3e-05,
+      "loss": 1.1697,
+      "step": 75
+    },
+    {
+      "epoch": 0.272108843537415,
+      "grad_norm": 0.4814499616622925,
+      "learning_rate": 2.999905043303196e-05,
+      "loss": 1.1659,
+      "step": 80
+    },
+    {
+      "epoch": 0.2891156462585034,
+      "grad_norm": 0.5074186325073242,
+      "learning_rate": 2.999620185235149e-05,
+      "loss": 1.0924,
+      "step": 85
+    },
+    {
+      "epoch": 0.30612244897959184,
+      "grad_norm": 0.592327892780304,
+      "learning_rate": 2.9991454618614338e-05,
+      "loss": 1.0793,
+      "step": 90
+    },
+    {
+      "epoch": 0.3231292517006803,
+      "grad_norm": 0.5456128120422363,
+      "learning_rate": 2.998480933286269e-05,
+      "loss": 1.0836,
+      "step": 95
+    },
+    {
+      "epoch": 0.3401360544217687,
+      "grad_norm": 0.6163952946662903,
+      "learning_rate": 2.9976266836449057e-05,
+      "loss": 1.0816,
+      "step": 100
+    },
+    {
+      "epoch": 0.35714285714285715,
+      "grad_norm": 0.5601843595504761,
+      "learning_rate": 2.9965828210929758e-05,
+      "loss": 1.091,
+      "step": 105
+    },
+    {
+      "epoch": 0.3741496598639456,
+      "grad_norm": 0.5945920944213867,
+      "learning_rate": 2.9953494777927995e-05,
+      "loss": 1.0349,
+      "step": 110
+    },
+    {
+      "epoch": 0.391156462585034,
+      "grad_norm": 0.5780587792396545,
+      "learning_rate": 2.993926809896651e-05,
+      "loss": 1.0389,
+      "step": 115
+    },
+    {
+      "epoch": 0.40816326530612246,
+      "grad_norm": 0.6240988969802856,
+      "learning_rate": 2.9923149975269885e-05,
+      "loss": 1.0521,
+      "step": 120
+    },
+    {
+      "epoch": 0.42517006802721086,
+      "grad_norm": 0.6076016426086426,
+      "learning_rate": 2.990514244753651e-05,
+      "loss": 0.9944,
+      "step": 125
+    },
+    {
+      "epoch": 0.4421768707482993,
+      "grad_norm": 0.6623475551605225,
+      "learning_rate": 2.988524779568018e-05,
+      "loss": 0.935,
+      "step": 130
+    },
+    {
+      "epoch": 0.45918367346938777,
+      "grad_norm": 0.6462083458900452,
+      "learning_rate": 2.9863468538541466e-05,
+      "loss": 1.0017,
+      "step": 135
+    },
+    {
+      "epoch": 0.47619047619047616,
+      "grad_norm": 0.6660473346710205,
+      "learning_rate": 2.9839807433568787e-05,
+      "loss": 0.8931,
+      "step": 140
+    },
+    {
+      "epoch": 0.4931972789115646,
+      "grad_norm": 0.680072546005249,
+      "learning_rate": 2.9814267476469304e-05,
+      "loss": 0.9786,
+      "step": 145
+    },
+    {
+      "epoch": 0.5102040816326531,
+      "grad_norm": 0.6932852268218994,
+      "learning_rate": 2.9786851900829633e-05,
+      "loss": 0.9335,
+      "step": 150
+    },
+    {
+      "epoch": 0.5272108843537415,
+      "grad_norm": 0.7095908522605896,
+      "learning_rate": 2.9757564177706448e-05,
+      "loss": 0.9278,
+      "step": 155
+    },
+    {
+      "epoch": 0.54421768707483,
+      "grad_norm": 0.7165976762771606,
+      "learning_rate": 2.972640801518701e-05,
+      "loss": 0.9029,
+      "step": 160
+    },
+    {
+      "epoch": 0.5612244897959183,
+      "grad_norm": 0.745604395866394,
+      "learning_rate": 2.969338735791968e-05,
+      "loss": 0.8675,
+      "step": 165
+    },
+    {
+      "epoch": 0.5782312925170068,
+      "grad_norm": 0.6811717748641968,
+      "learning_rate": 2.9658506386614525e-05,
+      "loss": 0.9125,
+      "step": 170
+    },
+    {
+      "epoch": 0.5952380952380952,
+      "grad_norm": 0.7211567163467407,
+      "learning_rate": 2.962176951751396e-05,
+      "loss": 0.8937,
+      "step": 175
+    },
+    {
+      "epoch": 0.6122448979591837,
+      "grad_norm": 0.7217546701431274,
+      "learning_rate": 2.958318140183364e-05,
+      "loss": 0.9474,
+      "step": 180
+    },
+    {
+      "epoch": 0.6292517006802721,
+      "grad_norm": 0.7711513042449951,
+      "learning_rate": 2.9542746925173566e-05,
+      "loss": 0.882,
+      "step": 185
+    },
+    {
+      "epoch": 0.6462585034013606,
+      "grad_norm": 0.8753324747085571,
+      "learning_rate": 2.9500471206899528e-05,
+      "loss": 0.874,
+      "step": 190
+    },
+    {
+      "epoch": 0.6632653061224489,
+      "grad_norm": 0.873315691947937,
+      "learning_rate": 2.945635959949494e-05,
+      "loss": 0.8252,
+      "step": 195
+    },
+    {
+      "epoch": 0.6802721088435374,
+      "grad_norm": 1.0285780429840088,
+      "learning_rate": 2.9410417687883173e-05,
+      "loss": 0.885,
+      "step": 200
+    },
+    {
+      "epoch": 0.6972789115646258,
+      "grad_norm": 0.8351210951805115,
+      "learning_rate": 2.936265128872046e-05,
+      "loss": 0.8756,
+      "step": 205
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 0.8613854646682739,
+      "learning_rate": 2.931306644965944e-05,
+      "loss": 0.8489,
+      "step": 210
+    },
+    {
+      "epoch": 0.7312925170068028,
+      "grad_norm": 0.9348616003990173,
+      "learning_rate": 2.9261669448583492e-05,
+      "loss": 0.8657,
+      "step": 215
+    },
+    {
+      "epoch": 0.7482993197278912,
+      "grad_norm": 1.0146417617797852,
+      "learning_rate": 2.9208466792811875e-05,
+      "loss": 0.815,
+      "step": 220
+    },
+    {
+      "epoch": 0.7653061224489796,
+      "grad_norm": 0.9285673499107361,
+      "learning_rate": 2.915346521827586e-05,
+      "loss": 0.8164,
+      "step": 225
+    },
+    {
+      "epoch": 0.782312925170068,
+      "grad_norm": 0.951054573059082,
+      "learning_rate": 2.9096671688665893e-05,
+      "loss": 0.8515,
+      "step": 230
+    },
+    {
+      "epoch": 0.7993197278911565,
+      "grad_norm": 0.9739980697631836,
+      "learning_rate": 2.9038093394549946e-05,
+      "loss": 0.7521,
+      "step": 235
+    },
+    {
+      "epoch": 0.8163265306122449,
+      "grad_norm": 1.0542875528335571,
+      "learning_rate": 2.8977737752463094e-05,
+      "loss": 0.8105,
+      "step": 240
+    },
+    {
+      "epoch": 0.8333333333333334,
+      "grad_norm": 0.9150990843772888,
+      "learning_rate": 2.891561240396855e-05,
+      "loss": 0.8032,
+      "step": 245
+    },
+    {
+      "epoch": 0.8503401360544217,
+      "grad_norm": 1.0160540342330933,
+      "learning_rate": 2.8851725214690155e-05,
+      "loss": 0.7096,
+      "step": 250
+    },
+    {
+      "epoch": 0.8673469387755102,
+      "grad_norm": 0.9452092051506042,
+      "learning_rate": 2.8786084273316524e-05,
+      "loss": 0.6958,
+      "step": 255
+    },
+    {
+      "epoch": 0.8843537414965986,
+      "grad_norm": 0.8861579895019531,
+      "learning_rate": 2.8718697890576944e-05,
+      "loss": 0.7085,
+      "step": 260
+    },
+    {
+      "epoch": 0.9013605442176871,
+      "grad_norm": 1.0629994869232178,
+      "learning_rate": 2.864957459818918e-05,
+      "loss": 0.6941,
+      "step": 265
+    },
+    {
+      "epoch": 0.9183673469387755,
+      "grad_norm": 0.9347429275512695,
+      "learning_rate": 2.8578723147779237e-05,
+      "loss": 0.689,
+      "step": 270
+    },
+    {
+      "epoch": 0.935374149659864,
+      "grad_norm": 1.0432939529418945,
+      "learning_rate": 2.850615250977339e-05,
+      "loss": 0.7042,
+      "step": 275
+    },
+    {
+      "epoch": 0.9523809523809523,
+      "grad_norm": 0.948538601398468,
+      "learning_rate": 2.843187187226239e-05,
+      "loss": 0.7113,
+      "step": 280
+    },
+    {
+      "epoch": 0.9693877551020408,
+      "grad_norm": 1.053505778312683,
+      "learning_rate": 2.835589063983821e-05,
+      "loss": 0.7173,
+      "step": 285
+    },
+    {
+      "epoch": 0.9863945578231292,
+      "grad_norm": 0.9371920824050903,
+      "learning_rate": 2.827821843240331e-05,
+      "loss": 0.6272,
+      "step": 290
+    },
+    {
+      "epoch": 1.0034013605442176,
+      "grad_norm": 1.1553608179092407,
+      "learning_rate": 2.8198865083952694e-05,
+      "loss": 0.6671,
+      "step": 295
+    },
+    {
+      "epoch": 1.0204081632653061,
+      "grad_norm": 1.0597134828567505,
+      "learning_rate": 2.811784064132883e-05,
+      "loss": 0.6007,
+      "step": 300
+    },
+    {
+      "epoch": 1.0374149659863945,
+      "grad_norm": 1.1430370807647705,
+      "learning_rate": 2.803515536294963e-05,
+      "loss": 0.6015,
+      "step": 305
+    },
+    {
+      "epoch": 1.054421768707483,
+      "grad_norm": 1.0705440044403076,
+      "learning_rate": 2.795081971750963e-05,
+      "loss": 0.5987,
+      "step": 310
+    },
+    {
+      "epoch": 1.0714285714285714,
+      "grad_norm": 1.1100271940231323,
+      "learning_rate": 2.786484438265459e-05,
+      "loss": 0.6058,
+      "step": 315
+    },
+    {
+      "epoch": 1.08843537414966,
+      "grad_norm": 1.1145716905593872,
+      "learning_rate": 2.7777240243629578e-05,
+      "loss": 0.5264,
+      "step": 320
+    },
+    {
+      "epoch": 1.1054421768707483,
+      "grad_norm": 1.0852850675582886,
+      "learning_rate": 2.7688018391900826e-05,
+      "loss": 0.549,
+      "step": 325
+    },
+    {
+      "epoch": 1.1224489795918366,
+      "grad_norm": 1.1767306327819824,
+      "learning_rate": 2.7597190123751422e-05,
+      "loss": 0.5997,
+      "step": 330
+    },
+    {
+      "epoch": 1.1394557823129252,
+      "grad_norm": 1.2128592729568481,
+      "learning_rate": 2.750476693885113e-05,
+      "loss": 0.5681,
+      "step": 335
+    },
+    {
+      "epoch": 1.1564625850340136,
+      "grad_norm": 1.313896894454956,
+      "learning_rate": 2.7410760538800408e-05,
+      "loss": 0.5698,
+      "step": 340
+    },
+    {
+      "epoch": 1.1734693877551021,
+      "grad_norm": 1.1703464984893799,
+      "learning_rate": 2.7315182825648895e-05,
+      "loss": 0.569,
+      "step": 345
+    },
+    {
+      "epoch": 1.1904761904761905,
+      "grad_norm": 1.1567600965499878,
+      "learning_rate": 2.7218045900388504e-05,
+      "loss": 0.5139,
+      "step": 350
+    },
+    {
+      "epoch": 1.2074829931972788,
+      "grad_norm": 1.097703456878662,
+      "learning_rate": 2.7119362061421303e-05,
+      "loss": 0.5093,
+      "step": 355
+    },
+    {
+      "epoch": 1.2244897959183674,
+      "grad_norm": 1.0932475328445435,
+      "learning_rate": 2.7019143803002465e-05,
+      "loss": 0.5731,
+      "step": 360
+    },
+    {
+      "epoch": 1.2414965986394557,
+      "grad_norm": 1.2174867391586304,
+      "learning_rate": 2.6917403813658364e-05,
+      "loss": 0.5077,
+      "step": 365
+    },
+    {
+      "epoch": 1.2585034013605443,
+      "grad_norm": 1.0711535215377808,
+      "learning_rate": 2.6814154974580092e-05,
+      "loss": 0.5205,
+      "step": 370
+    },
+    {
+      "epoch": 1.2755102040816326,
+      "grad_norm": 1.1222914457321167,
+      "learning_rate": 2.67094103579926e-05,
+      "loss": 0.5386,
+      "step": 375
+    },
+    {
+      "epoch": 1.2925170068027212,
+      "grad_norm": 1.1327167749404907,
+      "learning_rate": 2.6603183225499608e-05,
+      "loss": 0.5446,
+      "step": 380
+    },
+    {
+      "epoch": 1.3095238095238095,
+      "grad_norm": 1.2905243635177612,
+      "learning_rate": 2.6495487026404607e-05,
+      "loss": 0.5009,
+      "step": 385
+    },
+    {
+      "epoch": 1.3265306122448979,
+      "grad_norm": 1.059859037399292,
+      "learning_rate": 2.6386335396008033e-05,
+      "loss": 0.4772,
+      "step": 390
+    },
+    {
+      "epoch": 1.3435374149659864,
+      "grad_norm": 1.329107403755188,
+      "learning_rate": 2.6275742153880907e-05,
+      "loss": 0.5027,
+      "step": 395
+    },
+    {
+      "epoch": 1.3605442176870748,
+      "grad_norm": 1.1252979040145874,
+      "learning_rate": 2.6163721302115184e-05,
+      "loss": 0.4714,
+      "step": 400
+    },
+    {
+      "epoch": 1.3775510204081631,
+      "grad_norm": 1.1349155902862549,
+      "learning_rate": 2.6050287023550936e-05,
+      "loss": 0.4851,
+      "step": 405
+    },
+    {
+      "epoch": 1.3945578231292517,
+      "grad_norm": 1.0614500045776367,
+      "learning_rate": 2.59354536799807e-05,
+      "loss": 0.4893,
+      "step": 410
+    },
+    {
+      "epoch": 1.4115646258503403,
+      "grad_norm": 1.2005356550216675,
+      "learning_rate": 2.5819235810331115e-05,
+      "loss": 0.5233,
+      "step": 415
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 1.0654675960540771,
+      "learning_rate": 2.5701648128822205e-05,
+      "loss": 0.4876,
+      "step": 420
+    },
+    {
+      "epoch": 1.445578231292517,
+      "grad_norm": 1.3152744770050049,
+      "learning_rate": 2.55827055231044e-05,
+      "loss": 0.5113,
+      "step": 425
+    },
+    {
+      "epoch": 1.4625850340136055,
+      "grad_norm": 1.0852030515670776,
+      "learning_rate": 2.5462423052373628e-05,
+      "loss": 0.4794,
+      "step": 430
+    },
+    {
+      "epoch": 1.4795918367346939,
+      "grad_norm": 1.1766146421432495,
+      "learning_rate": 2.534081594546469e-05,
+      "loss": 0.4488,
+      "step": 435
+    },
+    {
+      "epoch": 1.4965986394557822,
+      "grad_norm": 1.2981992959976196,
+      "learning_rate": 2.5217899598923162e-05,
+      "loss": 0.4567,
+      "step": 440
+    },
+    {
+      "epoch": 1.5136054421768708,
+      "grad_norm": 1.0443862676620483,
+      "learning_rate": 2.5093689575056045e-05,
+      "loss": 0.4448,
+      "step": 445
+    },
+    {
+      "epoch": 1.5306122448979593,
+      "grad_norm": 1.130210280418396,
+      "learning_rate": 2.4968201599961445e-05,
+      "loss": 0.439,
+      "step": 450
+    },
+    {
+      "epoch": 1.5476190476190477,
+      "grad_norm": 1.3261239528656006,
+      "learning_rate": 2.4841451561537496e-05,
+      "loss": 0.4577,
+      "step": 455
+    },
+    {
+      "epoch": 1.564625850340136,
+      "grad_norm": 1.1579346656799316,
+      "learning_rate": 2.471345550747082e-05,
+      "loss": 0.4915,
+      "step": 460
+    },
+    {
+      "epoch": 1.5816326530612246,
+      "grad_norm": 1.1824404001235962,
+      "learning_rate": 2.4584229643204755e-05,
+      "loss": 0.4404,
+      "step": 465
+    },
+    {
+      "epoch": 1.598639455782313,
+      "grad_norm": 1.175586223602295,
+      "learning_rate": 2.4453790329887578e-05,
+      "loss": 0.4649,
+      "step": 470
+    },
+    {
+      "epoch": 1.6156462585034013,
+      "grad_norm": 1.5084452629089355,
+      "learning_rate": 2.4322154082301065e-05,
+      "loss": 0.4425,
+      "step": 475
+    },
+    {
+      "epoch": 1.6326530612244898,
+      "grad_norm": 1.2946665287017822,
+      "learning_rate": 2.4189337566769545e-05,
+      "loss": 0.47,
+      "step": 480
+    },
+    {
+      "epoch": 1.6496598639455784,
+      "grad_norm": 1.0975319147109985,
+      "learning_rate": 2.4055357599049807e-05,
+      "loss": 0.4747,
+      "step": 485
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 1.0334219932556152,
+      "learning_rate": 2.392023114220209e-05,
+      "loss": 0.4076,
+      "step": 490
+    },
+    {
+      "epoch": 1.683673469387755,
+      "grad_norm": 1.2870817184448242,
+      "learning_rate": 2.378397530444238e-05,
+      "loss": 0.4198,
+      "step": 495
+    },
+    {
+      "epoch": 1.7006802721088436,
+      "grad_norm": 1.0723446607589722,
+      "learning_rate": 2.3646607336976375e-05,
+      "loss": 0.4536,
+      "step": 500
+    },
+    {
+      "epoch": 1.717687074829932,
+      "grad_norm": 1.3169423341751099,
+      "learning_rate": 2.3508144631815326e-05,
+      "loss": 0.4213,
+      "step": 505
+    },
+    {
+      "epoch": 1.7346938775510203,
+      "grad_norm": 1.2261208295822144,
+      "learning_rate": 2.3368604719574055e-05,
+      "loss": 0.4414,
+      "step": 510
+    },
+    {
+      "epoch": 1.751700680272109,
+      "grad_norm": 1.2695997953414917,
+      "learning_rate": 2.322800526725141e-05,
+      "loss": 0.3781,
+      "step": 515
+    },
+    {
+      "epoch": 1.7687074829931972,
+      "grad_norm": 1.421828269958496,
+      "learning_rate": 2.308636407599347e-05,
+      "loss": 0.4473,
+      "step": 520
+    },
+    {
+      "epoch": 1.7857142857142856,
+      "grad_norm": 1.2838345766067505,
+      "learning_rate": 2.2943699078839783e-05,
+      "loss": 0.4126,
+      "step": 525
+    },
+    {
+      "epoch": 1.8027210884353742,
+      "grad_norm": 1.290840983390808,
+      "learning_rate": 2.2800028338452853e-05,
+      "loss": 0.4035,
+      "step": 530
+    },
+    {
+      "epoch": 1.8197278911564627,
+      "grad_norm": 1.3847336769104004,
+      "learning_rate": 2.2655370044831253e-05,
+      "loss": 0.378,
+      "step": 535
+    },
+    {
+      "epoch": 1.836734693877551,
+      "grad_norm": 1.149859070777893,
+      "learning_rate": 2.2509742513006633e-05,
+      "loss": 0.3737,
+      "step": 540
+    },
+    {
+      "epoch": 1.8537414965986394,
+      "grad_norm": 1.0945329666137695,
+      "learning_rate": 2.2363164180724828e-05,
+      "loss": 0.4016,
+      "step": 545
+    },
+    {
+      "epoch": 1.870748299319728,
+      "grad_norm": 1.2481576204299927,
+      "learning_rate": 2.2215653606111515e-05,
+      "loss": 0.4054,
+      "step": 550
+    },
+    {
+      "epoch": 1.8877551020408163,
+      "grad_norm": 1.065503478050232,
+      "learning_rate": 2.2067229465322578e-05,
+      "loss": 0.3944,
+      "step": 555
+    },
+    {
+      "epoch": 1.9047619047619047,
+      "grad_norm": 1.2339270114898682,
+      "learning_rate": 2.1917910550179527e-05,
+      "loss": 0.3505,
+      "step": 560
+    },
+    {
+      "epoch": 1.9217687074829932,
+      "grad_norm": 1.1712156534194946,
+      "learning_rate": 2.1767715765790303e-05,
+      "loss": 0.3497,
+      "step": 565
+    },
+    {
+      "epoch": 1.9387755102040818,
+      "grad_norm": 1.4303723573684692,
+      "learning_rate": 2.16166641281557e-05,
+      "loss": 0.3772,
+      "step": 570
+    },
+    {
+      "epoch": 1.95578231292517,
+      "grad_norm": 1.123719573020935,
+      "learning_rate": 2.1464774761761805e-05,
+      "loss": 0.3558,
+      "step": 575
+    },
+    {
+      "epoch": 1.9727891156462585,
+      "grad_norm": 1.1126965284347534,
+      "learning_rate": 2.131206689715863e-05,
+      "loss": 0.3763,
+      "step": 580
+    },
+    {
+      "epoch": 1.989795918367347,
+      "grad_norm": 1.0419843196868896,
+      "learning_rate": 2.1158559868525374e-05,
+      "loss": 0.3507,
+      "step": 585
+    },
+    {
+      "epoch": 2.006802721088435,
+      "grad_norm": 1.15863835811615,
+      "learning_rate": 2.1004273111222532e-05,
+      "loss": 0.3066,
+      "step": 590
+    },
+    {
+      "epoch": 2.0238095238095237,
+      "grad_norm": 1.1892127990722656,
+      "learning_rate": 2.0849226159331222e-05,
+      "loss": 0.2585,
+      "step": 595
+    },
+    {
+      "epoch": 2.0408163265306123,
+      "grad_norm": 1.0521714687347412,
+      "learning_rate": 2.069343864317998e-05,
+      "loss": 0.2707,
+      "step": 600
+    },
+    {
+      "epoch": 2.057823129251701,
+      "grad_norm": 1.334870457649231,
+      "learning_rate": 2.053693028685938e-05,
+      "loss": 0.2645,
+      "step": 605
+    },
+    {
+      "epoch": 2.074829931972789,
+      "grad_norm": 1.16738760471344,
+      "learning_rate": 2.0379720905724818e-05,
+      "loss": 0.2664,
+      "step": 610
+    },
+    {
+      "epoch": 2.0918367346938775,
+      "grad_norm": 1.2727965116500854,
+      "learning_rate": 2.022183040388767e-05,
+      "loss": 0.2688,
+      "step": 615
+    },
+    {
+      "epoch": 2.108843537414966,
+      "grad_norm": 1.0208539962768555,
+      "learning_rate": 2.006327877169529e-05,
+      "loss": 0.2724,
+      "step": 620
+    },
+    {
+      "epoch": 2.1258503401360542,
+      "grad_norm": 1.7100193500518799,
+      "learning_rate": 1.9904086083200035e-05,
+      "loss": 0.2936,
+      "step": 625
+    },
+    {
+      "epoch": 2.142857142857143,
+      "grad_norm": 1.143485426902771,
+      "learning_rate": 1.9744272493617703e-05,
+      "loss": 0.2851,
+      "step": 630
+    },
+    {
+      "epoch": 2.1598639455782314,
+      "grad_norm": 1.1511744260787964,
+      "learning_rate": 1.9583858236775733e-05,
+      "loss": 0.2602,
+      "step": 635
+    },
+    {
+      "epoch": 2.17687074829932,
+      "grad_norm": 1.3126564025878906,
+      "learning_rate": 1.9422863622551376e-05,
+      "loss": 0.2915,
+      "step": 640
+    },
+    {
+      "epoch": 2.193877551020408,
+      "grad_norm": 1.1826177835464478,
+      "learning_rate": 1.926130903430034e-05,
+      "loss": 0.2847,
+      "step": 645
+    },
+    {
+      "epoch": 2.2108843537414966,
+      "grad_norm": 1.19844651222229,
+      "learning_rate": 1.9099214926276024e-05,
+      "loss": 0.3337,
+      "step": 650
+    },
+    {
+      "epoch": 2.227891156462585,
+      "grad_norm": 1.5924816131591797,
+      "learning_rate": 1.8936601821039858e-05,
+      "loss": 0.252,
+      "step": 655
+    },
+    {
+      "epoch": 2.2448979591836733,
+      "grad_norm": 1.3154438734054565,
+      "learning_rate": 1.8773490306862945e-05,
+      "loss": 0.2506,
+      "step": 660
+    },
+    {
+      "epoch": 2.261904761904762,
+      "grad_norm": 1.09384024143219,
+      "learning_rate": 1.860990103511941e-05,
+      "loss": 0.2947,
+      "step": 665
+    },
+    {
+      "epoch": 2.2789115646258504,
+      "grad_norm": 1.2651005983352661,
+      "learning_rate": 1.8445854717671768e-05,
+      "loss": 0.2845,
+      "step": 670
+    },
+    {
+      "epoch": 2.295918367346939,
+      "grad_norm": 1.198618769645691,
+      "learning_rate": 1.828137212424858e-05,
+      "loss": 0.2675,
+      "step": 675
+    },
+    {
+      "epoch": 2.312925170068027,
+      "grad_norm": 1.225216269493103,
+      "learning_rate": 1.8116474079814855e-05,
+      "loss": 0.255,
+      "step": 680
+    },
+    {
+      "epoch": 2.3299319727891157,
+      "grad_norm": 1.2420880794525146,
+      "learning_rate": 1.79511814619354e-05,
+      "loss": 0.2276,
+      "step": 685
+    },
+    {
+      "epoch": 2.3469387755102042,
+      "grad_norm": 1.2797255516052246,
+      "learning_rate": 1.7785515198131556e-05,
+      "loss": 0.2761,
+      "step": 690
+    },
+    {
+      "epoch": 2.3639455782312924,
+      "grad_norm": 1.1373989582061768,
+      "learning_rate": 1.7619496263231557e-05,
+      "loss": 0.2634,
+      "step": 695
+    },
+    {
+      "epoch": 2.380952380952381,
+      "grad_norm": 1.0364551544189453,
+      "learning_rate": 1.745314567671496e-05,
+      "loss": 0.278,
+      "step": 700
+    },
+    {
+      "epoch": 2.3979591836734695,
+      "grad_norm": 1.3783962726593018,
+      "learning_rate": 1.7286484500051383e-05,
+      "loss": 0.2569,
+      "step": 705
+    },
+    {
+      "epoch": 2.4149659863945576,
+      "grad_norm": 1.2774587869644165,
+      "learning_rate": 1.7119533834033907e-05,
+      "loss": 0.2515,
+      "step": 710
+    },
+    {
+      "epoch": 2.431972789115646,
+      "grad_norm": 1.2328591346740723,
+      "learning_rate": 1.6952314816107576e-05,
+      "loss": 0.2344,
+      "step": 715
+    },
+    {
+      "epoch": 2.4489795918367347,
+      "grad_norm": 1.1517112255096436,
+      "learning_rate": 1.678484861769316e-05,
+      "loss": 0.2649,
+      "step": 720
+    },
+    {
+      "epoch": 2.4659863945578233,
+      "grad_norm": 1.1749553680419922,
+      "learning_rate": 1.661715644150671e-05,
+      "loss": 0.2284,
+      "step": 725
+    },
+    {
+      "epoch": 2.4829931972789114,
+      "grad_norm": 1.298263669013977,
+      "learning_rate": 1.644925951887505e-05,
+      "loss": 0.2178,
+      "step": 730
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 1.0253585577011108,
+      "learning_rate": 1.6281179107047765e-05,
+      "loss": 0.2464,
+      "step": 735
+    },
+    {
+      "epoch": 2.5170068027210886,
+      "grad_norm": 1.2435274124145508,
+      "learning_rate": 1.6112936486505785e-05,
+      "loss": 0.2129,
+      "step": 740
+    },
+    {
+      "epoch": 2.534013605442177,
+      "grad_norm": 1.1542493104934692,
+      "learning_rate": 1.5944552958267118e-05,
+      "loss": 0.2152,
+      "step": 745
+    },
+    {
+      "epoch": 2.5510204081632653,
+      "grad_norm": 1.0975438356399536,
+      "learning_rate": 1.5776049841189958e-05,
+      "loss": 0.2312,
+      "step": 750
+    },
+    {
+      "epoch": 2.568027210884354,
+      "grad_norm": 1.0609310865402222,
+      "learning_rate": 1.5607448469273495e-05,
+      "loss": 0.2426,
+      "step": 755
+    },
+    {
+      "epoch": 2.5850340136054424,
+      "grad_norm": 1.144705057144165,
+      "learning_rate": 1.543877018895687e-05,
+      "loss": 0.2047,
+      "step": 760
+    },
+    {
+      "epoch": 2.6020408163265305,
+      "grad_norm": 1.1797125339508057,
+      "learning_rate": 1.5270036356416515e-05,
+      "loss": 0.2153,
+      "step": 765
+    },
+    {
+      "epoch": 2.619047619047619,
+      "grad_norm": 1.1653424501419067,
+      "learning_rate": 1.5101268334862263e-05,
+      "loss": 0.2314,
+      "step": 770
+    },
+    {
+      "epoch": 2.6360544217687076,
+      "grad_norm": 1.0752824544906616,
+      "learning_rate": 1.4932487491832584e-05,
+      "loss": 0.2224,
+      "step": 775
+    },
+    {
+      "epoch": 2.6530612244897958,
+      "grad_norm": 1.1791971921920776,
+      "learning_rate": 1.4763715196489263e-05,
+      "loss": 0.2004,
+      "step": 780
+    },
+    {
+      "epoch": 2.6700680272108843,
+      "grad_norm": 1.1845297813415527,
+      "learning_rate": 1.4594972816911873e-05,
+      "loss": 0.2555,
+      "step": 785
+    },
+    {
+      "epoch": 2.687074829931973,
+      "grad_norm": 1.2175406217575073,
+      "learning_rate": 1.4426281717392377e-05,
+      "loss": 0.235,
+      "step": 790
+    },
+    {
+      "epoch": 2.704081632653061,
+      "grad_norm": 1.1495968103408813,
+      "learning_rate": 1.4257663255730234e-05,
+      "loss": 0.2297,
+      "step": 795
+    },
+    {
+      "epoch": 2.7210884353741496,
+      "grad_norm": 1.2075597047805786,
+      "learning_rate": 1.4089138780528287e-05,
+      "loss": 0.1798,
+      "step": 800
+    },
+    {
+      "epoch": 2.738095238095238,
+      "grad_norm": 1.117581844329834,
+      "learning_rate": 1.392072962848988e-05,
+      "loss": 0.2476,
+      "step": 805
+    },
+    {
+      "epoch": 2.7551020408163263,
+      "grad_norm": 1.2183245420455933,
+      "learning_rate": 1.3752457121717383e-05,
+      "loss": 0.2281,
+      "step": 810
+    },
+    {
+      "epoch": 2.772108843537415,
+      "grad_norm": 1.198248267173767,
+      "learning_rate": 1.3584342565012677e-05,
+      "loss": 0.2031,
+      "step": 815
+    },
+    {
+      "epoch": 2.7891156462585034,
+      "grad_norm": 1.257199764251709,
+      "learning_rate": 1.341640724317975e-05,
+      "loss": 0.2295,
+      "step": 820
+    },
+    {
+      "epoch": 2.806122448979592,
+      "grad_norm": 1.1676485538482666,
+      "learning_rate": 1.324867241832985e-05,
+      "loss": 0.1996,
+      "step": 825
+    },
+    {
+      "epoch": 2.8231292517006805,
+      "grad_norm": 0.9781966209411621,
+      "learning_rate": 1.3081159327189524e-05,
+      "loss": 0.2001,
+      "step": 830
+    },
+    {
+      "epoch": 2.8401360544217686,
+      "grad_norm": 1.2090445756912231,
+      "learning_rate": 1.2913889178411837e-05,
+      "loss": 0.2193,
+      "step": 835
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 1.130637764930725,
+      "learning_rate": 1.2746883149891203e-05,
+      "loss": 0.1832,
+      "step": 840
+    },
+    {
+      "epoch": 2.8741496598639458,
+      "grad_norm": 1.2970364093780518,
+      "learning_rate": 1.2580162386082028e-05,
+      "loss": 0.2086,
+      "step": 845
+    },
+    {
+      "epoch": 2.891156462585034,
+      "grad_norm": 1.0856074094772339,
+      "learning_rate": 1.2413747995321692e-05,
+      "loss": 0.1941,
+      "step": 850
+    },
+    {
+      "epoch": 2.9081632653061225,
+      "grad_norm": 1.1722519397735596,
+      "learning_rate": 1.2247661047157968e-05,
+      "loss": 0.1959,
+      "step": 855
+    },
+    {
+      "epoch": 2.925170068027211,
+      "grad_norm": 1.1799602508544922,
+      "learning_rate": 1.208192256968151e-05,
+      "loss": 0.1968,
+      "step": 860
+    },
+    {
+      "epoch": 2.942176870748299,
+      "grad_norm": 1.1089502573013306,
+      "learning_rate": 1.191655354686346e-05,
+      "loss": 0.2048,
+      "step": 865
+    },
+    {
+      "epoch": 2.9591836734693877,
+      "grad_norm": 1.1572333574295044,
+      "learning_rate": 1.175157491589869e-05,
+      "loss": 0.2141,
+      "step": 870
+    },
+    {
+      "epoch": 2.9761904761904763,
+      "grad_norm": 1.2635926008224487,
+      "learning_rate": 1.1587007564554991e-05,
+      "loss": 0.1973,
+      "step": 875
+    },
+    {
+      "epoch": 2.9931972789115644,
+      "grad_norm": 1.1943570375442505,
+      "learning_rate": 1.1422872328528473e-05,
+      "loss": 0.2053,
+      "step": 880
+    },
+    {
+      "epoch": 3.010204081632653,
+      "grad_norm": 1.0038658380508423,
+      "learning_rate": 1.1259189988805599e-05,
+      "loss": 0.1666,
+      "step": 885
+    },
+    {
+      "epoch": 3.0272108843537415,
+      "grad_norm": 1.3621819019317627,
+      "learning_rate": 1.1095981269032097e-05,
+      "loss": 0.1502,
+      "step": 890
+    },
+    {
+      "epoch": 3.04421768707483,
+      "grad_norm": 1.0059318542480469,
+      "learning_rate": 1.0933266832889206e-05,
+      "loss": 0.1665,
+      "step": 895
+    },
+    {
+      "epoch": 3.061224489795918,
+      "grad_norm": 1.1601946353912354,
+      "learning_rate": 1.077106728147743e-05,
+      "loss": 0.1801,
+      "step": 900
+    },
+    {
+      "epoch": 3.078231292517007,
+      "grad_norm": 1.3520936965942383,
+      "learning_rate": 1.0609403150708261e-05,
+      "loss": 0.1305,
+      "step": 905
+    },
+    {
+      "epoch": 3.0952380952380953,
+      "grad_norm": 1.0658321380615234,
+      "learning_rate": 1.0448294908704173e-05,
+      "loss": 0.127,
+      "step": 910
+    },
+    {
+      "epoch": 3.1122448979591835,
+      "grad_norm": 1.2596958875656128,
+      "learning_rate": 1.028776295320714e-05,
+      "loss": 0.1592,
+      "step": 915
+    },
+    {
+      "epoch": 3.129251700680272,
+      "grad_norm": 0.9525283575057983,
+      "learning_rate": 1.0127827608996144e-05,
+      "loss": 0.1419,
+      "step": 920
+    },
+    {
+      "epoch": 3.1462585034013606,
+      "grad_norm": 1.0077766180038452,
+      "learning_rate": 9.968509125313823e-06,
+      "loss": 0.1314,
+      "step": 925
+    },
+    {
+      "epoch": 3.163265306122449,
+      "grad_norm": 1.0083332061767578,
+      "learning_rate": 9.80982767330278e-06,
+      "loss": 0.1479,
+      "step": 930
+    },
+    {
+      "epoch": 3.1802721088435373,
+      "grad_norm": 0.9684851169586182,
+      "learning_rate": 9.651803343451729e-06,
+      "loss": 0.1385,
+      "step": 935
+    },
+    {
+      "epoch": 3.197278911564626,
+      "grad_norm": 1.190239429473877,
+      "learning_rate": 9.494456143051827e-06,
+      "loss": 0.1361,
+      "step": 940
+    },
+    {
+      "epoch": 3.2142857142857144,
+      "grad_norm": 0.989656925201416,
+      "learning_rate": 9.337805993663618e-06,
+      "loss": 0.1486,
+      "step": 945
+    },
+    {
+      "epoch": 3.2312925170068025,
+      "grad_norm": 1.1670643091201782,
+      "learning_rate": 9.181872728594747e-06,
+      "loss": 0.1476,
+      "step": 950
+    },
+    {
+      "epoch": 3.248299319727891,
+      "grad_norm": 1.0360350608825684,
+      "learning_rate": 9.02667609038892e-06,
+      "loss": 0.1355,
+      "step": 955
+    },
+    {
+      "epoch": 3.2653061224489797,
+      "grad_norm": 1.2193771600723267,
+      "learning_rate": 8.872235728326284e-06,
+      "loss": 0.1634,
+      "step": 960
+    },
+    {
+      "epoch": 3.282312925170068,
+      "grad_norm": 0.9404177665710449,
+      "learning_rate": 8.718571195935696e-06,
+      "loss": 0.144,
+      "step": 965
+    },
+    {
+      "epoch": 3.2993197278911564,
+      "grad_norm": 0.910885751247406,
+      "learning_rate": 8.565701948519034e-06,
+      "loss": 0.1308,
+      "step": 970
+    },
+    {
+      "epoch": 3.316326530612245,
+      "grad_norm": 1.0395634174346924,
+      "learning_rate": 8.413647340688e-06,
+      "loss": 0.1292,
+      "step": 975
+    },
+    {
+      "epoch": 3.3333333333333335,
+      "grad_norm": 0.9528372287750244,
+      "learning_rate": 8.262426623913663e-06,
+      "loss": 0.1329,
+      "step": 980
+    },
+    {
+      "epoch": 3.3503401360544216,
+      "grad_norm": 1.2525593042373657,
+      "learning_rate": 8.112058944089003e-06,
+      "loss": 0.1458,
+      "step": 985
+    },
+    {
+      "epoch": 3.36734693877551,
+      "grad_norm": 1.2458698749542236,
+      "learning_rate": 7.96256333910496e-06,
+      "loss": 0.1381,
+      "step": 990
+    },
+    {
+      "epoch": 3.3843537414965987,
+      "grad_norm": 1.031166434288025,
+      "learning_rate": 7.81395873643996e-06,
+      "loss": 0.1378,
+      "step": 995
+    },
+    {
+      "epoch": 3.4013605442176873,
+      "grad_norm": 1.1444342136383057,
+      "learning_rate": 7.666263950763607e-06,
+      "loss": 0.1256,
+      "step": 1000
+    },
+    {
+      "epoch": 3.4183673469387754,
+      "grad_norm": 1.0373079776763916,
+      "learning_rate": 7.519497681554551e-06,
+      "loss": 0.1287,
+      "step": 1005
+    },
+    {
+      "epoch": 3.435374149659864,
+      "grad_norm": 1.1225919723510742,
+      "learning_rate": 7.373678510732955e-06,
+      "loss": 0.1363,
+      "step": 1010
+    },
+    {
+      "epoch": 3.4523809523809526,
+      "grad_norm": 1.0068609714508057,
+      "learning_rate": 7.228824900307881e-06,
+      "loss": 0.1227,
+      "step": 1015
+    },
+    {
+      "epoch": 3.4693877551020407,
+      "grad_norm": 1.1868841648101807,
+      "learning_rate": 7.0849551900398065e-06,
+      "loss": 0.1162,
+      "step": 1020
+    },
+    {
+      "epoch": 3.4863945578231292,
+      "grad_norm": 0.9918032288551331,
+      "learning_rate": 6.94208759511868e-06,
+      "loss": 0.1289,
+      "step": 1025
+    },
+    {
+      "epoch": 3.503401360544218,
+      "grad_norm": 1.0155752897262573,
+      "learning_rate": 6.800240203857696e-06,
+      "loss": 0.138,
+      "step": 1030
+    },
+    {
+      "epoch": 3.520408163265306,
+      "grad_norm": 0.9242071509361267,
+      "learning_rate": 6.659430975403156e-06,
+      "loss": 0.1375,
+      "step": 1035
+    },
+    {
+      "epoch": 3.5374149659863945,
+      "grad_norm": 1.0767987966537476,
+      "learning_rate": 6.519677737460664e-06,
+      "loss": 0.1146,
+      "step": 1040
+    },
+    {
+      "epoch": 3.554421768707483,
+      "grad_norm": 0.9681980013847351,
+      "learning_rate": 6.380998184038025e-06,
+      "loss": 0.1365,
+      "step": 1045
+    },
+    {
+      "epoch": 3.571428571428571,
+      "grad_norm": 0.9982250332832336,
+      "learning_rate": 6.243409873204984e-06,
+      "loss": 0.1307,
+      "step": 1050
+    },
+    {
+      "epoch": 3.5884353741496597,
+      "grad_norm": 0.9535261392593384,
+      "learning_rate": 6.106930224870213e-06,
+      "loss": 0.1305,
+      "step": 1055
+    },
+    {
+      "epoch": 3.6054421768707483,
+      "grad_norm": 0.9636424779891968,
+      "learning_rate": 5.9715765185758395e-06,
+      "loss": 0.1205,
+      "step": 1060
+    },
+    {
+      "epoch": 3.622448979591837,
+      "grad_norm": 1.060290813446045,
+      "learning_rate": 5.8373658913096586e-06,
+      "loss": 0.1068,
+      "step": 1065
+    },
+    {
+      "epoch": 3.6394557823129254,
+      "grad_norm": 0.9654757380485535,
+      "learning_rate": 5.704315335335468e-06,
+      "loss": 0.142,
+      "step": 1070
+    },
+    {
+      "epoch": 3.6564625850340136,
+      "grad_norm": 0.9803591966629028,
+      "learning_rate": 5.572441696041653e-06,
+      "loss": 0.121,
+      "step": 1075
+    },
+    {
+      "epoch": 3.673469387755102,
+      "grad_norm": 0.9896225333213806,
+      "learning_rate": 5.4417616698084785e-06,
+      "loss": 0.1346,
+      "step": 1080
+    },
+    {
+      "epoch": 3.6904761904761907,
+      "grad_norm": 0.9817864298820496,
+      "learning_rate": 5.3122918018941044e-06,
+      "loss": 0.1276,
+      "step": 1085
+    },
+    {
+      "epoch": 3.707482993197279,
+      "grad_norm": 0.9709451794624329,
+      "learning_rate": 5.184048484339855e-06,
+      "loss": 0.1121,
+      "step": 1090
+    },
+    {
+      "epoch": 3.7244897959183674,
+      "grad_norm": 0.9238762259483337,
+      "learning_rate": 5.057047953894831e-06,
+      "loss": 0.1198,
+      "step": 1095
+    },
+    {
+      "epoch": 3.741496598639456,
+      "grad_norm": 0.9318645000457764,
+      "learning_rate": 4.931306289960181e-06,
+      "loss": 0.1254,
+      "step": 1100
+    },
+    {
+      "epoch": 3.758503401360544,
+      "grad_norm": 0.9292995929718018,
+      "learning_rate": 4.806839412553321e-06,
+      "loss": 0.1129,
+      "step": 1105
+    },
+    {
+      "epoch": 3.7755102040816326,
+      "grad_norm": 0.9565867185592651,
+      "learning_rate": 4.6836630802922955e-06,
+      "loss": 0.1088,
+      "step": 1110
+    },
+    {
+      "epoch": 3.792517006802721,
+      "grad_norm": 1.0380818843841553,
+      "learning_rate": 4.5617928884006234e-06,
+      "loss": 0.1002,
+      "step": 1115
+    },
+    {
+      "epoch": 3.8095238095238093,
+      "grad_norm": 1.0153783559799194,
+      "learning_rate": 4.441244266732786e-06,
+      "loss": 0.1002,
+      "step": 1120
+    },
+    {
+      "epoch": 3.826530612244898,
+      "grad_norm": 0.9576224088668823,
+      "learning_rate": 4.3220324778206735e-06,
+      "loss": 0.1272,
+      "step": 1125
+    },
+    {
+      "epoch": 3.8435374149659864,
+      "grad_norm": 0.9821186661720276,
+      "learning_rate": 4.204172614941214e-06,
+      "loss": 0.1186,
+      "step": 1130
+    },
+    {
+      "epoch": 3.8605442176870746,
+      "grad_norm": 0.9095916152000427,
+      "learning_rate": 4.087679600205425e-06,
+      "loss": 0.1312,
+      "step": 1135
+    },
+    {
+      "epoch": 3.877551020408163,
+      "grad_norm": 0.8640242218971252,
+      "learning_rate": 3.9725681826691525e-06,
+      "loss": 0.1124,
+      "step": 1140
+    },
+    {
+      "epoch": 3.8945578231292517,
+      "grad_norm": 0.9114555716514587,
+      "learning_rate": 3.858852936465688e-06,
+      "loss": 0.1027,
+      "step": 1145
+    },
+    {
+      "epoch": 3.9115646258503403,
+      "grad_norm": 1.0309470891952515,
+      "learning_rate": 3.7465482589605713e-06,
+      "loss": 0.1207,
+      "step": 1150
+    },
+    {
+      "epoch": 3.928571428571429,
+      "grad_norm": 0.7389628887176514,
+      "learning_rate": 3.6356683689287566e-06,
+      "loss": 0.1037,
+      "step": 1155
+    },
+    {
+      "epoch": 3.945578231292517,
+      "grad_norm": 0.9068615436553955,
+      "learning_rate": 3.5262273047543787e-06,
+      "loss": 0.1144,
+      "step": 1160
+    },
+    {
+      "epoch": 3.9625850340136055,
+      "grad_norm": 0.9940776824951172,
+      "learning_rate": 3.418238922653357e-06,
+      "loss": 0.1152,
+      "step": 1165
+    },
+    {
+      "epoch": 3.979591836734694,
+      "grad_norm": 0.9465267062187195,
+      "learning_rate": 3.3117168949191134e-06,
+      "loss": 0.1172,
+      "step": 1170
+    },
+    {
+      "epoch": 3.996598639455782,
+      "grad_norm": 0.8871757388114929,
+      "learning_rate": 3.206674708191502e-06,
+      "loss": 0.1064,
+      "step": 1175
+    },
+    {
+      "epoch": 4.01360544217687,
+      "grad_norm": 0.6294886469841003,
+      "learning_rate": 3.1031256617492843e-06,
+      "loss": 0.0905,
+      "step": 1180
+    },
+    {
+      "epoch": 4.030612244897959,
+      "grad_norm": 0.8446071147918701,
+      "learning_rate": 3.00108286582634e-06,
+      "loss": 0.0926,
+      "step": 1185
+    },
+    {
+      "epoch": 4.0476190476190474,
+      "grad_norm": 0.9118359088897705,
+      "learning_rate": 2.900559239951781e-06,
+      "loss": 0.1084,
+      "step": 1190
+    },
+    {
+      "epoch": 4.0646258503401365,
+      "grad_norm": 0.8552963733673096,
+      "learning_rate": 2.8015675113142285e-06,
+      "loss": 0.0788,
+      "step": 1195
+    },
+    {
+      "epoch": 4.081632653061225,
+      "grad_norm": 0.7456033229827881,
+      "learning_rate": 2.704120213150423e-06,
+      "loss": 0.0892,
+      "step": 1200
+    },
+    {
+      "epoch": 4.098639455782313,
+      "grad_norm": 0.8816990256309509,
+      "learning_rate": 2.6082296831584367e-06,
+      "loss": 0.0843,
+      "step": 1205
+    },
+    {
+      "epoch": 4.115646258503402,
+      "grad_norm": 0.8171160817146301,
+      "learning_rate": 2.5139080619355798e-06,
+      "loss": 0.101,
+      "step": 1210
+    },
+    {
+      "epoch": 4.13265306122449,
+      "grad_norm": 0.8481881618499756,
+      "learning_rate": 2.4211672914412947e-06,
+      "loss": 0.0843,
+      "step": 1215
+    },
+    {
+      "epoch": 4.149659863945578,
+      "grad_norm": 0.7491337060928345,
+      "learning_rate": 2.3300191134852193e-06,
+      "loss": 0.0937,
+      "step": 1220
+    },
+    {
+      "epoch": 4.166666666666667,
+      "grad_norm": 0.8872839212417603,
+      "learning_rate": 2.2404750682405546e-06,
+      "loss": 0.0772,
+      "step": 1225
+    },
+    {
+      "epoch": 4.183673469387755,
+      "grad_norm": 0.8537187576293945,
+      "learning_rate": 2.152546492782973e-06,
+      "loss": 0.0948,
+      "step": 1230
+    },
+    {
+      "epoch": 4.200680272108843,
+      "grad_norm": 0.8278796076774597,
+      "learning_rate": 2.0662445196552444e-06,
+      "loss": 0.0882,
+      "step": 1235
+    },
+    {
+      "epoch": 4.217687074829932,
+      "grad_norm": 0.8022066354751587,
+      "learning_rate": 1.981580075457763e-06,
+      "loss": 0.0883,
+      "step": 1240
+    },
+    {
+      "epoch": 4.23469387755102,
+      "grad_norm": 0.8415008783340454,
+      "learning_rate": 1.8985638794651467e-06,
+      "loss": 0.1037,
+      "step": 1245
+    },
+    {
+      "epoch": 4.2517006802721085,
+      "grad_norm": 0.7159227132797241,
+      "learning_rate": 1.8172064422690577e-06,
+      "loss": 0.0793,
+      "step": 1250
+    },
+    {
+      "epoch": 4.2687074829931975,
+      "grad_norm": 0.8320397138595581,
+      "learning_rate": 1.737518064447493e-06,
+      "loss": 0.0773,
+      "step": 1255
+    },
+    {
+      "epoch": 4.285714285714286,
+      "grad_norm": 0.8053768277168274,
+      "learning_rate": 1.659508835260632e-06,
+      "loss": 0.1018,
+      "step": 1260
+    },
+    {
+      "epoch": 4.302721088435375,
+      "grad_norm": 0.914559006690979,
+      "learning_rate": 1.5831886313734383e-06,
+      "loss": 0.0853,
+      "step": 1265
+    },
+    {
+      "epoch": 4.319727891156463,
+      "grad_norm": 0.7460253834724426,
+      "learning_rate": 1.5085671156051905e-06,
+      "loss": 0.0821,
+      "step": 1270
+    },
+    {
+      "epoch": 4.336734693877551,
+      "grad_norm": 0.6933921575546265,
+      "learning_rate": 1.4356537357060857e-06,
+      "loss": 0.0819,
+      "step": 1275
+    },
+    {
+      "epoch": 4.35374149659864,
+      "grad_norm": 0.8233038783073425,
+      "learning_rate": 1.36445772316107e-06,
+      "loss": 0.0865,
+      "step": 1280
+    },
+    {
+      "epoch": 4.370748299319728,
+      "grad_norm": 0.8251399397850037,
+      "learning_rate": 1.2949880920210461e-06,
+      "loss": 0.0852,
+      "step": 1285
+    },
+    {
+      "epoch": 4.387755102040816,
+      "grad_norm": 0.7801070213317871,
+      "learning_rate": 1.2272536377616134e-06,
+      "loss": 0.0803,
+      "step": 1290
+    },
+    {
+      "epoch": 4.404761904761905,
+      "grad_norm": 0.9840799570083618,
+      "learning_rate": 1.1612629361694877e-06,
+      "loss": 0.0809,
+      "step": 1295
+    },
+    {
+      "epoch": 4.421768707482993,
+      "grad_norm": 0.7471717596054077,
+      "learning_rate": 1.097024342256731e-06,
+      "loss": 0.084,
+      "step": 1300
+    },
+    {
+      "epoch": 4.438775510204081,
+      "grad_norm": 0.7413614988327026,
+      "learning_rate": 1.034545989202917e-06,
+      "loss": 0.0836,
+      "step": 1305
+    },
+    {
+      "epoch": 4.45578231292517,
+      "grad_norm": 0.877647876739502,
+      "learning_rate": 9.738357873254178e-07,
+      "loss": 0.0986,
+      "step": 1310
+    },
+    {
+      "epoch": 4.4727891156462585,
+      "grad_norm": 0.9114505052566528,
+      "learning_rate": 9.149014230778774e-07,
+      "loss": 0.0859,
+      "step": 1315
+    },
+    {
+      "epoch": 4.489795918367347,
+      "grad_norm": 0.8558387756347656,
+      "learning_rate": 8.577503580770402e-07,
+      "loss": 0.0838,
+      "step": 1320
+    },
+    {
+      "epoch": 4.506802721088436,
+      "grad_norm": 0.8477901220321655,
+      "learning_rate": 8.023898281580406e-07,
+      "loss": 0.0871,
+      "step": 1325
+    },
+    {
+      "epoch": 4.523809523809524,
+      "grad_norm": 0.755340039730072,
+      "learning_rate": 7.488268424582839e-07,
+      "loss": 0.0878,
+      "step": 1330
+    },
+    {
+      "epoch": 4.540816326530612,
+      "grad_norm": 0.6771763563156128,
+      "learning_rate": 6.970681825300285e-07,
+      "loss": 0.0753,
+      "step": 1335
+    },
+    {
+      "epoch": 4.557823129251701,
+      "grad_norm": 0.8752419948577881,
+      "learning_rate": 6.471204014817767e-07,
+      "loss": 0.0919,
+      "step": 1340
+    },
+    {
+      "epoch": 4.574829931972789,
+      "grad_norm": 0.7726633548736572,
+      "learning_rate": 5.989898231485974e-07,
+      "loss": 0.096,
+      "step": 1345
+    },
+    {
+      "epoch": 4.591836734693878,
+      "grad_norm": 0.7407664060592651,
+      "learning_rate": 5.526825412914683e-07,
+      "loss": 0.0877,
+      "step": 1350
+    },
+    {
+      "epoch": 4.608843537414966,
+      "grad_norm": 0.8129183650016785,
+      "learning_rate": 5.08204418825754e-07,
+      "loss": 0.0887,
+      "step": 1355
+    },
+    {
+      "epoch": 4.625850340136054,
+      "grad_norm": 0.7982159852981567,
+      "learning_rate": 4.65561087078909e-07,
+      "loss": 0.0961,
+      "step": 1360
+    },
+    {
+      "epoch": 4.642857142857143,
+      "grad_norm": 0.774387776851654,
+      "learning_rate": 4.2475794507749246e-07,
+      "loss": 0.0852,
+      "step": 1365
+    },
+    {
+      "epoch": 4.659863945578231,
+      "grad_norm": 0.7400962114334106,
+      "learning_rate": 3.8580015886362173e-07,
+      "loss": 0.085,
+      "step": 1370
+    },
+    {
+      "epoch": 4.6768707482993195,
+      "grad_norm": 0.7531066536903381,
+      "learning_rate": 3.486926608408858e-07,
+      "loss": 0.0967,
+      "step": 1375
+    },
+    {
+      "epoch": 4.6938775510204085,
+      "grad_norm": 0.8446415066719055,
+      "learning_rate": 3.134401491498695e-07,
+      "loss": 0.094,
+      "step": 1380
+    },
+    {
+      "epoch": 4.710884353741497,
+      "grad_norm": 0.9277853965759277,
+      "learning_rate": 2.8004708707332626e-07,
+      "loss": 0.0921,
+      "step": 1385
+    },
+    {
+      "epoch": 4.727891156462585,
+      "grad_norm": 0.7945823073387146,
+      "learning_rate": 2.48517702471085e-07,
+      "loss": 0.0898,
+      "step": 1390
+    },
+    {
+      "epoch": 4.744897959183674,
+      "grad_norm": 0.7594613432884216,
+      "learning_rate": 2.1885598724476152e-07,
+      "loss": 0.0907,
+      "step": 1395
+    },
+    {
+      "epoch": 4.761904761904762,
+      "grad_norm": 0.682223916053772,
+      "learning_rate": 1.9106569683235485e-07,
+      "loss": 0.0916,
+      "step": 1400
+    },
+    {
+      "epoch": 4.77891156462585,
+      "grad_norm": 0.8218616843223572,
+      "learning_rate": 1.6515034973277886e-07,
+      "loss": 0.0666,
+      "step": 1405
+    },
+    {
+      "epoch": 4.795918367346939,
+      "grad_norm": 0.7728404402732849,
+      "learning_rate": 1.4111322706036878e-07,
+      "loss": 0.0821,
+      "step": 1410
+    },
+    {
+      "epoch": 4.812925170068027,
+      "grad_norm": 0.7478455901145935,
+      "learning_rate": 1.1895737212948843e-07,
+      "loss": 0.0775,
+      "step": 1415
+    },
+    {
+      "epoch": 4.829931972789115,
+      "grad_norm": 0.8629809617996216,
+      "learning_rate": 9.868559006920186e-08,
+      "loss": 0.089,
+      "step": 1420
+    },
+    {
+      "epoch": 4.846938775510204,
+      "grad_norm": 0.7577793002128601,
+      "learning_rate": 8.030044746812737e-08,
+      "loss": 0.0857,
+      "step": 1425
+    },
+    {
+      "epoch": 4.863945578231292,
+      "grad_norm": 0.8693432211875916,
+      "learning_rate": 6.380427204947914e-08,
+      "loss": 0.0969,
+      "step": 1430
+    },
+    {
+      "epoch": 4.880952380952381,
+      "grad_norm": 0.8243268132209778,
+      "learning_rate": 4.91991523763613e-08,
+      "loss": 0.0988,
+      "step": 1435
+    },
+    {
+      "epoch": 4.8979591836734695,
+      "grad_norm": 0.8658847808837891,
+      "learning_rate": 3.6486937587334456e-08,
+      "loss": 0.0847,
+      "step": 1440
+    },
+    {
+      "epoch": 4.914965986394558,
+      "grad_norm": 0.7587132453918457,
+      "learning_rate": 2.566923716229963e-08,
+      "loss": 0.0888,
+      "step": 1445
+    },
+    {
+      "epoch": 4.931972789115647,
+      "grad_norm": 0.7976932525634766,
+      "learning_rate": 1.6747420718722927e-08,
+      "loss": 0.0961,
+      "step": 1450
+    },
+    {
+      "epoch": 4.948979591836735,
+      "grad_norm": 0.9054105281829834,
+      "learning_rate": 9.722617838227587e-09,
+      "loss": 0.0875,
+      "step": 1455
+    },
+    {
+      "epoch": 4.965986394557823,
+      "grad_norm": 0.7605534791946411,
+      "learning_rate": 4.595717923585041e-09,
+      "loss": 0.077,
+      "step": 1460
+    },
+    {
+      "epoch": 4.982993197278912,
+      "grad_norm": 0.594926118850708,
+      "learning_rate": 1.3673700861033256e-09,
+      "loss": 0.0767,
+      "step": 1465
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.7051914930343628,
+      "learning_rate": 3.79830634444911e-11,
+      "loss": 0.0909,
+      "step": 1470
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1470,
+  "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": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.1488542864176579e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

48_128_e5_3e-5/checkpoint-1470/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1c73a3d3ae42383cf0a6ea1e73607dc0924ec4b32199fdb2961f1af8855e261
+size 7736

48_128_e5_3e-5/checkpoint-1470/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)

48_128_e5_3e-5/checkpoint-294/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

48_128_e5_3e-5/checkpoint-294/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": [
+    "o_proj",
+    "down_proj",
+    "v_proj",
+    "gate_proj",
+    "k_proj",
+    "q_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

48_128_e5_3e-5/checkpoint-294/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b915da442d3d910d514df07282e5ef0429cfc4e144a4efb896689949709ada5
+size 791751704

48_128_e5_3e-5/checkpoint-294/latest

@@ -0,0 +1 @@
+global_step294

48_128_e5_3e-5/checkpoint-294/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:611f178fdba89369762d66481b31ee9b6117fcaace663dce37935eeec47b047f
+size 15920

48_128_e5_3e-5/checkpoint-294/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:492fd222ab9e5cdbfe7def5828e427213d70e7febc107aaca6ee7fbc8c34db74
+size 15920

48_128_e5_3e-5/checkpoint-294/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:431dd49b17bbf40109934f63f3ebe0a6a681fbd56239205b4081e8e7efdcde01
+size 15920