Upload folder using huggingface_hub

38_128_e5_3e-5/checkpoint-1347/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

38_128_e5_3e-5/checkpoint-1347/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": [
+    "q_proj",
+    "v_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "k_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

38_128_e5_3e-5/checkpoint-1347/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f750a199d892efc6934a760dea2723a39e3aa0c02b3a1627abcc2f0da73867d
+size 791751704

38_128_e5_3e-5/checkpoint-1347/latest

@@ -0,0 +1 @@
+global_step1347

38_128_e5_3e-5/checkpoint-1347/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72d3e78c9e6b443720ec75ba59bd28e4ffe5574d07b7623d2275cb0b91f54fff
+size 15920

38_128_e5_3e-5/checkpoint-1347/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f55324196474861a436491a25deb8b2bf4acd65bbb580ac4f88a667307409662
+size 15920

38_128_e5_3e-5/checkpoint-1347/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:294ca3e9b5cf15560c15734a048cfbdb6eea8006a8a45321de8c5131b0eab3aa
+size 15920

38_128_e5_3e-5/checkpoint-1347/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd0d9d30eb728c7cc0cfa18c96421637a22b75d5ecc02c8675ef8ee9c66183c2
+size 15920

38_128_e5_3e-5/checkpoint-1347/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:391ffce156eab8f99bdcc44140c482456fc942ab6f19515470aa653347c57bcd
+size 15920

38_128_e5_3e-5/checkpoint-1347/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d22f3f0d6d7a7138c5a6ce4f9cacf9f2f6a7b65b7db980863270e2e970e9e685
+size 15920

38_128_e5_3e-5/checkpoint-1347/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:44769e2c10844dd06c6b4ed60ca1fedcffd29e68f6d6cd6d00c52c05b7fdd265
+size 15920

38_128_e5_3e-5/checkpoint-1347/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba5ef1dafe9ba29b6740d31f34d68c7b5828db1a1b673cadfbbf920c4e835160
+size 15920

38_128_e5_3e-5/checkpoint-1347/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:66bc449c7075f5965deae03dbd4305353959dc64a48282954e8cdf5bc078b019
+size 1064

38_128_e5_3e-5/checkpoint-1347/special_tokens_map.json

@@ -0,0 +1,45 @@
+{
+  "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": "<reponame>",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

38_128_e5_3e-5/checkpoint-1347/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": "<reponame>",
+  "padding_side": "left",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>",
+  "vocab_size": 49152
+}

38_128_e5_3e-5/checkpoint-1347/trainer_state.json

@@ -0,0 +1,1917 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1347,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011135857461024499,
+      "grad_norm": 1.3413572311401367,
+      "learning_rate": 1.0619469026548673e-06,
+      "loss": 1.3355,
+      "step": 5
+    },
+    {
+      "epoch": 0.022271714922048998,
+      "grad_norm": 1.0000957250595093,
+      "learning_rate": 2.389380530973451e-06,
+      "loss": 1.4079,
+      "step": 10
+    },
+    {
+      "epoch": 0.0334075723830735,
+      "grad_norm": 0.7801820039749146,
+      "learning_rate": 3.7168141592920353e-06,
+      "loss": 1.3474,
+      "step": 15
+    },
+    {
+      "epoch": 0.044543429844097995,
+      "grad_norm": 0.5836945176124573,
+      "learning_rate": 5.04424778761062e-06,
+      "loss": 1.3387,
+      "step": 20
+    },
+    {
+      "epoch": 0.0556792873051225,
+      "grad_norm": 0.6259394288063049,
+      "learning_rate": 6.371681415929204e-06,
+      "loss": 1.3457,
+      "step": 25
+    },
+    {
+      "epoch": 0.066815144766147,
+      "grad_norm": 0.6279721856117249,
+      "learning_rate": 7.699115044247788e-06,
+      "loss": 1.345,
+      "step": 30
+    },
+    {
+      "epoch": 0.0779510022271715,
+      "grad_norm": 0.5959689021110535,
+      "learning_rate": 9.026548672566373e-06,
+      "loss": 1.3064,
+      "step": 35
+    },
+    {
+      "epoch": 0.08908685968819599,
+      "grad_norm": 0.4760344624519348,
+      "learning_rate": 1.0353982300884955e-05,
+      "loss": 1.1918,
+      "step": 40
+    },
+    {
+      "epoch": 0.10022271714922049,
+      "grad_norm": 0.5409045219421387,
+      "learning_rate": 1.168141592920354e-05,
+      "loss": 1.2243,
+      "step": 45
+    },
+    {
+      "epoch": 0.111358574610245,
+      "grad_norm": 0.6069848537445068,
+      "learning_rate": 1.3008849557522125e-05,
+      "loss": 1.2276,
+      "step": 50
+    },
+    {
+      "epoch": 0.12249443207126949,
+      "grad_norm": 0.5506272315979004,
+      "learning_rate": 1.433628318584071e-05,
+      "loss": 1.2671,
+      "step": 55
+    },
+    {
+      "epoch": 0.133630289532294,
+      "grad_norm": 0.4604385793209076,
+      "learning_rate": 1.5663716814159295e-05,
+      "loss": 1.2268,
+      "step": 60
+    },
+    {
+      "epoch": 0.1447661469933185,
+      "grad_norm": 0.5579900741577148,
+      "learning_rate": 1.6991150442477876e-05,
+      "loss": 1.2215,
+      "step": 65
+    },
+    {
+      "epoch": 0.155902004454343,
+      "grad_norm": 0.48294851183891296,
+      "learning_rate": 1.8318584070796458e-05,
+      "loss": 1.2418,
+      "step": 70
+    },
+    {
+      "epoch": 0.16703786191536749,
+      "grad_norm": 0.5626055598258972,
+      "learning_rate": 1.9646017699115046e-05,
+      "loss": 1.185,
+      "step": 75
+    },
+    {
+      "epoch": 0.17817371937639198,
+      "grad_norm": 0.5725001096725464,
+      "learning_rate": 2.0973451327433628e-05,
+      "loss": 1.2337,
+      "step": 80
+    },
+    {
+      "epoch": 0.18930957683741648,
+      "grad_norm": 0.5308484435081482,
+      "learning_rate": 2.2300884955752213e-05,
+      "loss": 1.164,
+      "step": 85
+    },
+    {
+      "epoch": 0.20044543429844097,
+      "grad_norm": 0.5047495365142822,
+      "learning_rate": 2.3628318584070798e-05,
+      "loss": 1.1862,
+      "step": 90
+    },
+    {
+      "epoch": 0.21158129175946547,
+      "grad_norm": 0.5524989366531372,
+      "learning_rate": 2.495575221238938e-05,
+      "loss": 1.1503,
+      "step": 95
+    },
+    {
+      "epoch": 0.22271714922049,
+      "grad_norm": 0.5751224160194397,
+      "learning_rate": 2.6283185840707968e-05,
+      "loss": 1.1819,
+      "step": 100
+    },
+    {
+      "epoch": 0.23385300668151449,
+      "grad_norm": 0.590943455696106,
+      "learning_rate": 2.761061946902655e-05,
+      "loss": 1.1177,
+      "step": 105
+    },
+    {
+      "epoch": 0.24498886414253898,
+      "grad_norm": 0.624424159526825,
+      "learning_rate": 2.893805309734513e-05,
+      "loss": 1.1612,
+      "step": 110
+    },
+    {
+      "epoch": 0.2561247216035635,
+      "grad_norm": 0.7900341153144836,
+      "learning_rate": 2.9999983715046248e-05,
+      "loss": 1.1556,
+      "step": 115
+    },
+    {
+      "epoch": 0.267260579064588,
+      "grad_norm": 0.717691957950592,
+      "learning_rate": 2.9999413745377787e-05,
+      "loss": 1.1074,
+      "step": 120
+    },
+    {
+      "epoch": 0.27839643652561247,
+      "grad_norm": 0.5590342283248901,
+      "learning_rate": 2.9998029563381436e-05,
+      "loss": 1.1106,
+      "step": 125
+    },
+    {
+      "epoch": 0.289532293986637,
+      "grad_norm": 0.5695084929466248,
+      "learning_rate": 2.999583124419468e-05,
+      "loss": 1.0615,
+      "step": 130
+    },
+    {
+      "epoch": 0.30066815144766146,
+      "grad_norm": 0.6339223384857178,
+      "learning_rate": 2.999281890714874e-05,
+      "loss": 1.1041,
+      "step": 135
+    },
+    {
+      "epoch": 0.311804008908686,
+      "grad_norm": 0.6401590704917908,
+      "learning_rate": 2.9988992715762148e-05,
+      "loss": 1.0891,
+      "step": 140
+    },
+    {
+      "epoch": 0.32293986636971045,
+      "grad_norm": 0.6139689087867737,
+      "learning_rate": 2.9984352877731836e-05,
+      "loss": 1.0935,
+      "step": 145
+    },
+    {
+      "epoch": 0.33407572383073497,
+      "grad_norm": 0.5202445387840271,
+      "learning_rate": 2.9978899644921866e-05,
+      "loss": 1.0622,
+      "step": 150
+    },
+    {
+      "epoch": 0.34521158129175944,
+      "grad_norm": 0.5589560866355896,
+      "learning_rate": 2.9972633313349764e-05,
+      "loss": 1.0604,
+      "step": 155
+    },
+    {
+      "epoch": 0.35634743875278396,
+      "grad_norm": 0.6837779879570007,
+      "learning_rate": 2.9965554223170455e-05,
+      "loss": 1.008,
+      "step": 160
+    },
+    {
+      "epoch": 0.3674832962138085,
+      "grad_norm": 0.7174135446548462,
+      "learning_rate": 2.9957662758657787e-05,
+      "loss": 1.0292,
+      "step": 165
+    },
+    {
+      "epoch": 0.37861915367483295,
+      "grad_norm": 0.6325995922088623,
+      "learning_rate": 2.9948959348183686e-05,
+      "loss": 1.0019,
+      "step": 170
+    },
+    {
+      "epoch": 0.3897550111358575,
+      "grad_norm": 0.6495500802993774,
+      "learning_rate": 2.993944446419489e-05,
+      "loss": 0.9838,
+      "step": 175
+    },
+    {
+      "epoch": 0.40089086859688194,
+      "grad_norm": 0.641270101070404,
+      "learning_rate": 2.9929118623187307e-05,
+      "loss": 0.9553,
+      "step": 180
+    },
+    {
+      "epoch": 0.41202672605790647,
+      "grad_norm": 0.6926126480102539,
+      "learning_rate": 2.9917982385677982e-05,
+      "loss": 1.0296,
+      "step": 185
+    },
+    {
+      "epoch": 0.42316258351893093,
+      "grad_norm": 0.7776742577552795,
+      "learning_rate": 2.990603635617466e-05,
+      "loss": 1.0223,
+      "step": 190
+    },
+    {
+      "epoch": 0.43429844097995546,
+      "grad_norm": 0.6762447357177734,
+      "learning_rate": 2.9893281183142996e-05,
+      "loss": 1.053,
+      "step": 195
+    },
+    {
+      "epoch": 0.44543429844098,
+      "grad_norm": 0.70978844165802,
+      "learning_rate": 2.9879717558971317e-05,
+      "loss": 0.9561,
+      "step": 200
+    },
+    {
+      "epoch": 0.45657015590200445,
+      "grad_norm": 0.7382838726043701,
+      "learning_rate": 2.986534621993307e-05,
+      "loss": 0.9923,
+      "step": 205
+    },
+    {
+      "epoch": 0.46770601336302897,
+      "grad_norm": 0.7762662768363953,
+      "learning_rate": 2.9850167946146838e-05,
+      "loss": 1.0125,
+      "step": 210
+    },
+    {
+      "epoch": 0.47884187082405344,
+      "grad_norm": 0.7446727752685547,
+      "learning_rate": 2.9834183561533997e-05,
+      "loss": 0.9445,
+      "step": 215
+    },
+    {
+      "epoch": 0.48997772828507796,
+      "grad_norm": 0.78370201587677,
+      "learning_rate": 2.9817393933773994e-05,
+      "loss": 0.9717,
+      "step": 220
+    },
+    {
+      "epoch": 0.5011135857461024,
+      "grad_norm": 0.6878297328948975,
+      "learning_rate": 2.9799799974257248e-05,
+      "loss": 0.9373,
+      "step": 225
+    },
+    {
+      "epoch": 0.512249443207127,
+      "grad_norm": 0.690732479095459,
+      "learning_rate": 2.9781402638035656e-05,
+      "loss": 0.9109,
+      "step": 230
+    },
+    {
+      "epoch": 0.5233853006681515,
+      "grad_norm": 0.8460608124732971,
+      "learning_rate": 2.9762202923770795e-05,
+      "loss": 0.9548,
+      "step": 235
+    },
+    {
+      "epoch": 0.534521158129176,
+      "grad_norm": 0.8983225226402283,
+      "learning_rate": 2.9742201873679656e-05,
+      "loss": 0.9547,
+      "step": 240
+    },
+    {
+      "epoch": 0.5456570155902004,
+      "grad_norm": 0.8119720816612244,
+      "learning_rate": 2.9721400573478117e-05,
+      "loss": 0.9748,
+      "step": 245
+    },
+    {
+      "epoch": 0.5567928730512249,
+      "grad_norm": 0.8376508951187134,
+      "learning_rate": 2.9699800152321975e-05,
+      "loss": 0.8978,
+      "step": 250
+    },
+    {
+      "epoch": 0.5679287305122495,
+      "grad_norm": 0.7981753349304199,
+      "learning_rate": 2.967740178274567e-05,
+      "loss": 0.8872,
+      "step": 255
+    },
+    {
+      "epoch": 0.579064587973274,
+      "grad_norm": 0.7605271339416504,
+      "learning_rate": 2.9654206680598638e-05,
+      "loss": 0.8954,
+      "step": 260
+    },
+    {
+      "epoch": 0.5902004454342984,
+      "grad_norm": 0.9272874593734741,
+      "learning_rate": 2.9630216104979288e-05,
+      "loss": 0.8797,
+      "step": 265
+    },
+    {
+      "epoch": 0.6013363028953229,
+      "grad_norm": 0.8306118249893188,
+      "learning_rate": 2.9605431358166687e-05,
+      "loss": 0.9045,
+      "step": 270
+    },
+    {
+      "epoch": 0.6124721603563474,
+      "grad_norm": 0.849515438079834,
+      "learning_rate": 2.957985378554984e-05,
+      "loss": 0.8929,
+      "step": 275
+    },
+    {
+      "epoch": 0.623608017817372,
+      "grad_norm": 0.9004138708114624,
+      "learning_rate": 2.955348477555467e-05,
+      "loss": 0.9355,
+      "step": 280
+    },
+    {
+      "epoch": 0.6347438752783965,
+      "grad_norm": 0.836115837097168,
+      "learning_rate": 2.9526325759568665e-05,
+      "loss": 0.8795,
+      "step": 285
+    },
+    {
+      "epoch": 0.6458797327394209,
+      "grad_norm": 0.8236976861953735,
+      "learning_rate": 2.9498378211863145e-05,
+      "loss": 0.8241,
+      "step": 290
+    },
+    {
+      "epoch": 0.6570155902004454,
+      "grad_norm": 0.817470908164978,
+      "learning_rate": 2.9469643649513264e-05,
+      "loss": 0.8541,
+      "step": 295
+    },
+    {
+      "epoch": 0.6681514476614699,
+      "grad_norm": 0.9785251021385193,
+      "learning_rate": 2.9440123632315647e-05,
+      "loss": 0.8715,
+      "step": 300
+    },
+    {
+      "epoch": 0.6792873051224945,
+      "grad_norm": 0.9281602501869202,
+      "learning_rate": 2.9409819762703715e-05,
+      "loss": 0.7914,
+      "step": 305
+    },
+    {
+      "epoch": 0.6904231625835189,
+      "grad_norm": 0.8825387358665466,
+      "learning_rate": 2.9378733685660707e-05,
+      "loss": 0.8527,
+      "step": 310
+    },
+    {
+      "epoch": 0.7015590200445434,
+      "grad_norm": 0.9314531683921814,
+      "learning_rate": 2.934686708863039e-05,
+      "loss": 0.8966,
+      "step": 315
+    },
+    {
+      "epoch": 0.7126948775055679,
+      "grad_norm": 0.9172025322914124,
+      "learning_rate": 2.9314221701425445e-05,
+      "loss": 0.8472,
+      "step": 320
+    },
+    {
+      "epoch": 0.7238307349665924,
+      "grad_norm": 0.9198710918426514,
+      "learning_rate": 2.9280799296133577e-05,
+      "loss": 0.878,
+      "step": 325
+    },
+    {
+      "epoch": 0.734966592427617,
+      "grad_norm": 0.8762511610984802,
+      "learning_rate": 2.9246601687021324e-05,
+      "loss": 0.7845,
+      "step": 330
+    },
+    {
+      "epoch": 0.7461024498886414,
+      "grad_norm": 0.8782876133918762,
+      "learning_rate": 2.9211630730435564e-05,
+      "loss": 0.7928,
+      "step": 335
+    },
+    {
+      "epoch": 0.7572383073496659,
+      "grad_norm": 0.9156160950660706,
+      "learning_rate": 2.9175888324702762e-05,
+      "loss": 0.8467,
+      "step": 340
+    },
+    {
+      "epoch": 0.7683741648106904,
+      "grad_norm": 1.0404008626937866,
+      "learning_rate": 2.9139376410025897e-05,
+      "loss": 0.7927,
+      "step": 345
+    },
+    {
+      "epoch": 0.779510022271715,
+      "grad_norm": 1.0743643045425415,
+      "learning_rate": 2.9102096968379178e-05,
+      "loss": 0.8123,
+      "step": 350
+    },
+    {
+      "epoch": 0.7906458797327395,
+      "grad_norm": 0.9985504150390625,
+      "learning_rate": 2.9064052023400417e-05,
+      "loss": 0.7735,
+      "step": 355
+    },
+    {
+      "epoch": 0.8017817371937639,
+      "grad_norm": 0.9021556973457336,
+      "learning_rate": 2.9025243640281226e-05,
+      "loss": 0.8006,
+      "step": 360
+    },
+    {
+      "epoch": 0.8129175946547884,
+      "grad_norm": 0.9394658803939819,
+      "learning_rate": 2.8985673925654853e-05,
+      "loss": 0.7879,
+      "step": 365
+    },
+    {
+      "epoch": 0.8240534521158129,
+      "grad_norm": 0.88721764087677,
+      "learning_rate": 2.8945345027481884e-05,
+      "loss": 0.7225,
+      "step": 370
+    },
+    {
+      "epoch": 0.8351893095768375,
+      "grad_norm": 0.9523298740386963,
+      "learning_rate": 2.8904259134933627e-05,
+      "loss": 0.8202,
+      "step": 375
+    },
+    {
+      "epoch": 0.8463251670378619,
+      "grad_norm": 0.9752704501152039,
+      "learning_rate": 2.8862418478273246e-05,
+      "loss": 0.7871,
+      "step": 380
+    },
+    {
+      "epoch": 0.8574610244988864,
+      "grad_norm": 0.9238730072975159,
+      "learning_rate": 2.881982532873476e-05,
+      "loss": 0.7216,
+      "step": 385
+    },
+    {
+      "epoch": 0.8685968819599109,
+      "grad_norm": 0.8588601350784302,
+      "learning_rate": 2.8776481998399676e-05,
+      "loss": 0.7146,
+      "step": 390
+    },
+    {
+      "epoch": 0.8797327394209354,
+      "grad_norm": 0.8526139259338379,
+      "learning_rate": 2.873239084007156e-05,
+      "loss": 0.8105,
+      "step": 395
+    },
+    {
+      "epoch": 0.89086859688196,
+      "grad_norm": 1.0353056192398071,
+      "learning_rate": 2.868755424714825e-05,
+      "loss": 0.789,
+      "step": 400
+    },
+    {
+      "epoch": 0.9020044543429844,
+      "grad_norm": 0.8390752077102661,
+      "learning_rate": 2.8641974653491997e-05,
+      "loss": 0.7338,
+      "step": 405
+    },
+    {
+      "epoch": 0.9131403118040089,
+      "grad_norm": 0.9766253232955933,
+      "learning_rate": 2.85956545332973e-05,
+      "loss": 0.7782,
+      "step": 410
+    },
+    {
+      "epoch": 0.9242761692650334,
+      "grad_norm": 0.9131956100463867,
+      "learning_rate": 2.8548596400956614e-05,
+      "loss": 0.734,
+      "step": 415
+    },
+    {
+      "epoch": 0.9354120267260579,
+      "grad_norm": 0.979026734828949,
+      "learning_rate": 2.850080281092389e-05,
+      "loss": 0.7538,
+      "step": 420
+    },
+    {
+      "epoch": 0.9465478841870824,
+      "grad_norm": 1.0852129459381104,
+      "learning_rate": 2.845227635757587e-05,
+      "loss": 0.7072,
+      "step": 425
+    },
+    {
+      "epoch": 0.9576837416481069,
+      "grad_norm": 1.1135140657424927,
+      "learning_rate": 2.8403019675071267e-05,
+      "loss": 0.7657,
+      "step": 430
+    },
+    {
+      "epoch": 0.9688195991091314,
+      "grad_norm": 0.8528003096580505,
+      "learning_rate": 2.8353035437207796e-05,
+      "loss": 0.7036,
+      "step": 435
+    },
+    {
+      "epoch": 0.9799554565701559,
+      "grad_norm": 0.9384568929672241,
+      "learning_rate": 2.8302326357277013e-05,
+      "loss": 0.6948,
+      "step": 440
+    },
+    {
+      "epoch": 0.9910913140311804,
+      "grad_norm": 1.1752519607543945,
+      "learning_rate": 2.825089518791704e-05,
+      "loss": 0.7052,
+      "step": 445
+    },
+    {
+      "epoch": 1.0022271714922049,
+      "grad_norm": 1.038319706916809,
+      "learning_rate": 2.8198744720963127e-05,
+      "loss": 0.6852,
+      "step": 450
+    },
+    {
+      "epoch": 1.0133630289532294,
+      "grad_norm": 1.007336974143982,
+      "learning_rate": 2.814587778729613e-05,
+      "loss": 0.6137,
+      "step": 455
+    },
+    {
+      "epoch": 1.024498886414254,
+      "grad_norm": 1.089078426361084,
+      "learning_rate": 2.809229725668881e-05,
+      "loss": 0.6511,
+      "step": 460
+    },
+    {
+      "epoch": 1.0356347438752784,
+      "grad_norm": 0.9254465699195862,
+      "learning_rate": 2.803800603765008e-05,
+      "loss": 0.601,
+      "step": 465
+    },
+    {
+      "epoch": 1.046770601336303,
+      "grad_norm": 1.0865051746368408,
+      "learning_rate": 2.798300707726712e-05,
+      "loss": 0.6846,
+      "step": 470
+    },
+    {
+      "epoch": 1.0579064587973275,
+      "grad_norm": 0.9619767665863037,
+      "learning_rate": 2.7927303361045376e-05,
+      "loss": 0.659,
+      "step": 475
+    },
+    {
+      "epoch": 1.069042316258352,
+      "grad_norm": 1.1050430536270142,
+      "learning_rate": 2.7870897912746532e-05,
+      "loss": 0.5382,
+      "step": 480
+    },
+    {
+      "epoch": 1.0801781737193763,
+      "grad_norm": 1.060831069946289,
+      "learning_rate": 2.7813793794224348e-05,
+      "loss": 0.5903,
+      "step": 485
+    },
+    {
+      "epoch": 1.0913140311804008,
+      "grad_norm": 1.0150086879730225,
+      "learning_rate": 2.775599410525845e-05,
+      "loss": 0.577,
+      "step": 490
+    },
+    {
+      "epoch": 1.1024498886414253,
+      "grad_norm": 1.1610509157180786,
+      "learning_rate": 2.769750198338609e-05,
+      "loss": 0.5755,
+      "step": 495
+    },
+    {
+      "epoch": 1.1135857461024499,
+      "grad_norm": 1.022871494293213,
+      "learning_rate": 2.7638320603731817e-05,
+      "loss": 0.5687,
+      "step": 500
+    },
+    {
+      "epoch": 1.1247216035634744,
+      "grad_norm": 1.0842667818069458,
+      "learning_rate": 2.7578453178835108e-05,
+      "loss": 0.5426,
+      "step": 505
+    },
+    {
+      "epoch": 1.135857461024499,
+      "grad_norm": 1.1093308925628662,
+      "learning_rate": 2.7517902958476003e-05,
+      "loss": 0.5733,
+      "step": 510
+    },
+    {
+      "epoch": 1.1469933184855234,
+      "grad_norm": 1.0264402627944946,
+      "learning_rate": 2.745667322949869e-05,
+      "loss": 0.5749,
+      "step": 515
+    },
+    {
+      "epoch": 1.158129175946548,
+      "grad_norm": 1.2180768251419067,
+      "learning_rate": 2.7394767315633086e-05,
+      "loss": 0.5767,
+      "step": 520
+    },
+    {
+      "epoch": 1.1692650334075725,
+      "grad_norm": 1.0007996559143066,
+      "learning_rate": 2.7332188577314407e-05,
+      "loss": 0.5277,
+      "step": 525
+    },
+    {
+      "epoch": 1.1804008908685968,
+      "grad_norm": 0.9211612343788147,
+      "learning_rate": 2.7268940411500768e-05,
+      "loss": 0.6127,
+      "step": 530
+    },
+    {
+      "epoch": 1.1915367483296213,
+      "grad_norm": 1.1937363147735596,
+      "learning_rate": 2.720502625148877e-05,
+      "loss": 0.5765,
+      "step": 535
+    },
+    {
+      "epoch": 1.2026726057906458,
+      "grad_norm": 1.041098713874817,
+      "learning_rate": 2.7140449566727156e-05,
+      "loss": 0.5372,
+      "step": 540
+    },
+    {
+      "epoch": 1.2138084632516704,
+      "grad_norm": 1.0834497213363647,
+      "learning_rate": 2.707521386262845e-05,
+      "loss": 0.6064,
+      "step": 545
+    },
+    {
+      "epoch": 1.2249443207126949,
+      "grad_norm": 1.024735450744629,
+      "learning_rate": 2.700932268037868e-05,
+      "loss": 0.5451,
+      "step": 550
+    },
+    {
+      "epoch": 1.2360801781737194,
+      "grad_norm": 1.0640431642532349,
+      "learning_rate": 2.6942779596745172e-05,
+      "loss": 0.535,
+      "step": 555
+    },
+    {
+      "epoch": 1.247216035634744,
+      "grad_norm": 1.0001004934310913,
+      "learning_rate": 2.6875588223882363e-05,
+      "loss": 0.5336,
+      "step": 560
+    },
+    {
+      "epoch": 1.2583518930957684,
+      "grad_norm": 1.328310251235962,
+      "learning_rate": 2.680775220913575e-05,
+      "loss": 0.5681,
+      "step": 565
+    },
+    {
+      "epoch": 1.269487750556793,
+      "grad_norm": 1.0832302570343018,
+      "learning_rate": 2.6739275234843894e-05,
+      "loss": 0.6001,
+      "step": 570
+    },
+    {
+      "epoch": 1.2806236080178173,
+      "grad_norm": 1.1307528018951416,
+      "learning_rate": 2.66701610181385e-05,
+      "loss": 0.5613,
+      "step": 575
+    },
+    {
+      "epoch": 1.2917594654788418,
+      "grad_norm": 1.1129206418991089,
+      "learning_rate": 2.6600413310742696e-05,
+      "loss": 0.5266,
+      "step": 580
+    },
+    {
+      "epoch": 1.3028953229398663,
+      "grad_norm": 1.110201358795166,
+      "learning_rate": 2.653003589876733e-05,
+      "loss": 0.5429,
+      "step": 585
+    },
+    {
+      "epoch": 1.3140311804008908,
+      "grad_norm": 1.1400443315505981,
+      "learning_rate": 2.645903260250548e-05,
+      "loss": 0.5909,
+      "step": 590
+    },
+    {
+      "epoch": 1.3251670378619154,
+      "grad_norm": 1.065966248512268,
+      "learning_rate": 2.6387407276225057e-05,
+      "loss": 0.5312,
+      "step": 595
+    },
+    {
+      "epoch": 1.3363028953229399,
+      "grad_norm": 1.266574501991272,
+      "learning_rate": 2.6315163807959603e-05,
+      "loss": 0.5414,
+      "step": 600
+    },
+    {
+      "epoch": 1.3474387527839644,
+      "grad_norm": 0.9977100491523743,
+      "learning_rate": 2.6242306119297205e-05,
+      "loss": 0.5039,
+      "step": 605
+    },
+    {
+      "epoch": 1.358574610244989,
+      "grad_norm": 1.1061787605285645,
+      "learning_rate": 2.6168838165167674e-05,
+      "loss": 0.4875,
+      "step": 610
+    },
+    {
+      "epoch": 1.3697104677060135,
+      "grad_norm": 1.107053279876709,
+      "learning_rate": 2.6094763933627794e-05,
+      "loss": 0.5168,
+      "step": 615
+    },
+    {
+      "epoch": 1.3808463251670378,
+      "grad_norm": 1.0395705699920654,
+      "learning_rate": 2.602008744564489e-05,
+      "loss": 0.482,
+      "step": 620
+    },
+    {
+      "epoch": 1.3919821826280623,
+      "grad_norm": 0.9968161582946777,
+      "learning_rate": 2.594481275487854e-05,
+      "loss": 0.5275,
+      "step": 625
+    },
+    {
+      "epoch": 1.4031180400890868,
+      "grad_norm": 1.1658473014831543,
+      "learning_rate": 2.5868943947460514e-05,
+      "loss": 0.5031,
+      "step": 630
+    },
+    {
+      "epoch": 1.4142538975501113,
+      "grad_norm": 1.0978684425354004,
+      "learning_rate": 2.5792485141773007e-05,
+      "loss": 0.5183,
+      "step": 635
+    },
+    {
+      "epoch": 1.4253897550111359,
+      "grad_norm": 1.0857203006744385,
+      "learning_rate": 2.5715440488225028e-05,
+      "loss": 0.4841,
+      "step": 640
+    },
+    {
+      "epoch": 1.4365256124721604,
+      "grad_norm": 1.1584542989730835,
+      "learning_rate": 2.5637814169027158e-05,
+      "loss": 0.4944,
+      "step": 645
+    },
+    {
+      "epoch": 1.447661469933185,
+      "grad_norm": 1.1438153982162476,
+      "learning_rate": 2.555961039796449e-05,
+      "loss": 0.4495,
+      "step": 650
+    },
+    {
+      "epoch": 1.4587973273942094,
+      "grad_norm": 1.2044782638549805,
+      "learning_rate": 2.5480833420167904e-05,
+      "loss": 0.4588,
+      "step": 655
+    },
+    {
+      "epoch": 1.469933184855234,
+      "grad_norm": 1.299019694328308,
+      "learning_rate": 2.5401487511883628e-05,
+      "loss": 0.5053,
+      "step": 660
+    },
+    {
+      "epoch": 1.4810690423162582,
+      "grad_norm": 0.9981383681297302,
+      "learning_rate": 2.5321576980241115e-05,
+      "loss": 0.525,
+      "step": 665
+    },
+    {
+      "epoch": 1.492204899777283,
+      "grad_norm": 1.142288088798523,
+      "learning_rate": 2.5241106163019227e-05,
+      "loss": 0.5095,
+      "step": 670
+    },
+    {
+      "epoch": 1.5033407572383073,
+      "grad_norm": 1.0118094682693481,
+      "learning_rate": 2.5160079428410795e-05,
+      "loss": 0.4554,
+      "step": 675
+    },
+    {
+      "epoch": 1.5144766146993318,
+      "grad_norm": 1.1396214962005615,
+      "learning_rate": 2.507850117478547e-05,
+      "loss": 0.4246,
+      "step": 680
+    },
+    {
+      "epoch": 1.5256124721603563,
+      "grad_norm": 0.9885154962539673,
+      "learning_rate": 2.4996375830450964e-05,
+      "loss": 0.4702,
+      "step": 685
+    },
+    {
+      "epoch": 1.5367483296213809,
+      "grad_norm": 1.0633879899978638,
+      "learning_rate": 2.4913707853412712e-05,
+      "loss": 0.4636,
+      "step": 690
+    },
+    {
+      "epoch": 1.5478841870824054,
+      "grad_norm": 1.0840256214141846,
+      "learning_rate": 2.483050173113183e-05,
+      "loss": 0.4692,
+      "step": 695
+    },
+    {
+      "epoch": 1.5590200445434297,
+      "grad_norm": 1.0847933292388916,
+      "learning_rate": 2.4746761980281552e-05,
+      "loss": 0.4398,
+      "step": 700
+    },
+    {
+      "epoch": 1.5701559020044544,
+      "grad_norm": 1.1708725690841675,
+      "learning_rate": 2.466249314650203e-05,
+      "loss": 0.447,
+      "step": 705
+    },
+    {
+      "epoch": 1.5812917594654787,
+      "grad_norm": 1.2069448232650757,
+      "learning_rate": 2.4577699804153603e-05,
+      "loss": 0.4652,
+      "step": 710
+    },
+    {
+      "epoch": 1.5924276169265035,
+      "grad_norm": 1.042110562324524,
+      "learning_rate": 2.4492386556068478e-05,
+      "loss": 0.4583,
+      "step": 715
+    },
+    {
+      "epoch": 1.6035634743875278,
+      "grad_norm": 1.1450542211532593,
+      "learning_rate": 2.440655803330088e-05,
+      "loss": 0.4379,
+      "step": 720
+    },
+    {
+      "epoch": 1.6146993318485523,
+      "grad_norm": 1.1880165338516235,
+      "learning_rate": 2.432021889487565e-05,
+      "loss": 0.4419,
+      "step": 725
+    },
+    {
+      "epoch": 1.6258351893095768,
+      "grad_norm": 1.0804628133773804,
+      "learning_rate": 2.4233373827535354e-05,
+      "loss": 0.4545,
+      "step": 730
+    },
+    {
+      "epoch": 1.6369710467706013,
+      "grad_norm": 1.0696831941604614,
+      "learning_rate": 2.414602754548587e-05,
+      "loss": 0.4793,
+      "step": 735
+    },
+    {
+      "epoch": 1.6481069042316259,
+      "grad_norm": 1.0015273094177246,
+      "learning_rate": 2.405818479014049e-05,
+      "loss": 0.4868,
+      "step": 740
+    },
+    {
+      "epoch": 1.6592427616926502,
+      "grad_norm": 1.0753366947174072,
+      "learning_rate": 2.3969850329862533e-05,
+      "loss": 0.4124,
+      "step": 745
+    },
+    {
+      "epoch": 1.670378619153675,
+      "grad_norm": 1.079838514328003,
+      "learning_rate": 2.3881028959706507e-05,
+      "loss": 0.4305,
+      "step": 750
+    },
+    {
+      "epoch": 1.6815144766146992,
+      "grad_norm": 1.1108442544937134,
+      "learning_rate": 2.3791725501157824e-05,
+      "loss": 0.4135,
+      "step": 755
+    },
+    {
+      "epoch": 1.692650334075724,
+      "grad_norm": 1.0079882144927979,
+      "learning_rate": 2.3701944801871073e-05,
+      "loss": 0.4424,
+      "step": 760
+    },
+    {
+      "epoch": 1.7037861915367483,
+      "grad_norm": 1.353743553161621,
+      "learning_rate": 2.3611691735406878e-05,
+      "loss": 0.4547,
+      "step": 765
+    },
+    {
+      "epoch": 1.7149220489977728,
+      "grad_norm": 1.1483458280563354,
+      "learning_rate": 2.3520971200967337e-05,
+      "loss": 0.4032,
+      "step": 770
+    },
+    {
+      "epoch": 1.7260579064587973,
+      "grad_norm": 1.1146272420883179,
+      "learning_rate": 2.342978812313009e-05,
+      "loss": 0.416,
+      "step": 775
+    },
+    {
+      "epoch": 1.7371937639198218,
+      "grad_norm": 1.1441491842269897,
+      "learning_rate": 2.3338147451580997e-05,
+      "loss": 0.4376,
+      "step": 780
+    },
+    {
+      "epoch": 1.7483296213808464,
+      "grad_norm": 1.0915056467056274,
+      "learning_rate": 2.324605416084544e-05,
+      "loss": 0.4071,
+      "step": 785
+    },
+    {
+      "epoch": 1.7594654788418709,
+      "grad_norm": 1.3740445375442505,
+      "learning_rate": 2.315351325001832e-05,
+      "loss": 0.422,
+      "step": 790
+    },
+    {
+      "epoch": 1.7706013363028954,
+      "grad_norm": 1.111710786819458,
+      "learning_rate": 2.3060529742492656e-05,
+      "loss": 0.3774,
+      "step": 795
+    },
+    {
+      "epoch": 1.7817371937639197,
+      "grad_norm": 1.1952409744262695,
+      "learning_rate": 2.2967108685686934e-05,
+      "loss": 0.4169,
+      "step": 800
+    },
+    {
+      "epoch": 1.7928730512249444,
+      "grad_norm": 1.2465108633041382,
+      "learning_rate": 2.2873255150771096e-05,
+      "loss": 0.3993,
+      "step": 805
+    },
+    {
+      "epoch": 1.8040089086859687,
+      "grad_norm": 1.1177217960357666,
+      "learning_rate": 2.277897423239126e-05,
+      "loss": 0.4199,
+      "step": 810
+    },
+    {
+      "epoch": 1.8151447661469933,
+      "grad_norm": 1.0628764629364014,
+      "learning_rate": 2.2684271048393194e-05,
+      "loss": 0.388,
+      "step": 815
+    },
+    {
+      "epoch": 1.8262806236080178,
+      "grad_norm": 1.1902976036071777,
+      "learning_rate": 2.2589150739544477e-05,
+      "loss": 0.3985,
+      "step": 820
+    },
+    {
+      "epoch": 1.8374164810690423,
+      "grad_norm": 1.1216514110565186,
+      "learning_rate": 2.2493618469255453e-05,
+      "loss": 0.379,
+      "step": 825
+    },
+    {
+      "epoch": 1.8485523385300668,
+      "grad_norm": 1.1403833627700806,
+      "learning_rate": 2.2397679423298945e-05,
+      "loss": 0.4247,
+      "step": 830
+    },
+    {
+      "epoch": 1.8596881959910914,
+      "grad_norm": 1.1960326433181763,
+      "learning_rate": 2.230133880952876e-05,
+      "loss": 0.3967,
+      "step": 835
+    },
+    {
+      "epoch": 1.8708240534521159,
+      "grad_norm": 1.2178454399108887,
+      "learning_rate": 2.2204601857596987e-05,
+      "loss": 0.4054,
+      "step": 840
+    },
+    {
+      "epoch": 1.8819599109131402,
+      "grad_norm": 1.148664116859436,
+      "learning_rate": 2.2107473818670107e-05,
+      "loss": 0.3996,
+      "step": 845
+    },
+    {
+      "epoch": 1.893095768374165,
+      "grad_norm": 1.2493960857391357,
+      "learning_rate": 2.2009959965143972e-05,
+      "loss": 0.4157,
+      "step": 850
+    },
+    {
+      "epoch": 1.9042316258351892,
+      "grad_norm": 1.1501412391662598,
+      "learning_rate": 2.1912065590357574e-05,
+      "loss": 0.4131,
+      "step": 855
+    },
+    {
+      "epoch": 1.9153674832962138,
+      "grad_norm": 1.0085930824279785,
+      "learning_rate": 2.1813796008305724e-05,
+      "loss": 0.4247,
+      "step": 860
+    },
+    {
+      "epoch": 1.9265033407572383,
+      "grad_norm": 1.2748278379440308,
+      "learning_rate": 2.1715156553350576e-05,
+      "loss": 0.4016,
+      "step": 865
+    },
+    {
+      "epoch": 1.9376391982182628,
+      "grad_norm": 1.1798368692398071,
+      "learning_rate": 2.1616152579932095e-05,
+      "loss": 0.3665,
+      "step": 870
+    },
+    {
+      "epoch": 1.9487750556792873,
+      "grad_norm": 1.21760892868042,
+      "learning_rate": 2.1516789462277372e-05,
+      "loss": 0.367,
+      "step": 875
+    },
+    {
+      "epoch": 1.9599109131403119,
+      "grad_norm": 1.129085898399353,
+      "learning_rate": 2.1417072594108894e-05,
+      "loss": 0.3782,
+      "step": 880
+    },
+    {
+      "epoch": 1.9710467706013364,
+      "grad_norm": 1.136759877204895,
+      "learning_rate": 2.1317007388351774e-05,
+      "loss": 0.404,
+      "step": 885
+    },
+    {
+      "epoch": 1.9821826280623607,
+      "grad_norm": 1.1734161376953125,
+      "learning_rate": 2.1216599276839922e-05,
+      "loss": 0.3595,
+      "step": 890
+    },
+    {
+      "epoch": 1.9933184855233854,
+      "grad_norm": 1.1954755783081055,
+      "learning_rate": 2.1115853710021174e-05,
+      "loss": 0.3842,
+      "step": 895
+    },
+    {
+      "epoch": 2.0044543429844097,
+      "grad_norm": 1.0263651609420776,
+      "learning_rate": 2.101477615666143e-05,
+      "loss": 0.3263,
+      "step": 900
+    },
+    {
+      "epoch": 2.0155902004454345,
+      "grad_norm": 1.2437479496002197,
+      "learning_rate": 2.0913372103547805e-05,
+      "loss": 0.3192,
+      "step": 905
+    },
+    {
+      "epoch": 2.0267260579064588,
+      "grad_norm": 1.1094865798950195,
+      "learning_rate": 2.081164705519078e-05,
+      "loss": 0.3165,
+      "step": 910
+    },
+    {
+      "epoch": 2.0378619153674835,
+      "grad_norm": 1.2660256624221802,
+      "learning_rate": 2.07096065335254e-05,
+      "loss": 0.2691,
+      "step": 915
+    },
+    {
+      "epoch": 2.048997772828508,
+      "grad_norm": 1.1449397802352905,
+      "learning_rate": 2.0607256077611533e-05,
+      "loss": 0.2808,
+      "step": 920
+    },
+    {
+      "epoch": 2.060133630289532,
+      "grad_norm": 1.2465488910675049,
+      "learning_rate": 2.0504601243333182e-05,
+      "loss": 0.2933,
+      "step": 925
+    },
+    {
+      "epoch": 2.071269487750557,
+      "grad_norm": 1.0938855409622192,
+      "learning_rate": 2.040164760309691e-05,
+      "loss": 0.2703,
+      "step": 930
+    },
+    {
+      "epoch": 2.082405345211581,
+      "grad_norm": 1.1405394077301025,
+      "learning_rate": 2.0298400745529352e-05,
+      "loss": 0.2926,
+      "step": 935
+    },
+    {
+      "epoch": 2.093541202672606,
+      "grad_norm": 1.2811548709869385,
+      "learning_rate": 2.0194866275173827e-05,
+      "loss": 0.2965,
+      "step": 940
+    },
+    {
+      "epoch": 2.10467706013363,
+      "grad_norm": 1.2539308071136475,
+      "learning_rate": 2.0091049812186146e-05,
+      "loss": 0.3216,
+      "step": 945
+    },
+    {
+      "epoch": 2.115812917594655,
+      "grad_norm": 1.1019704341888428,
+      "learning_rate": 1.9986956992029487e-05,
+      "loss": 0.2967,
+      "step": 950
+    },
+    {
+      "epoch": 2.1269487750556793,
+      "grad_norm": 1.0621236562728882,
+      "learning_rate": 1.9882593465168522e-05,
+      "loss": 0.2511,
+      "step": 955
+    },
+    {
+      "epoch": 2.138084632516704,
+      "grad_norm": 1.1347256898880005,
+      "learning_rate": 1.9777964896762676e-05,
+      "loss": 0.3121,
+      "step": 960
+    },
+    {
+      "epoch": 2.1492204899777283,
+      "grad_norm": 1.0914628505706787,
+      "learning_rate": 1.9673076966358606e-05,
+      "loss": 0.249,
+      "step": 965
+    },
+    {
+      "epoch": 2.1603563474387526,
+      "grad_norm": 1.3239182233810425,
+      "learning_rate": 1.9567935367581913e-05,
+      "loss": 0.2597,
+      "step": 970
+    },
+    {
+      "epoch": 2.1714922048997773,
+      "grad_norm": 1.1536415815353394,
+      "learning_rate": 1.9462545807828046e-05,
+      "loss": 0.2602,
+      "step": 975
+    },
+    {
+      "epoch": 2.1826280623608016,
+      "grad_norm": 1.1208415031433105,
+      "learning_rate": 1.935691400795252e-05,
+      "loss": 0.2929,
+      "step": 980
+    },
+    {
+      "epoch": 2.1937639198218264,
+      "grad_norm": 1.0744553804397583,
+      "learning_rate": 1.925104570196036e-05,
+      "loss": 0.2335,
+      "step": 985
+    },
+    {
+      "epoch": 2.2048997772828507,
+      "grad_norm": 1.1932042837142944,
+      "learning_rate": 1.914494663669482e-05,
+      "loss": 0.3014,
+      "step": 990
+    },
+    {
+      "epoch": 2.2160356347438754,
+      "grad_norm": 1.2387852668762207,
+      "learning_rate": 1.9038622571525464e-05,
+      "loss": 0.2754,
+      "step": 995
+    },
+    {
+      "epoch": 2.2271714922048997,
+      "grad_norm": 1.1342064142227173,
+      "learning_rate": 1.8932079278035503e-05,
+      "loss": 0.2704,
+      "step": 1000
+    },
+    {
+      "epoch": 2.2383073496659245,
+      "grad_norm": 1.0542978048324585,
+      "learning_rate": 1.882532253970852e-05,
+      "loss": 0.2913,
+      "step": 1005
+    },
+    {
+      "epoch": 2.249443207126949,
+      "grad_norm": 1.4599226713180542,
+      "learning_rate": 1.871835815161449e-05,
+      "loss": 0.3044,
+      "step": 1010
+    },
+    {
+      "epoch": 2.260579064587973,
+      "grad_norm": 1.1367197036743164,
+      "learning_rate": 1.861119192009526e-05,
+      "loss": 0.2821,
+      "step": 1015
+    },
+    {
+      "epoch": 2.271714922048998,
+      "grad_norm": 1.1666250228881836,
+      "learning_rate": 1.850382966244929e-05,
+      "loss": 0.2834,
+      "step": 1020
+    },
+    {
+      "epoch": 2.282850779510022,
+      "grad_norm": 1.2641099691390991,
+      "learning_rate": 1.8396277206615953e-05,
+      "loss": 0.2586,
+      "step": 1025
+    },
+    {
+      "epoch": 2.293986636971047,
+      "grad_norm": 1.1135034561157227,
+      "learning_rate": 1.8288540390859124e-05,
+      "loss": 0.2727,
+      "step": 1030
+    },
+    {
+      "epoch": 2.305122494432071,
+      "grad_norm": 1.1536855697631836,
+      "learning_rate": 1.8180625063450278e-05,
+      "loss": 0.2767,
+      "step": 1035
+    },
+    {
+      "epoch": 2.316258351893096,
+      "grad_norm": 1.176110029220581,
+      "learning_rate": 1.8072537082351028e-05,
+      "loss": 0.2583,
+      "step": 1040
+    },
+    {
+      "epoch": 2.3273942093541202,
+      "grad_norm": 1.2295929193496704,
+      "learning_rate": 1.796428231489514e-05,
+      "loss": 0.2559,
+      "step": 1045
+    },
+    {
+      "epoch": 2.338530066815145,
+      "grad_norm": 1.1238468885421753,
+      "learning_rate": 1.7855866637470027e-05,
+      "loss": 0.2523,
+      "step": 1050
+    },
+    {
+      "epoch": 2.3496659242761693,
+      "grad_norm": 1.1042718887329102,
+      "learning_rate": 1.7747295935197776e-05,
+      "loss": 0.2766,
+      "step": 1055
+    },
+    {
+      "epoch": 2.3608017817371936,
+      "grad_norm": 1.2366911172866821,
+      "learning_rate": 1.7638576101615678e-05,
+      "loss": 0.2461,
+      "step": 1060
+    },
+    {
+      "epoch": 2.3719376391982183,
+      "grad_norm": 1.1563042402267456,
+      "learning_rate": 1.7529713038356313e-05,
+      "loss": 0.2362,
+      "step": 1065
+    },
+    {
+      "epoch": 2.3830734966592426,
+      "grad_norm": 1.2360973358154297,
+      "learning_rate": 1.7420712654827182e-05,
+      "loss": 0.2508,
+      "step": 1070
+    },
+    {
+      "epoch": 2.3942093541202674,
+      "grad_norm": 1.1910051107406616,
+      "learning_rate": 1.7311580867889935e-05,
+      "loss": 0.2282,
+      "step": 1075
+    },
+    {
+      "epoch": 2.4053452115812917,
+      "grad_norm": 1.0605477094650269,
+      "learning_rate": 1.7202323601539197e-05,
+      "loss": 0.223,
+      "step": 1080
+    },
+    {
+      "epoch": 2.4164810690423164,
+      "grad_norm": 1.1978503465652466,
+      "learning_rate": 1.7092946786580964e-05,
+      "loss": 0.262,
+      "step": 1085
+    },
+    {
+      "epoch": 2.4276169265033407,
+      "grad_norm": 1.2242716550827026,
+      "learning_rate": 1.6983456360310716e-05,
+      "loss": 0.2493,
+      "step": 1090
+    },
+    {
+      "epoch": 2.4387527839643655,
+      "grad_norm": 1.0966243743896484,
+      "learning_rate": 1.687385826619105e-05,
+      "loss": 0.2252,
+      "step": 1095
+    },
+    {
+      "epoch": 2.4498886414253898,
+      "grad_norm": 1.2636704444885254,
+      "learning_rate": 1.6764158453529124e-05,
+      "loss": 0.2501,
+      "step": 1100
+    },
+    {
+      "epoch": 2.461024498886414,
+      "grad_norm": 1.1845978498458862,
+      "learning_rate": 1.665436287715365e-05,
+      "loss": 0.246,
+      "step": 1105
+    },
+    {
+      "epoch": 2.472160356347439,
+      "grad_norm": 0.9880180954933167,
+      "learning_rate": 1.6544477497091698e-05,
+      "loss": 0.236,
+      "step": 1110
+    },
+    {
+      "epoch": 2.483296213808463,
+      "grad_norm": 1.1788389682769775,
+      "learning_rate": 1.6434508278245134e-05,
+      "loss": 0.2377,
+      "step": 1115
+    },
+    {
+      "epoch": 2.494432071269488,
+      "grad_norm": 1.3312638998031616,
+      "learning_rate": 1.632446119006685e-05,
+      "loss": 0.2576,
+      "step": 1120
+    },
+    {
+      "epoch": 2.505567928730512,
+      "grad_norm": 1.2232873439788818,
+      "learning_rate": 1.6214342206236703e-05,
+      "loss": 0.2258,
+      "step": 1125
+    },
+    {
+      "epoch": 2.516703786191537,
+      "grad_norm": 1.1207743883132935,
+      "learning_rate": 1.6104157304337264e-05,
+      "loss": 0.2096,
+      "step": 1130
+    },
+    {
+      "epoch": 2.527839643652561,
+      "grad_norm": 1.002036690711975,
+      "learning_rate": 1.5993912465529337e-05,
+      "loss": 0.2574,
+      "step": 1135
+    },
+    {
+      "epoch": 2.538975501113586,
+      "grad_norm": 1.3266305923461914,
+      "learning_rate": 1.588361367422726e-05,
+      "loss": 0.2267,
+      "step": 1140
+    },
+    {
+      "epoch": 2.5501113585746102,
+      "grad_norm": 1.281622290611267,
+      "learning_rate": 1.5773266917774087e-05,
+      "loss": 0.2529,
+      "step": 1145
+    },
+    {
+      "epoch": 2.5612472160356345,
+      "grad_norm": 1.1964174509048462,
+      "learning_rate": 1.5662878186116556e-05,
+      "loss": 0.2393,
+      "step": 1150
+    },
+    {
+      "epoch": 2.5723830734966593,
+      "grad_norm": 1.0871187448501587,
+      "learning_rate": 1.555245347147995e-05,
+      "loss": 0.2623,
+      "step": 1155
+    },
+    {
+      "epoch": 2.5835189309576836,
+      "grad_norm": 1.1345094442367554,
+      "learning_rate": 1.54419987680428e-05,
+      "loss": 0.2343,
+      "step": 1160
+    },
+    {
+      "epoch": 2.5946547884187083,
+      "grad_norm": 1.1741135120391846,
+      "learning_rate": 1.533152007161152e-05,
+      "loss": 0.231,
+      "step": 1165
+    },
+    {
+      "epoch": 2.6057906458797326,
+      "grad_norm": 1.1888442039489746,
+      "learning_rate": 1.5221023379294952e-05,
+      "loss": 0.2437,
+      "step": 1170
+    },
+    {
+      "epoch": 2.6169265033407574,
+      "grad_norm": 1.1484440565109253,
+      "learning_rate": 1.5110514689178788e-05,
+      "loss": 0.238,
+      "step": 1175
+    },
+    {
+      "epoch": 2.6280623608017817,
+      "grad_norm": 1.2487529516220093,
+      "learning_rate": 1.5e-05,
+      "loss": 0.2416,
+      "step": 1180
+    },
+    {
+      "epoch": 2.6391982182628064,
+      "grad_norm": 1.165150761604309,
+      "learning_rate": 1.4889485310821213e-05,
+      "loss": 0.2356,
+      "step": 1185
+    },
+    {
+      "epoch": 2.6503340757238307,
+      "grad_norm": 1.1523540019989014,
+      "learning_rate": 1.477897662070505e-05,
+      "loss": 0.2144,
+      "step": 1190
+    },
+    {
+      "epoch": 2.661469933184855,
+      "grad_norm": 1.0735657215118408,
+      "learning_rate": 1.4668479928388478e-05,
+      "loss": 0.2212,
+      "step": 1195
+    },
+    {
+      "epoch": 2.6726057906458798,
+      "grad_norm": 1.110089898109436,
+      "learning_rate": 1.4558001231957202e-05,
+      "loss": 0.21,
+      "step": 1200
+    },
+    {
+      "epoch": 2.683741648106904,
+      "grad_norm": 1.2104490995407104,
+      "learning_rate": 1.444754652852005e-05,
+      "loss": 0.2157,
+      "step": 1205
+    },
+    {
+      "epoch": 2.694877505567929,
+      "grad_norm": 1.134554386138916,
+      "learning_rate": 1.4337121813883443e-05,
+      "loss": 0.2285,
+      "step": 1210
+    },
+    {
+      "epoch": 2.706013363028953,
+      "grad_norm": 1.0749125480651855,
+      "learning_rate": 1.4226733082225917e-05,
+      "loss": 0.2279,
+      "step": 1215
+    },
+    {
+      "epoch": 2.717149220489978,
+      "grad_norm": 1.1136120557785034,
+      "learning_rate": 1.4116386325772745e-05,
+      "loss": 0.2124,
+      "step": 1220
+    },
+    {
+      "epoch": 2.728285077951002,
+      "grad_norm": 1.1267787218093872,
+      "learning_rate": 1.4006087534470667e-05,
+      "loss": 0.2265,
+      "step": 1225
+    },
+    {
+      "epoch": 2.739420935412027,
+      "grad_norm": 1.248026728630066,
+      "learning_rate": 1.3895842695662738e-05,
+      "loss": 0.2335,
+      "step": 1230
+    },
+    {
+      "epoch": 2.750556792873051,
+      "grad_norm": 1.3174666166305542,
+      "learning_rate": 1.37856577937633e-05,
+      "loss": 0.2214,
+      "step": 1235
+    },
+    {
+      "epoch": 2.7616926503340755,
+      "grad_norm": 0.9793177247047424,
+      "learning_rate": 1.367553880993315e-05,
+      "loss": 0.1913,
+      "step": 1240
+    },
+    {
+      "epoch": 2.7728285077951003,
+      "grad_norm": 1.211518406867981,
+      "learning_rate": 1.3565491721754868e-05,
+      "loss": 0.201,
+      "step": 1245
+    },
+    {
+      "epoch": 2.7839643652561246,
+      "grad_norm": 1.1160845756530762,
+      "learning_rate": 1.3455522502908305e-05,
+      "loss": 0.2075,
+      "step": 1250
+    },
+    {
+      "epoch": 2.7951002227171493,
+      "grad_norm": 1.0458290576934814,
+      "learning_rate": 1.334563712284635e-05,
+      "loss": 0.2067,
+      "step": 1255
+    },
+    {
+      "epoch": 2.8062360801781736,
+      "grad_norm": 1.0362893342971802,
+      "learning_rate": 1.3235841546470884e-05,
+      "loss": 0.1904,
+      "step": 1260
+    },
+    {
+      "epoch": 2.8173719376391984,
+      "grad_norm": 1.207885980606079,
+      "learning_rate": 1.312614173380895e-05,
+      "loss": 0.1989,
+      "step": 1265
+    },
+    {
+      "epoch": 2.8285077951002227,
+      "grad_norm": 1.2045843601226807,
+      "learning_rate": 1.301654363968929e-05,
+      "loss": 0.2144,
+      "step": 1270
+    },
+    {
+      "epoch": 2.8396436525612474,
+      "grad_norm": 1.1977723836898804,
+      "learning_rate": 1.2907053213419035e-05,
+      "loss": 0.2294,
+      "step": 1275
+    },
+    {
+      "epoch": 2.8507795100222717,
+      "grad_norm": 1.1156507730484009,
+      "learning_rate": 1.2797676398460808e-05,
+      "loss": 0.1677,
+      "step": 1280
+    },
+    {
+      "epoch": 2.861915367483296,
+      "grad_norm": 1.168018102645874,
+      "learning_rate": 1.268841913211007e-05,
+      "loss": 0.2166,
+      "step": 1285
+    },
+    {
+      "epoch": 2.8730512249443207,
+      "grad_norm": 1.1159573793411255,
+      "learning_rate": 1.2579287345172824e-05,
+      "loss": 0.19,
+      "step": 1290
+    },
+    {
+      "epoch": 2.884187082405345,
+      "grad_norm": 1.0215970277786255,
+      "learning_rate": 1.2470286961643688e-05,
+      "loss": 0.1599,
+      "step": 1295
+    },
+    {
+      "epoch": 2.89532293986637,
+      "grad_norm": 1.1599113941192627,
+      "learning_rate": 1.2361423898384323e-05,
+      "loss": 0.2201,
+      "step": 1300
+    },
+    {
+      "epoch": 2.906458797327394,
+      "grad_norm": 1.1932123899459839,
+      "learning_rate": 1.2252704064802225e-05,
+      "loss": 0.1998,
+      "step": 1305
+    },
+    {
+      "epoch": 2.917594654788419,
+      "grad_norm": 0.9487115740776062,
+      "learning_rate": 1.2144133362529972e-05,
+      "loss": 0.2255,
+      "step": 1310
+    },
+    {
+      "epoch": 2.928730512249443,
+      "grad_norm": 1.045422077178955,
+      "learning_rate": 1.2035717685104866e-05,
+      "loss": 0.1796,
+      "step": 1315
+    },
+    {
+      "epoch": 2.939866369710468,
+      "grad_norm": 1.1371259689331055,
+      "learning_rate": 1.1927462917648973e-05,
+      "loss": 0.1721,
+      "step": 1320
+    },
+    {
+      "epoch": 2.951002227171492,
+      "grad_norm": 1.2436094284057617,
+      "learning_rate": 1.1819374936549723e-05,
+      "loss": 0.1794,
+      "step": 1325
+    },
+    {
+      "epoch": 2.9621380846325165,
+      "grad_norm": 1.2351981401443481,
+      "learning_rate": 1.1711459609140875e-05,
+      "loss": 0.2143,
+      "step": 1330
+    },
+    {
+      "epoch": 2.9732739420935412,
+      "grad_norm": 0.973818302154541,
+      "learning_rate": 1.1603722793384046e-05,
+      "loss": 0.1637,
+      "step": 1335
+    },
+    {
+      "epoch": 2.984409799554566,
+      "grad_norm": 0.9122995138168335,
+      "learning_rate": 1.1496170337550713e-05,
+      "loss": 0.1779,
+      "step": 1340
+    },
+    {
+      "epoch": 2.9955456570155903,
+      "grad_norm": 1.0754543542861938,
+      "learning_rate": 1.1388808079904745e-05,
+      "loss": 0.1809,
+      "step": 1345
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2245,
+  "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": 1.864228049118036e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

38_128_e5_3e-5/checkpoint-1347/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28546b29d51193caf8c0ea9b7bbe68af481f6dc06b3ac3e27fa68da70bce3d29
+size 7736

38_128_e5_3e-5/checkpoint-1347/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)

38_128_e5_3e-5/checkpoint-1796/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

38_128_e5_3e-5/checkpoint-1796/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": [
+    "q_proj",
+    "v_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "k_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

38_128_e5_3e-5/checkpoint-1796/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:071fc7224699ca0dc5313bf69478c152f06a5f06341f0622d7bdbfc31cf639d2
+size 791751704

38_128_e5_3e-5/checkpoint-1796/latest

@@ -0,0 +1 @@
+global_step1796

38_128_e5_3e-5/checkpoint-1796/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a176ec84065127aa3a7a767a71662c110bbb168b7a43b0d99abdd3a7bf91295
+size 15920

38_128_e5_3e-5/checkpoint-1796/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0939afdd15692c0935bafb14aa4fb7625639805d56394844992a8c85c419bcc4
+size 15920

38_128_e5_3e-5/checkpoint-1796/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1e376956fa27b290547a35714d7ff8c7c307d10e9744c52482e04668a492e51
+size 15920

38_128_e5_3e-5/checkpoint-1796/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a59454ce290e1607d29c9f3bda47b63357714ee160440e59bfbdd2d47a0adf0
+size 15920

38_128_e5_3e-5/checkpoint-1796/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7240db1e8be89b7e34bfd984216832d36516c92ce4754da5f158de4695a283b1
+size 15920

38_128_e5_3e-5/checkpoint-1796/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e0aa5ab3966dc9a2686041fb3d02b91cc8ab643f0b6e1ac9637ace613009034
+size 15920

38_128_e5_3e-5/checkpoint-1796/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:be260b48211253c7bad837562413ad023feaa919a33865dfbbf6e39a8e5dbf3e
+size 15920

38_128_e5_3e-5/checkpoint-1796/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e9a98366eff43f10e47406c4a3ef57da684979bb005e26c7f23c50f6cdb3e7f
+size 15920

38_128_e5_3e-5/checkpoint-1796/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:133d6ab241fbc6e3e648d308f5e462540d552ba63cab6025e4ad1b75b902363d
+size 1064

38_128_e5_3e-5/checkpoint-1796/special_tokens_map.json

@@ -0,0 +1,45 @@
+{
+  "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": "<reponame>",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

38_128_e5_3e-5/checkpoint-1796/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": "<reponame>",
+  "padding_side": "left",
+  "tokenizer_class": "GPT2Tokenizer",
+  "unk_token": "<|endoftext|>",
+  "vocab_size": 49152
+}

38_128_e5_3e-5/checkpoint-1796/trainer_state.json

@@ -0,0 +1,2547 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1796,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011135857461024499,
+      "grad_norm": 1.3413572311401367,
+      "learning_rate": 1.0619469026548673e-06,
+      "loss": 1.3355,
+      "step": 5
+    },
+    {
+      "epoch": 0.022271714922048998,
+      "grad_norm": 1.0000957250595093,
+      "learning_rate": 2.389380530973451e-06,
+      "loss": 1.4079,
+      "step": 10
+    },
+    {
+      "epoch": 0.0334075723830735,
+      "grad_norm": 0.7801820039749146,
+      "learning_rate": 3.7168141592920353e-06,
+      "loss": 1.3474,
+      "step": 15
+    },
+    {
+      "epoch": 0.044543429844097995,
+      "grad_norm": 0.5836945176124573,
+      "learning_rate": 5.04424778761062e-06,
+      "loss": 1.3387,
+      "step": 20
+    },
+    {
+      "epoch": 0.0556792873051225,
+      "grad_norm": 0.6259394288063049,
+      "learning_rate": 6.371681415929204e-06,
+      "loss": 1.3457,
+      "step": 25
+    },
+    {
+      "epoch": 0.066815144766147,
+      "grad_norm": 0.6279721856117249,
+      "learning_rate": 7.699115044247788e-06,
+      "loss": 1.345,
+      "step": 30
+    },
+    {
+      "epoch": 0.0779510022271715,
+      "grad_norm": 0.5959689021110535,
+      "learning_rate": 9.026548672566373e-06,
+      "loss": 1.3064,
+      "step": 35
+    },
+    {
+      "epoch": 0.08908685968819599,
+      "grad_norm": 0.4760344624519348,
+      "learning_rate": 1.0353982300884955e-05,
+      "loss": 1.1918,
+      "step": 40
+    },
+    {
+      "epoch": 0.10022271714922049,
+      "grad_norm": 0.5409045219421387,
+      "learning_rate": 1.168141592920354e-05,
+      "loss": 1.2243,
+      "step": 45
+    },
+    {
+      "epoch": 0.111358574610245,
+      "grad_norm": 0.6069848537445068,
+      "learning_rate": 1.3008849557522125e-05,
+      "loss": 1.2276,
+      "step": 50
+    },
+    {
+      "epoch": 0.12249443207126949,
+      "grad_norm": 0.5506272315979004,
+      "learning_rate": 1.433628318584071e-05,
+      "loss": 1.2671,
+      "step": 55
+    },
+    {
+      "epoch": 0.133630289532294,
+      "grad_norm": 0.4604385793209076,
+      "learning_rate": 1.5663716814159295e-05,
+      "loss": 1.2268,
+      "step": 60
+    },
+    {
+      "epoch": 0.1447661469933185,
+      "grad_norm": 0.5579900741577148,
+      "learning_rate": 1.6991150442477876e-05,
+      "loss": 1.2215,
+      "step": 65
+    },
+    {
+      "epoch": 0.155902004454343,
+      "grad_norm": 0.48294851183891296,
+      "learning_rate": 1.8318584070796458e-05,
+      "loss": 1.2418,
+      "step": 70
+    },
+    {
+      "epoch": 0.16703786191536749,
+      "grad_norm": 0.5626055598258972,
+      "learning_rate": 1.9646017699115046e-05,
+      "loss": 1.185,
+      "step": 75
+    },
+    {
+      "epoch": 0.17817371937639198,
+      "grad_norm": 0.5725001096725464,
+      "learning_rate": 2.0973451327433628e-05,
+      "loss": 1.2337,
+      "step": 80
+    },
+    {
+      "epoch": 0.18930957683741648,
+      "grad_norm": 0.5308484435081482,
+      "learning_rate": 2.2300884955752213e-05,
+      "loss": 1.164,
+      "step": 85
+    },
+    {
+      "epoch": 0.20044543429844097,
+      "grad_norm": 0.5047495365142822,
+      "learning_rate": 2.3628318584070798e-05,
+      "loss": 1.1862,
+      "step": 90
+    },
+    {
+      "epoch": 0.21158129175946547,
+      "grad_norm": 0.5524989366531372,
+      "learning_rate": 2.495575221238938e-05,
+      "loss": 1.1503,
+      "step": 95
+    },
+    {
+      "epoch": 0.22271714922049,
+      "grad_norm": 0.5751224160194397,
+      "learning_rate": 2.6283185840707968e-05,
+      "loss": 1.1819,
+      "step": 100
+    },
+    {
+      "epoch": 0.23385300668151449,
+      "grad_norm": 0.590943455696106,
+      "learning_rate": 2.761061946902655e-05,
+      "loss": 1.1177,
+      "step": 105
+    },
+    {
+      "epoch": 0.24498886414253898,
+      "grad_norm": 0.624424159526825,
+      "learning_rate": 2.893805309734513e-05,
+      "loss": 1.1612,
+      "step": 110
+    },
+    {
+      "epoch": 0.2561247216035635,
+      "grad_norm": 0.7900341153144836,
+      "learning_rate": 2.9999983715046248e-05,
+      "loss": 1.1556,
+      "step": 115
+    },
+    {
+      "epoch": 0.267260579064588,
+      "grad_norm": 0.717691957950592,
+      "learning_rate": 2.9999413745377787e-05,
+      "loss": 1.1074,
+      "step": 120
+    },
+    {
+      "epoch": 0.27839643652561247,
+      "grad_norm": 0.5590342283248901,
+      "learning_rate": 2.9998029563381436e-05,
+      "loss": 1.1106,
+      "step": 125
+    },
+    {
+      "epoch": 0.289532293986637,
+      "grad_norm": 0.5695084929466248,
+      "learning_rate": 2.999583124419468e-05,
+      "loss": 1.0615,
+      "step": 130
+    },
+    {
+      "epoch": 0.30066815144766146,
+      "grad_norm": 0.6339223384857178,
+      "learning_rate": 2.999281890714874e-05,
+      "loss": 1.1041,
+      "step": 135
+    },
+    {
+      "epoch": 0.311804008908686,
+      "grad_norm": 0.6401590704917908,
+      "learning_rate": 2.9988992715762148e-05,
+      "loss": 1.0891,
+      "step": 140
+    },
+    {
+      "epoch": 0.32293986636971045,
+      "grad_norm": 0.6139689087867737,
+      "learning_rate": 2.9984352877731836e-05,
+      "loss": 1.0935,
+      "step": 145
+    },
+    {
+      "epoch": 0.33407572383073497,
+      "grad_norm": 0.5202445387840271,
+      "learning_rate": 2.9978899644921866e-05,
+      "loss": 1.0622,
+      "step": 150
+    },
+    {
+      "epoch": 0.34521158129175944,
+      "grad_norm": 0.5589560866355896,
+      "learning_rate": 2.9972633313349764e-05,
+      "loss": 1.0604,
+      "step": 155
+    },
+    {
+      "epoch": 0.35634743875278396,
+      "grad_norm": 0.6837779879570007,
+      "learning_rate": 2.9965554223170455e-05,
+      "loss": 1.008,
+      "step": 160
+    },
+    {
+      "epoch": 0.3674832962138085,
+      "grad_norm": 0.7174135446548462,
+      "learning_rate": 2.9957662758657787e-05,
+      "loss": 1.0292,
+      "step": 165
+    },
+    {
+      "epoch": 0.37861915367483295,
+      "grad_norm": 0.6325995922088623,
+      "learning_rate": 2.9948959348183686e-05,
+      "loss": 1.0019,
+      "step": 170
+    },
+    {
+      "epoch": 0.3897550111358575,
+      "grad_norm": 0.6495500802993774,
+      "learning_rate": 2.993944446419489e-05,
+      "loss": 0.9838,
+      "step": 175
+    },
+    {
+      "epoch": 0.40089086859688194,
+      "grad_norm": 0.641270101070404,
+      "learning_rate": 2.9929118623187307e-05,
+      "loss": 0.9553,
+      "step": 180
+    },
+    {
+      "epoch": 0.41202672605790647,
+      "grad_norm": 0.6926126480102539,
+      "learning_rate": 2.9917982385677982e-05,
+      "loss": 1.0296,
+      "step": 185
+    },
+    {
+      "epoch": 0.42316258351893093,
+      "grad_norm": 0.7776742577552795,
+      "learning_rate": 2.990603635617466e-05,
+      "loss": 1.0223,
+      "step": 190
+    },
+    {
+      "epoch": 0.43429844097995546,
+      "grad_norm": 0.6762447357177734,
+      "learning_rate": 2.9893281183142996e-05,
+      "loss": 1.053,
+      "step": 195
+    },
+    {
+      "epoch": 0.44543429844098,
+      "grad_norm": 0.70978844165802,
+      "learning_rate": 2.9879717558971317e-05,
+      "loss": 0.9561,
+      "step": 200
+    },
+    {
+      "epoch": 0.45657015590200445,
+      "grad_norm": 0.7382838726043701,
+      "learning_rate": 2.986534621993307e-05,
+      "loss": 0.9923,
+      "step": 205
+    },
+    {
+      "epoch": 0.46770601336302897,
+      "grad_norm": 0.7762662768363953,
+      "learning_rate": 2.9850167946146838e-05,
+      "loss": 1.0125,
+      "step": 210
+    },
+    {
+      "epoch": 0.47884187082405344,
+      "grad_norm": 0.7446727752685547,
+      "learning_rate": 2.9834183561533997e-05,
+      "loss": 0.9445,
+      "step": 215
+    },
+    {
+      "epoch": 0.48997772828507796,
+      "grad_norm": 0.78370201587677,
+      "learning_rate": 2.9817393933773994e-05,
+      "loss": 0.9717,
+      "step": 220
+    },
+    {
+      "epoch": 0.5011135857461024,
+      "grad_norm": 0.6878297328948975,
+      "learning_rate": 2.9799799974257248e-05,
+      "loss": 0.9373,
+      "step": 225
+    },
+    {
+      "epoch": 0.512249443207127,
+      "grad_norm": 0.690732479095459,
+      "learning_rate": 2.9781402638035656e-05,
+      "loss": 0.9109,
+      "step": 230
+    },
+    {
+      "epoch": 0.5233853006681515,
+      "grad_norm": 0.8460608124732971,
+      "learning_rate": 2.9762202923770795e-05,
+      "loss": 0.9548,
+      "step": 235
+    },
+    {
+      "epoch": 0.534521158129176,
+      "grad_norm": 0.8983225226402283,
+      "learning_rate": 2.9742201873679656e-05,
+      "loss": 0.9547,
+      "step": 240
+    },
+    {
+      "epoch": 0.5456570155902004,
+      "grad_norm": 0.8119720816612244,
+      "learning_rate": 2.9721400573478117e-05,
+      "loss": 0.9748,
+      "step": 245
+    },
+    {
+      "epoch": 0.5567928730512249,
+      "grad_norm": 0.8376508951187134,
+      "learning_rate": 2.9699800152321975e-05,
+      "loss": 0.8978,
+      "step": 250
+    },
+    {
+      "epoch": 0.5679287305122495,
+      "grad_norm": 0.7981753349304199,
+      "learning_rate": 2.967740178274567e-05,
+      "loss": 0.8872,
+      "step": 255
+    },
+    {
+      "epoch": 0.579064587973274,
+      "grad_norm": 0.7605271339416504,
+      "learning_rate": 2.9654206680598638e-05,
+      "loss": 0.8954,
+      "step": 260
+    },
+    {
+      "epoch": 0.5902004454342984,
+      "grad_norm": 0.9272874593734741,
+      "learning_rate": 2.9630216104979288e-05,
+      "loss": 0.8797,
+      "step": 265
+    },
+    {
+      "epoch": 0.6013363028953229,
+      "grad_norm": 0.8306118249893188,
+      "learning_rate": 2.9605431358166687e-05,
+      "loss": 0.9045,
+      "step": 270
+    },
+    {
+      "epoch": 0.6124721603563474,
+      "grad_norm": 0.849515438079834,
+      "learning_rate": 2.957985378554984e-05,
+      "loss": 0.8929,
+      "step": 275
+    },
+    {
+      "epoch": 0.623608017817372,
+      "grad_norm": 0.9004138708114624,
+      "learning_rate": 2.955348477555467e-05,
+      "loss": 0.9355,
+      "step": 280
+    },
+    {
+      "epoch": 0.6347438752783965,
+      "grad_norm": 0.836115837097168,
+      "learning_rate": 2.9526325759568665e-05,
+      "loss": 0.8795,
+      "step": 285
+    },
+    {
+      "epoch": 0.6458797327394209,
+      "grad_norm": 0.8236976861953735,
+      "learning_rate": 2.9498378211863145e-05,
+      "loss": 0.8241,
+      "step": 290
+    },
+    {
+      "epoch": 0.6570155902004454,
+      "grad_norm": 0.817470908164978,
+      "learning_rate": 2.9469643649513264e-05,
+      "loss": 0.8541,
+      "step": 295
+    },
+    {
+      "epoch": 0.6681514476614699,
+      "grad_norm": 0.9785251021385193,
+      "learning_rate": 2.9440123632315647e-05,
+      "loss": 0.8715,
+      "step": 300
+    },
+    {
+      "epoch": 0.6792873051224945,
+      "grad_norm": 0.9281602501869202,
+      "learning_rate": 2.9409819762703715e-05,
+      "loss": 0.7914,
+      "step": 305
+    },
+    {
+      "epoch": 0.6904231625835189,
+      "grad_norm": 0.8825387358665466,
+      "learning_rate": 2.9378733685660707e-05,
+      "loss": 0.8527,
+      "step": 310
+    },
+    {
+      "epoch": 0.7015590200445434,
+      "grad_norm": 0.9314531683921814,
+      "learning_rate": 2.934686708863039e-05,
+      "loss": 0.8966,
+      "step": 315
+    },
+    {
+      "epoch": 0.7126948775055679,
+      "grad_norm": 0.9172025322914124,
+      "learning_rate": 2.9314221701425445e-05,
+      "loss": 0.8472,
+      "step": 320
+    },
+    {
+      "epoch": 0.7238307349665924,
+      "grad_norm": 0.9198710918426514,
+      "learning_rate": 2.9280799296133577e-05,
+      "loss": 0.878,
+      "step": 325
+    },
+    {
+      "epoch": 0.734966592427617,
+      "grad_norm": 0.8762511610984802,
+      "learning_rate": 2.9246601687021324e-05,
+      "loss": 0.7845,
+      "step": 330
+    },
+    {
+      "epoch": 0.7461024498886414,
+      "grad_norm": 0.8782876133918762,
+      "learning_rate": 2.9211630730435564e-05,
+      "loss": 0.7928,
+      "step": 335
+    },
+    {
+      "epoch": 0.7572383073496659,
+      "grad_norm": 0.9156160950660706,
+      "learning_rate": 2.9175888324702762e-05,
+      "loss": 0.8467,
+      "step": 340
+    },
+    {
+      "epoch": 0.7683741648106904,
+      "grad_norm": 1.0404008626937866,
+      "learning_rate": 2.9139376410025897e-05,
+      "loss": 0.7927,
+      "step": 345
+    },
+    {
+      "epoch": 0.779510022271715,
+      "grad_norm": 1.0743643045425415,
+      "learning_rate": 2.9102096968379178e-05,
+      "loss": 0.8123,
+      "step": 350
+    },
+    {
+      "epoch": 0.7906458797327395,
+      "grad_norm": 0.9985504150390625,
+      "learning_rate": 2.9064052023400417e-05,
+      "loss": 0.7735,
+      "step": 355
+    },
+    {
+      "epoch": 0.8017817371937639,
+      "grad_norm": 0.9021556973457336,
+      "learning_rate": 2.9025243640281226e-05,
+      "loss": 0.8006,
+      "step": 360
+    },
+    {
+      "epoch": 0.8129175946547884,
+      "grad_norm": 0.9394658803939819,
+      "learning_rate": 2.8985673925654853e-05,
+      "loss": 0.7879,
+      "step": 365
+    },
+    {
+      "epoch": 0.8240534521158129,
+      "grad_norm": 0.88721764087677,
+      "learning_rate": 2.8945345027481884e-05,
+      "loss": 0.7225,
+      "step": 370
+    },
+    {
+      "epoch": 0.8351893095768375,
+      "grad_norm": 0.9523298740386963,
+      "learning_rate": 2.8904259134933627e-05,
+      "loss": 0.8202,
+      "step": 375
+    },
+    {
+      "epoch": 0.8463251670378619,
+      "grad_norm": 0.9752704501152039,
+      "learning_rate": 2.8862418478273246e-05,
+      "loss": 0.7871,
+      "step": 380
+    },
+    {
+      "epoch": 0.8574610244988864,
+      "grad_norm": 0.9238730072975159,
+      "learning_rate": 2.881982532873476e-05,
+      "loss": 0.7216,
+      "step": 385
+    },
+    {
+      "epoch": 0.8685968819599109,
+      "grad_norm": 0.8588601350784302,
+      "learning_rate": 2.8776481998399676e-05,
+      "loss": 0.7146,
+      "step": 390
+    },
+    {
+      "epoch": 0.8797327394209354,
+      "grad_norm": 0.8526139259338379,
+      "learning_rate": 2.873239084007156e-05,
+      "loss": 0.8105,
+      "step": 395
+    },
+    {
+      "epoch": 0.89086859688196,
+      "grad_norm": 1.0353056192398071,
+      "learning_rate": 2.868755424714825e-05,
+      "loss": 0.789,
+      "step": 400
+    },
+    {
+      "epoch": 0.9020044543429844,
+      "grad_norm": 0.8390752077102661,
+      "learning_rate": 2.8641974653491997e-05,
+      "loss": 0.7338,
+      "step": 405
+    },
+    {
+      "epoch": 0.9131403118040089,
+      "grad_norm": 0.9766253232955933,
+      "learning_rate": 2.85956545332973e-05,
+      "loss": 0.7782,
+      "step": 410
+    },
+    {
+      "epoch": 0.9242761692650334,
+      "grad_norm": 0.9131956100463867,
+      "learning_rate": 2.8548596400956614e-05,
+      "loss": 0.734,
+      "step": 415
+    },
+    {
+      "epoch": 0.9354120267260579,
+      "grad_norm": 0.979026734828949,
+      "learning_rate": 2.850080281092389e-05,
+      "loss": 0.7538,
+      "step": 420
+    },
+    {
+      "epoch": 0.9465478841870824,
+      "grad_norm": 1.0852129459381104,
+      "learning_rate": 2.845227635757587e-05,
+      "loss": 0.7072,
+      "step": 425
+    },
+    {
+      "epoch": 0.9576837416481069,
+      "grad_norm": 1.1135140657424927,
+      "learning_rate": 2.8403019675071267e-05,
+      "loss": 0.7657,
+      "step": 430
+    },
+    {
+      "epoch": 0.9688195991091314,
+      "grad_norm": 0.8528003096580505,
+      "learning_rate": 2.8353035437207796e-05,
+      "loss": 0.7036,
+      "step": 435
+    },
+    {
+      "epoch": 0.9799554565701559,
+      "grad_norm": 0.9384568929672241,
+      "learning_rate": 2.8302326357277013e-05,
+      "loss": 0.6948,
+      "step": 440
+    },
+    {
+      "epoch": 0.9910913140311804,
+      "grad_norm": 1.1752519607543945,
+      "learning_rate": 2.825089518791704e-05,
+      "loss": 0.7052,
+      "step": 445
+    },
+    {
+      "epoch": 1.0022271714922049,
+      "grad_norm": 1.038319706916809,
+      "learning_rate": 2.8198744720963127e-05,
+      "loss": 0.6852,
+      "step": 450
+    },
+    {
+      "epoch": 1.0133630289532294,
+      "grad_norm": 1.007336974143982,
+      "learning_rate": 2.814587778729613e-05,
+      "loss": 0.6137,
+      "step": 455
+    },
+    {
+      "epoch": 1.024498886414254,
+      "grad_norm": 1.089078426361084,
+      "learning_rate": 2.809229725668881e-05,
+      "loss": 0.6511,
+      "step": 460
+    },
+    {
+      "epoch": 1.0356347438752784,
+      "grad_norm": 0.9254465699195862,
+      "learning_rate": 2.803800603765008e-05,
+      "loss": 0.601,
+      "step": 465
+    },
+    {
+      "epoch": 1.046770601336303,
+      "grad_norm": 1.0865051746368408,
+      "learning_rate": 2.798300707726712e-05,
+      "loss": 0.6846,
+      "step": 470
+    },
+    {
+      "epoch": 1.0579064587973275,
+      "grad_norm": 0.9619767665863037,
+      "learning_rate": 2.7927303361045376e-05,
+      "loss": 0.659,
+      "step": 475
+    },
+    {
+      "epoch": 1.069042316258352,
+      "grad_norm": 1.1050430536270142,
+      "learning_rate": 2.7870897912746532e-05,
+      "loss": 0.5382,
+      "step": 480
+    },
+    {
+      "epoch": 1.0801781737193763,
+      "grad_norm": 1.060831069946289,
+      "learning_rate": 2.7813793794224348e-05,
+      "loss": 0.5903,
+      "step": 485
+    },
+    {
+      "epoch": 1.0913140311804008,
+      "grad_norm": 1.0150086879730225,
+      "learning_rate": 2.775599410525845e-05,
+      "loss": 0.577,
+      "step": 490
+    },
+    {
+      "epoch": 1.1024498886414253,
+      "grad_norm": 1.1610509157180786,
+      "learning_rate": 2.769750198338609e-05,
+      "loss": 0.5755,
+      "step": 495
+    },
+    {
+      "epoch": 1.1135857461024499,
+      "grad_norm": 1.022871494293213,
+      "learning_rate": 2.7638320603731817e-05,
+      "loss": 0.5687,
+      "step": 500
+    },
+    {
+      "epoch": 1.1247216035634744,
+      "grad_norm": 1.0842667818069458,
+      "learning_rate": 2.7578453178835108e-05,
+      "loss": 0.5426,
+      "step": 505
+    },
+    {
+      "epoch": 1.135857461024499,
+      "grad_norm": 1.1093308925628662,
+      "learning_rate": 2.7517902958476003e-05,
+      "loss": 0.5733,
+      "step": 510
+    },
+    {
+      "epoch": 1.1469933184855234,
+      "grad_norm": 1.0264402627944946,
+      "learning_rate": 2.745667322949869e-05,
+      "loss": 0.5749,
+      "step": 515
+    },
+    {
+      "epoch": 1.158129175946548,
+      "grad_norm": 1.2180768251419067,
+      "learning_rate": 2.7394767315633086e-05,
+      "loss": 0.5767,
+      "step": 520
+    },
+    {
+      "epoch": 1.1692650334075725,
+      "grad_norm": 1.0007996559143066,
+      "learning_rate": 2.7332188577314407e-05,
+      "loss": 0.5277,
+      "step": 525
+    },
+    {
+      "epoch": 1.1804008908685968,
+      "grad_norm": 0.9211612343788147,
+      "learning_rate": 2.7268940411500768e-05,
+      "loss": 0.6127,
+      "step": 530
+    },
+    {
+      "epoch": 1.1915367483296213,
+      "grad_norm": 1.1937363147735596,
+      "learning_rate": 2.720502625148877e-05,
+      "loss": 0.5765,
+      "step": 535
+    },
+    {
+      "epoch": 1.2026726057906458,
+      "grad_norm": 1.041098713874817,
+      "learning_rate": 2.7140449566727156e-05,
+      "loss": 0.5372,
+      "step": 540
+    },
+    {
+      "epoch": 1.2138084632516704,
+      "grad_norm": 1.0834497213363647,
+      "learning_rate": 2.707521386262845e-05,
+      "loss": 0.6064,
+      "step": 545
+    },
+    {
+      "epoch": 1.2249443207126949,
+      "grad_norm": 1.024735450744629,
+      "learning_rate": 2.700932268037868e-05,
+      "loss": 0.5451,
+      "step": 550
+    },
+    {
+      "epoch": 1.2360801781737194,
+      "grad_norm": 1.0640431642532349,
+      "learning_rate": 2.6942779596745172e-05,
+      "loss": 0.535,
+      "step": 555
+    },
+    {
+      "epoch": 1.247216035634744,
+      "grad_norm": 1.0001004934310913,
+      "learning_rate": 2.6875588223882363e-05,
+      "loss": 0.5336,
+      "step": 560
+    },
+    {
+      "epoch": 1.2583518930957684,
+      "grad_norm": 1.328310251235962,
+      "learning_rate": 2.680775220913575e-05,
+      "loss": 0.5681,
+      "step": 565
+    },
+    {
+      "epoch": 1.269487750556793,
+      "grad_norm": 1.0832302570343018,
+      "learning_rate": 2.6739275234843894e-05,
+      "loss": 0.6001,
+      "step": 570
+    },
+    {
+      "epoch": 1.2806236080178173,
+      "grad_norm": 1.1307528018951416,
+      "learning_rate": 2.66701610181385e-05,
+      "loss": 0.5613,
+      "step": 575
+    },
+    {
+      "epoch": 1.2917594654788418,
+      "grad_norm": 1.1129206418991089,
+      "learning_rate": 2.6600413310742696e-05,
+      "loss": 0.5266,
+      "step": 580
+    },
+    {
+      "epoch": 1.3028953229398663,
+      "grad_norm": 1.110201358795166,
+      "learning_rate": 2.653003589876733e-05,
+      "loss": 0.5429,
+      "step": 585
+    },
+    {
+      "epoch": 1.3140311804008908,
+      "grad_norm": 1.1400443315505981,
+      "learning_rate": 2.645903260250548e-05,
+      "loss": 0.5909,
+      "step": 590
+    },
+    {
+      "epoch": 1.3251670378619154,
+      "grad_norm": 1.065966248512268,
+      "learning_rate": 2.6387407276225057e-05,
+      "loss": 0.5312,
+      "step": 595
+    },
+    {
+      "epoch": 1.3363028953229399,
+      "grad_norm": 1.266574501991272,
+      "learning_rate": 2.6315163807959603e-05,
+      "loss": 0.5414,
+      "step": 600
+    },
+    {
+      "epoch": 1.3474387527839644,
+      "grad_norm": 0.9977100491523743,
+      "learning_rate": 2.6242306119297205e-05,
+      "loss": 0.5039,
+      "step": 605
+    },
+    {
+      "epoch": 1.358574610244989,
+      "grad_norm": 1.1061787605285645,
+      "learning_rate": 2.6168838165167674e-05,
+      "loss": 0.4875,
+      "step": 610
+    },
+    {
+      "epoch": 1.3697104677060135,
+      "grad_norm": 1.107053279876709,
+      "learning_rate": 2.6094763933627794e-05,
+      "loss": 0.5168,
+      "step": 615
+    },
+    {
+      "epoch": 1.3808463251670378,
+      "grad_norm": 1.0395705699920654,
+      "learning_rate": 2.602008744564489e-05,
+      "loss": 0.482,
+      "step": 620
+    },
+    {
+      "epoch": 1.3919821826280623,
+      "grad_norm": 0.9968161582946777,
+      "learning_rate": 2.594481275487854e-05,
+      "loss": 0.5275,
+      "step": 625
+    },
+    {
+      "epoch": 1.4031180400890868,
+      "grad_norm": 1.1658473014831543,
+      "learning_rate": 2.5868943947460514e-05,
+      "loss": 0.5031,
+      "step": 630
+    },
+    {
+      "epoch": 1.4142538975501113,
+      "grad_norm": 1.0978684425354004,
+      "learning_rate": 2.5792485141773007e-05,
+      "loss": 0.5183,
+      "step": 635
+    },
+    {
+      "epoch": 1.4253897550111359,
+      "grad_norm": 1.0857203006744385,
+      "learning_rate": 2.5715440488225028e-05,
+      "loss": 0.4841,
+      "step": 640
+    },
+    {
+      "epoch": 1.4365256124721604,
+      "grad_norm": 1.1584542989730835,
+      "learning_rate": 2.5637814169027158e-05,
+      "loss": 0.4944,
+      "step": 645
+    },
+    {
+      "epoch": 1.447661469933185,
+      "grad_norm": 1.1438153982162476,
+      "learning_rate": 2.555961039796449e-05,
+      "loss": 0.4495,
+      "step": 650
+    },
+    {
+      "epoch": 1.4587973273942094,
+      "grad_norm": 1.2044782638549805,
+      "learning_rate": 2.5480833420167904e-05,
+      "loss": 0.4588,
+      "step": 655
+    },
+    {
+      "epoch": 1.469933184855234,
+      "grad_norm": 1.299019694328308,
+      "learning_rate": 2.5401487511883628e-05,
+      "loss": 0.5053,
+      "step": 660
+    },
+    {
+      "epoch": 1.4810690423162582,
+      "grad_norm": 0.9981383681297302,
+      "learning_rate": 2.5321576980241115e-05,
+      "loss": 0.525,
+      "step": 665
+    },
+    {
+      "epoch": 1.492204899777283,
+      "grad_norm": 1.142288088798523,
+      "learning_rate": 2.5241106163019227e-05,
+      "loss": 0.5095,
+      "step": 670
+    },
+    {
+      "epoch": 1.5033407572383073,
+      "grad_norm": 1.0118094682693481,
+      "learning_rate": 2.5160079428410795e-05,
+      "loss": 0.4554,
+      "step": 675
+    },
+    {
+      "epoch": 1.5144766146993318,
+      "grad_norm": 1.1396214962005615,
+      "learning_rate": 2.507850117478547e-05,
+      "loss": 0.4246,
+      "step": 680
+    },
+    {
+      "epoch": 1.5256124721603563,
+      "grad_norm": 0.9885154962539673,
+      "learning_rate": 2.4996375830450964e-05,
+      "loss": 0.4702,
+      "step": 685
+    },
+    {
+      "epoch": 1.5367483296213809,
+      "grad_norm": 1.0633879899978638,
+      "learning_rate": 2.4913707853412712e-05,
+      "loss": 0.4636,
+      "step": 690
+    },
+    {
+      "epoch": 1.5478841870824054,
+      "grad_norm": 1.0840256214141846,
+      "learning_rate": 2.483050173113183e-05,
+      "loss": 0.4692,
+      "step": 695
+    },
+    {
+      "epoch": 1.5590200445434297,
+      "grad_norm": 1.0847933292388916,
+      "learning_rate": 2.4746761980281552e-05,
+      "loss": 0.4398,
+      "step": 700
+    },
+    {
+      "epoch": 1.5701559020044544,
+      "grad_norm": 1.1708725690841675,
+      "learning_rate": 2.466249314650203e-05,
+      "loss": 0.447,
+      "step": 705
+    },
+    {
+      "epoch": 1.5812917594654787,
+      "grad_norm": 1.2069448232650757,
+      "learning_rate": 2.4577699804153603e-05,
+      "loss": 0.4652,
+      "step": 710
+    },
+    {
+      "epoch": 1.5924276169265035,
+      "grad_norm": 1.042110562324524,
+      "learning_rate": 2.4492386556068478e-05,
+      "loss": 0.4583,
+      "step": 715
+    },
+    {
+      "epoch": 1.6035634743875278,
+      "grad_norm": 1.1450542211532593,
+      "learning_rate": 2.440655803330088e-05,
+      "loss": 0.4379,
+      "step": 720
+    },
+    {
+      "epoch": 1.6146993318485523,
+      "grad_norm": 1.1880165338516235,
+      "learning_rate": 2.432021889487565e-05,
+      "loss": 0.4419,
+      "step": 725
+    },
+    {
+      "epoch": 1.6258351893095768,
+      "grad_norm": 1.0804628133773804,
+      "learning_rate": 2.4233373827535354e-05,
+      "loss": 0.4545,
+      "step": 730
+    },
+    {
+      "epoch": 1.6369710467706013,
+      "grad_norm": 1.0696831941604614,
+      "learning_rate": 2.414602754548587e-05,
+      "loss": 0.4793,
+      "step": 735
+    },
+    {
+      "epoch": 1.6481069042316259,
+      "grad_norm": 1.0015273094177246,
+      "learning_rate": 2.405818479014049e-05,
+      "loss": 0.4868,
+      "step": 740
+    },
+    {
+      "epoch": 1.6592427616926502,
+      "grad_norm": 1.0753366947174072,
+      "learning_rate": 2.3969850329862533e-05,
+      "loss": 0.4124,
+      "step": 745
+    },
+    {
+      "epoch": 1.670378619153675,
+      "grad_norm": 1.079838514328003,
+      "learning_rate": 2.3881028959706507e-05,
+      "loss": 0.4305,
+      "step": 750
+    },
+    {
+      "epoch": 1.6815144766146992,
+      "grad_norm": 1.1108442544937134,
+      "learning_rate": 2.3791725501157824e-05,
+      "loss": 0.4135,
+      "step": 755
+    },
+    {
+      "epoch": 1.692650334075724,
+      "grad_norm": 1.0079882144927979,
+      "learning_rate": 2.3701944801871073e-05,
+      "loss": 0.4424,
+      "step": 760
+    },
+    {
+      "epoch": 1.7037861915367483,
+      "grad_norm": 1.353743553161621,
+      "learning_rate": 2.3611691735406878e-05,
+      "loss": 0.4547,
+      "step": 765
+    },
+    {
+      "epoch": 1.7149220489977728,
+      "grad_norm": 1.1483458280563354,
+      "learning_rate": 2.3520971200967337e-05,
+      "loss": 0.4032,
+      "step": 770
+    },
+    {
+      "epoch": 1.7260579064587973,
+      "grad_norm": 1.1146272420883179,
+      "learning_rate": 2.342978812313009e-05,
+      "loss": 0.416,
+      "step": 775
+    },
+    {
+      "epoch": 1.7371937639198218,
+      "grad_norm": 1.1441491842269897,
+      "learning_rate": 2.3338147451580997e-05,
+      "loss": 0.4376,
+      "step": 780
+    },
+    {
+      "epoch": 1.7483296213808464,
+      "grad_norm": 1.0915056467056274,
+      "learning_rate": 2.324605416084544e-05,
+      "loss": 0.4071,
+      "step": 785
+    },
+    {
+      "epoch": 1.7594654788418709,
+      "grad_norm": 1.3740445375442505,
+      "learning_rate": 2.315351325001832e-05,
+      "loss": 0.422,
+      "step": 790
+    },
+    {
+      "epoch": 1.7706013363028954,
+      "grad_norm": 1.111710786819458,
+      "learning_rate": 2.3060529742492656e-05,
+      "loss": 0.3774,
+      "step": 795
+    },
+    {
+      "epoch": 1.7817371937639197,
+      "grad_norm": 1.1952409744262695,
+      "learning_rate": 2.2967108685686934e-05,
+      "loss": 0.4169,
+      "step": 800
+    },
+    {
+      "epoch": 1.7928730512249444,
+      "grad_norm": 1.2465108633041382,
+      "learning_rate": 2.2873255150771096e-05,
+      "loss": 0.3993,
+      "step": 805
+    },
+    {
+      "epoch": 1.8040089086859687,
+      "grad_norm": 1.1177217960357666,
+      "learning_rate": 2.277897423239126e-05,
+      "loss": 0.4199,
+      "step": 810
+    },
+    {
+      "epoch": 1.8151447661469933,
+      "grad_norm": 1.0628764629364014,
+      "learning_rate": 2.2684271048393194e-05,
+      "loss": 0.388,
+      "step": 815
+    },
+    {
+      "epoch": 1.8262806236080178,
+      "grad_norm": 1.1902976036071777,
+      "learning_rate": 2.2589150739544477e-05,
+      "loss": 0.3985,
+      "step": 820
+    },
+    {
+      "epoch": 1.8374164810690423,
+      "grad_norm": 1.1216514110565186,
+      "learning_rate": 2.2493618469255453e-05,
+      "loss": 0.379,
+      "step": 825
+    },
+    {
+      "epoch": 1.8485523385300668,
+      "grad_norm": 1.1403833627700806,
+      "learning_rate": 2.2397679423298945e-05,
+      "loss": 0.4247,
+      "step": 830
+    },
+    {
+      "epoch": 1.8596881959910914,
+      "grad_norm": 1.1960326433181763,
+      "learning_rate": 2.230133880952876e-05,
+      "loss": 0.3967,
+      "step": 835
+    },
+    {
+      "epoch": 1.8708240534521159,
+      "grad_norm": 1.2178454399108887,
+      "learning_rate": 2.2204601857596987e-05,
+      "loss": 0.4054,
+      "step": 840
+    },
+    {
+      "epoch": 1.8819599109131402,
+      "grad_norm": 1.148664116859436,
+      "learning_rate": 2.2107473818670107e-05,
+      "loss": 0.3996,
+      "step": 845
+    },
+    {
+      "epoch": 1.893095768374165,
+      "grad_norm": 1.2493960857391357,
+      "learning_rate": 2.2009959965143972e-05,
+      "loss": 0.4157,
+      "step": 850
+    },
+    {
+      "epoch": 1.9042316258351892,
+      "grad_norm": 1.1501412391662598,
+      "learning_rate": 2.1912065590357574e-05,
+      "loss": 0.4131,
+      "step": 855
+    },
+    {
+      "epoch": 1.9153674832962138,
+      "grad_norm": 1.0085930824279785,
+      "learning_rate": 2.1813796008305724e-05,
+      "loss": 0.4247,
+      "step": 860
+    },
+    {
+      "epoch": 1.9265033407572383,
+      "grad_norm": 1.2748278379440308,
+      "learning_rate": 2.1715156553350576e-05,
+      "loss": 0.4016,
+      "step": 865
+    },
+    {
+      "epoch": 1.9376391982182628,
+      "grad_norm": 1.1798368692398071,
+      "learning_rate": 2.1616152579932095e-05,
+      "loss": 0.3665,
+      "step": 870
+    },
+    {
+      "epoch": 1.9487750556792873,
+      "grad_norm": 1.21760892868042,
+      "learning_rate": 2.1516789462277372e-05,
+      "loss": 0.367,
+      "step": 875
+    },
+    {
+      "epoch": 1.9599109131403119,
+      "grad_norm": 1.129085898399353,
+      "learning_rate": 2.1417072594108894e-05,
+      "loss": 0.3782,
+      "step": 880
+    },
+    {
+      "epoch": 1.9710467706013364,
+      "grad_norm": 1.136759877204895,
+      "learning_rate": 2.1317007388351774e-05,
+      "loss": 0.404,
+      "step": 885
+    },
+    {
+      "epoch": 1.9821826280623607,
+      "grad_norm": 1.1734161376953125,
+      "learning_rate": 2.1216599276839922e-05,
+      "loss": 0.3595,
+      "step": 890
+    },
+    {
+      "epoch": 1.9933184855233854,
+      "grad_norm": 1.1954755783081055,
+      "learning_rate": 2.1115853710021174e-05,
+      "loss": 0.3842,
+      "step": 895
+    },
+    {
+      "epoch": 2.0044543429844097,
+      "grad_norm": 1.0263651609420776,
+      "learning_rate": 2.101477615666143e-05,
+      "loss": 0.3263,
+      "step": 900
+    },
+    {
+      "epoch": 2.0155902004454345,
+      "grad_norm": 1.2437479496002197,
+      "learning_rate": 2.0913372103547805e-05,
+      "loss": 0.3192,
+      "step": 905
+    },
+    {
+      "epoch": 2.0267260579064588,
+      "grad_norm": 1.1094865798950195,
+      "learning_rate": 2.081164705519078e-05,
+      "loss": 0.3165,
+      "step": 910
+    },
+    {
+      "epoch": 2.0378619153674835,
+      "grad_norm": 1.2660256624221802,
+      "learning_rate": 2.07096065335254e-05,
+      "loss": 0.2691,
+      "step": 915
+    },
+    {
+      "epoch": 2.048997772828508,
+      "grad_norm": 1.1449397802352905,
+      "learning_rate": 2.0607256077611533e-05,
+      "loss": 0.2808,
+      "step": 920
+    },
+    {
+      "epoch": 2.060133630289532,
+      "grad_norm": 1.2465488910675049,
+      "learning_rate": 2.0504601243333182e-05,
+      "loss": 0.2933,
+      "step": 925
+    },
+    {
+      "epoch": 2.071269487750557,
+      "grad_norm": 1.0938855409622192,
+      "learning_rate": 2.040164760309691e-05,
+      "loss": 0.2703,
+      "step": 930
+    },
+    {
+      "epoch": 2.082405345211581,
+      "grad_norm": 1.1405394077301025,
+      "learning_rate": 2.0298400745529352e-05,
+      "loss": 0.2926,
+      "step": 935
+    },
+    {
+      "epoch": 2.093541202672606,
+      "grad_norm": 1.2811548709869385,
+      "learning_rate": 2.0194866275173827e-05,
+      "loss": 0.2965,
+      "step": 940
+    },
+    {
+      "epoch": 2.10467706013363,
+      "grad_norm": 1.2539308071136475,
+      "learning_rate": 2.0091049812186146e-05,
+      "loss": 0.3216,
+      "step": 945
+    },
+    {
+      "epoch": 2.115812917594655,
+      "grad_norm": 1.1019704341888428,
+      "learning_rate": 1.9986956992029487e-05,
+      "loss": 0.2967,
+      "step": 950
+    },
+    {
+      "epoch": 2.1269487750556793,
+      "grad_norm": 1.0621236562728882,
+      "learning_rate": 1.9882593465168522e-05,
+      "loss": 0.2511,
+      "step": 955
+    },
+    {
+      "epoch": 2.138084632516704,
+      "grad_norm": 1.1347256898880005,
+      "learning_rate": 1.9777964896762676e-05,
+      "loss": 0.3121,
+      "step": 960
+    },
+    {
+      "epoch": 2.1492204899777283,
+      "grad_norm": 1.0914628505706787,
+      "learning_rate": 1.9673076966358606e-05,
+      "loss": 0.249,
+      "step": 965
+    },
+    {
+      "epoch": 2.1603563474387526,
+      "grad_norm": 1.3239182233810425,
+      "learning_rate": 1.9567935367581913e-05,
+      "loss": 0.2597,
+      "step": 970
+    },
+    {
+      "epoch": 2.1714922048997773,
+      "grad_norm": 1.1536415815353394,
+      "learning_rate": 1.9462545807828046e-05,
+      "loss": 0.2602,
+      "step": 975
+    },
+    {
+      "epoch": 2.1826280623608016,
+      "grad_norm": 1.1208415031433105,
+      "learning_rate": 1.935691400795252e-05,
+      "loss": 0.2929,
+      "step": 980
+    },
+    {
+      "epoch": 2.1937639198218264,
+      "grad_norm": 1.0744553804397583,
+      "learning_rate": 1.925104570196036e-05,
+      "loss": 0.2335,
+      "step": 985
+    },
+    {
+      "epoch": 2.2048997772828507,
+      "grad_norm": 1.1932042837142944,
+      "learning_rate": 1.914494663669482e-05,
+      "loss": 0.3014,
+      "step": 990
+    },
+    {
+      "epoch": 2.2160356347438754,
+      "grad_norm": 1.2387852668762207,
+      "learning_rate": 1.9038622571525464e-05,
+      "loss": 0.2754,
+      "step": 995
+    },
+    {
+      "epoch": 2.2271714922048997,
+      "grad_norm": 1.1342064142227173,
+      "learning_rate": 1.8932079278035503e-05,
+      "loss": 0.2704,
+      "step": 1000
+    },
+    {
+      "epoch": 2.2383073496659245,
+      "grad_norm": 1.0542978048324585,
+      "learning_rate": 1.882532253970852e-05,
+      "loss": 0.2913,
+      "step": 1005
+    },
+    {
+      "epoch": 2.249443207126949,
+      "grad_norm": 1.4599226713180542,
+      "learning_rate": 1.871835815161449e-05,
+      "loss": 0.3044,
+      "step": 1010
+    },
+    {
+      "epoch": 2.260579064587973,
+      "grad_norm": 1.1367197036743164,
+      "learning_rate": 1.861119192009526e-05,
+      "loss": 0.2821,
+      "step": 1015
+    },
+    {
+      "epoch": 2.271714922048998,
+      "grad_norm": 1.1666250228881836,
+      "learning_rate": 1.850382966244929e-05,
+      "loss": 0.2834,
+      "step": 1020
+    },
+    {
+      "epoch": 2.282850779510022,
+      "grad_norm": 1.2641099691390991,
+      "learning_rate": 1.8396277206615953e-05,
+      "loss": 0.2586,
+      "step": 1025
+    },
+    {
+      "epoch": 2.293986636971047,
+      "grad_norm": 1.1135034561157227,
+      "learning_rate": 1.8288540390859124e-05,
+      "loss": 0.2727,
+      "step": 1030
+    },
+    {
+      "epoch": 2.305122494432071,
+      "grad_norm": 1.1536855697631836,
+      "learning_rate": 1.8180625063450278e-05,
+      "loss": 0.2767,
+      "step": 1035
+    },
+    {
+      "epoch": 2.316258351893096,
+      "grad_norm": 1.176110029220581,
+      "learning_rate": 1.8072537082351028e-05,
+      "loss": 0.2583,
+      "step": 1040
+    },
+    {
+      "epoch": 2.3273942093541202,
+      "grad_norm": 1.2295929193496704,
+      "learning_rate": 1.796428231489514e-05,
+      "loss": 0.2559,
+      "step": 1045
+    },
+    {
+      "epoch": 2.338530066815145,
+      "grad_norm": 1.1238468885421753,
+      "learning_rate": 1.7855866637470027e-05,
+      "loss": 0.2523,
+      "step": 1050
+    },
+    {
+      "epoch": 2.3496659242761693,
+      "grad_norm": 1.1042718887329102,
+      "learning_rate": 1.7747295935197776e-05,
+      "loss": 0.2766,
+      "step": 1055
+    },
+    {
+      "epoch": 2.3608017817371936,
+      "grad_norm": 1.2366911172866821,
+      "learning_rate": 1.7638576101615678e-05,
+      "loss": 0.2461,
+      "step": 1060
+    },
+    {
+      "epoch": 2.3719376391982183,
+      "grad_norm": 1.1563042402267456,
+      "learning_rate": 1.7529713038356313e-05,
+      "loss": 0.2362,
+      "step": 1065
+    },
+    {
+      "epoch": 2.3830734966592426,
+      "grad_norm": 1.2360973358154297,
+      "learning_rate": 1.7420712654827182e-05,
+      "loss": 0.2508,
+      "step": 1070
+    },
+    {
+      "epoch": 2.3942093541202674,
+      "grad_norm": 1.1910051107406616,
+      "learning_rate": 1.7311580867889935e-05,
+      "loss": 0.2282,
+      "step": 1075
+    },
+    {
+      "epoch": 2.4053452115812917,
+      "grad_norm": 1.0605477094650269,
+      "learning_rate": 1.7202323601539197e-05,
+      "loss": 0.223,
+      "step": 1080
+    },
+    {
+      "epoch": 2.4164810690423164,
+      "grad_norm": 1.1978503465652466,
+      "learning_rate": 1.7092946786580964e-05,
+      "loss": 0.262,
+      "step": 1085
+    },
+    {
+      "epoch": 2.4276169265033407,
+      "grad_norm": 1.2242716550827026,
+      "learning_rate": 1.6983456360310716e-05,
+      "loss": 0.2493,
+      "step": 1090
+    },
+    {
+      "epoch": 2.4387527839643655,
+      "grad_norm": 1.0966243743896484,
+      "learning_rate": 1.687385826619105e-05,
+      "loss": 0.2252,
+      "step": 1095
+    },
+    {
+      "epoch": 2.4498886414253898,
+      "grad_norm": 1.2636704444885254,
+      "learning_rate": 1.6764158453529124e-05,
+      "loss": 0.2501,
+      "step": 1100
+    },
+    {
+      "epoch": 2.461024498886414,
+      "grad_norm": 1.1845978498458862,
+      "learning_rate": 1.665436287715365e-05,
+      "loss": 0.246,
+      "step": 1105
+    },
+    {
+      "epoch": 2.472160356347439,
+      "grad_norm": 0.9880180954933167,
+      "learning_rate": 1.6544477497091698e-05,
+      "loss": 0.236,
+      "step": 1110
+    },
+    {
+      "epoch": 2.483296213808463,
+      "grad_norm": 1.1788389682769775,
+      "learning_rate": 1.6434508278245134e-05,
+      "loss": 0.2377,
+      "step": 1115
+    },
+    {
+      "epoch": 2.494432071269488,
+      "grad_norm": 1.3312638998031616,
+      "learning_rate": 1.632446119006685e-05,
+      "loss": 0.2576,
+      "step": 1120
+    },
+    {
+      "epoch": 2.505567928730512,
+      "grad_norm": 1.2232873439788818,
+      "learning_rate": 1.6214342206236703e-05,
+      "loss": 0.2258,
+      "step": 1125
+    },
+    {
+      "epoch": 2.516703786191537,
+      "grad_norm": 1.1207743883132935,
+      "learning_rate": 1.6104157304337264e-05,
+      "loss": 0.2096,
+      "step": 1130
+    },
+    {
+      "epoch": 2.527839643652561,
+      "grad_norm": 1.002036690711975,
+      "learning_rate": 1.5993912465529337e-05,
+      "loss": 0.2574,
+      "step": 1135
+    },
+    {
+      "epoch": 2.538975501113586,
+      "grad_norm": 1.3266305923461914,
+      "learning_rate": 1.588361367422726e-05,
+      "loss": 0.2267,
+      "step": 1140
+    },
+    {
+      "epoch": 2.5501113585746102,
+      "grad_norm": 1.281622290611267,
+      "learning_rate": 1.5773266917774087e-05,
+      "loss": 0.2529,
+      "step": 1145
+    },
+    {
+      "epoch": 2.5612472160356345,
+      "grad_norm": 1.1964174509048462,
+      "learning_rate": 1.5662878186116556e-05,
+      "loss": 0.2393,
+      "step": 1150
+    },
+    {
+      "epoch": 2.5723830734966593,
+      "grad_norm": 1.0871187448501587,
+      "learning_rate": 1.555245347147995e-05,
+      "loss": 0.2623,
+      "step": 1155
+    },
+    {
+      "epoch": 2.5835189309576836,
+      "grad_norm": 1.1345094442367554,
+      "learning_rate": 1.54419987680428e-05,
+      "loss": 0.2343,
+      "step": 1160
+    },
+    {
+      "epoch": 2.5946547884187083,
+      "grad_norm": 1.1741135120391846,
+      "learning_rate": 1.533152007161152e-05,
+      "loss": 0.231,
+      "step": 1165
+    },
+    {
+      "epoch": 2.6057906458797326,
+      "grad_norm": 1.1888442039489746,
+      "learning_rate": 1.5221023379294952e-05,
+      "loss": 0.2437,
+      "step": 1170
+    },
+    {
+      "epoch": 2.6169265033407574,
+      "grad_norm": 1.1484440565109253,
+      "learning_rate": 1.5110514689178788e-05,
+      "loss": 0.238,
+      "step": 1175
+    },
+    {
+      "epoch": 2.6280623608017817,
+      "grad_norm": 1.2487529516220093,
+      "learning_rate": 1.5e-05,
+      "loss": 0.2416,
+      "step": 1180
+    },
+    {
+      "epoch": 2.6391982182628064,
+      "grad_norm": 1.165150761604309,
+      "learning_rate": 1.4889485310821213e-05,
+      "loss": 0.2356,
+      "step": 1185
+    },
+    {
+      "epoch": 2.6503340757238307,
+      "grad_norm": 1.1523540019989014,
+      "learning_rate": 1.477897662070505e-05,
+      "loss": 0.2144,
+      "step": 1190
+    },
+    {
+      "epoch": 2.661469933184855,
+      "grad_norm": 1.0735657215118408,
+      "learning_rate": 1.4668479928388478e-05,
+      "loss": 0.2212,
+      "step": 1195
+    },
+    {
+      "epoch": 2.6726057906458798,
+      "grad_norm": 1.110089898109436,
+      "learning_rate": 1.4558001231957202e-05,
+      "loss": 0.21,
+      "step": 1200
+    },
+    {
+      "epoch": 2.683741648106904,
+      "grad_norm": 1.2104490995407104,
+      "learning_rate": 1.444754652852005e-05,
+      "loss": 0.2157,
+      "step": 1205
+    },
+    {
+      "epoch": 2.694877505567929,
+      "grad_norm": 1.134554386138916,
+      "learning_rate": 1.4337121813883443e-05,
+      "loss": 0.2285,
+      "step": 1210
+    },
+    {
+      "epoch": 2.706013363028953,
+      "grad_norm": 1.0749125480651855,
+      "learning_rate": 1.4226733082225917e-05,
+      "loss": 0.2279,
+      "step": 1215
+    },
+    {
+      "epoch": 2.717149220489978,
+      "grad_norm": 1.1136120557785034,
+      "learning_rate": 1.4116386325772745e-05,
+      "loss": 0.2124,
+      "step": 1220
+    },
+    {
+      "epoch": 2.728285077951002,
+      "grad_norm": 1.1267787218093872,
+      "learning_rate": 1.4006087534470667e-05,
+      "loss": 0.2265,
+      "step": 1225
+    },
+    {
+      "epoch": 2.739420935412027,
+      "grad_norm": 1.248026728630066,
+      "learning_rate": 1.3895842695662738e-05,
+      "loss": 0.2335,
+      "step": 1230
+    },
+    {
+      "epoch": 2.750556792873051,
+      "grad_norm": 1.3174666166305542,
+      "learning_rate": 1.37856577937633e-05,
+      "loss": 0.2214,
+      "step": 1235
+    },
+    {
+      "epoch": 2.7616926503340755,
+      "grad_norm": 0.9793177247047424,
+      "learning_rate": 1.367553880993315e-05,
+      "loss": 0.1913,
+      "step": 1240
+    },
+    {
+      "epoch": 2.7728285077951003,
+      "grad_norm": 1.211518406867981,
+      "learning_rate": 1.3565491721754868e-05,
+      "loss": 0.201,
+      "step": 1245
+    },
+    {
+      "epoch": 2.7839643652561246,
+      "grad_norm": 1.1160845756530762,
+      "learning_rate": 1.3455522502908305e-05,
+      "loss": 0.2075,
+      "step": 1250
+    },
+    {
+      "epoch": 2.7951002227171493,
+      "grad_norm": 1.0458290576934814,
+      "learning_rate": 1.334563712284635e-05,
+      "loss": 0.2067,
+      "step": 1255
+    },
+    {
+      "epoch": 2.8062360801781736,
+      "grad_norm": 1.0362893342971802,
+      "learning_rate": 1.3235841546470884e-05,
+      "loss": 0.1904,
+      "step": 1260
+    },
+    {
+      "epoch": 2.8173719376391984,
+      "grad_norm": 1.207885980606079,
+      "learning_rate": 1.312614173380895e-05,
+      "loss": 0.1989,
+      "step": 1265
+    },
+    {
+      "epoch": 2.8285077951002227,
+      "grad_norm": 1.2045843601226807,
+      "learning_rate": 1.301654363968929e-05,
+      "loss": 0.2144,
+      "step": 1270
+    },
+    {
+      "epoch": 2.8396436525612474,
+      "grad_norm": 1.1977723836898804,
+      "learning_rate": 1.2907053213419035e-05,
+      "loss": 0.2294,
+      "step": 1275
+    },
+    {
+      "epoch": 2.8507795100222717,
+      "grad_norm": 1.1156507730484009,
+      "learning_rate": 1.2797676398460808e-05,
+      "loss": 0.1677,
+      "step": 1280
+    },
+    {
+      "epoch": 2.861915367483296,
+      "grad_norm": 1.168018102645874,
+      "learning_rate": 1.268841913211007e-05,
+      "loss": 0.2166,
+      "step": 1285
+    },
+    {
+      "epoch": 2.8730512249443207,
+      "grad_norm": 1.1159573793411255,
+      "learning_rate": 1.2579287345172824e-05,
+      "loss": 0.19,
+      "step": 1290
+    },
+    {
+      "epoch": 2.884187082405345,
+      "grad_norm": 1.0215970277786255,
+      "learning_rate": 1.2470286961643688e-05,
+      "loss": 0.1599,
+      "step": 1295
+    },
+    {
+      "epoch": 2.89532293986637,
+      "grad_norm": 1.1599113941192627,
+      "learning_rate": 1.2361423898384323e-05,
+      "loss": 0.2201,
+      "step": 1300
+    },
+    {
+      "epoch": 2.906458797327394,
+      "grad_norm": 1.1932123899459839,
+      "learning_rate": 1.2252704064802225e-05,
+      "loss": 0.1998,
+      "step": 1305
+    },
+    {
+      "epoch": 2.917594654788419,
+      "grad_norm": 0.9487115740776062,
+      "learning_rate": 1.2144133362529972e-05,
+      "loss": 0.2255,
+      "step": 1310
+    },
+    {
+      "epoch": 2.928730512249443,
+      "grad_norm": 1.045422077178955,
+      "learning_rate": 1.2035717685104866e-05,
+      "loss": 0.1796,
+      "step": 1315
+    },
+    {
+      "epoch": 2.939866369710468,
+      "grad_norm": 1.1371259689331055,
+      "learning_rate": 1.1927462917648973e-05,
+      "loss": 0.1721,
+      "step": 1320
+    },
+    {
+      "epoch": 2.951002227171492,
+      "grad_norm": 1.2436094284057617,
+      "learning_rate": 1.1819374936549723e-05,
+      "loss": 0.1794,
+      "step": 1325
+    },
+    {
+      "epoch": 2.9621380846325165,
+      "grad_norm": 1.2351981401443481,
+      "learning_rate": 1.1711459609140875e-05,
+      "loss": 0.2143,
+      "step": 1330
+    },
+    {
+      "epoch": 2.9732739420935412,
+      "grad_norm": 0.973818302154541,
+      "learning_rate": 1.1603722793384046e-05,
+      "loss": 0.1637,
+      "step": 1335
+    },
+    {
+      "epoch": 2.984409799554566,
+      "grad_norm": 0.9122995138168335,
+      "learning_rate": 1.1496170337550713e-05,
+      "loss": 0.1779,
+      "step": 1340
+    },
+    {
+      "epoch": 2.9955456570155903,
+      "grad_norm": 1.0754543542861938,
+      "learning_rate": 1.1388808079904745e-05,
+      "loss": 0.1809,
+      "step": 1345
+    },
+    {
+      "epoch": 3.0066815144766146,
+      "grad_norm": 0.9397510290145874,
+      "learning_rate": 1.1281641848385506e-05,
+      "loss": 0.1664,
+      "step": 1350
+    },
+    {
+      "epoch": 3.0178173719376393,
+      "grad_norm": 1.090684175491333,
+      "learning_rate": 1.1174677460291483e-05,
+      "loss": 0.1635,
+      "step": 1355
+    },
+    {
+      "epoch": 3.0289532293986636,
+      "grad_norm": 1.0511411428451538,
+      "learning_rate": 1.1067920721964493e-05,
+      "loss": 0.1308,
+      "step": 1360
+    },
+    {
+      "epoch": 3.0400890868596884,
+      "grad_norm": 1.0632355213165283,
+      "learning_rate": 1.0961377428474542e-05,
+      "loss": 0.1424,
+      "step": 1365
+    },
+    {
+      "epoch": 3.0512249443207127,
+      "grad_norm": 1.0594873428344727,
+      "learning_rate": 1.0855053363305186e-05,
+      "loss": 0.1445,
+      "step": 1370
+    },
+    {
+      "epoch": 3.062360801781737,
+      "grad_norm": 1.0620012283325195,
+      "learning_rate": 1.0748954298039644e-05,
+      "loss": 0.1341,
+      "step": 1375
+    },
+    {
+      "epoch": 3.0734966592427617,
+      "grad_norm": 1.1539349555969238,
+      "learning_rate": 1.0643085992047482e-05,
+      "loss": 0.1655,
+      "step": 1380
+    },
+    {
+      "epoch": 3.084632516703786,
+      "grad_norm": 0.9552104473114014,
+      "learning_rate": 1.0537454192171958e-05,
+      "loss": 0.1428,
+      "step": 1385
+    },
+    {
+      "epoch": 3.0957683741648108,
+      "grad_norm": 1.133872389793396,
+      "learning_rate": 1.0432064632418091e-05,
+      "loss": 0.1383,
+      "step": 1390
+    },
+    {
+      "epoch": 3.106904231625835,
+      "grad_norm": 0.9998437166213989,
+      "learning_rate": 1.0326923033641396e-05,
+      "loss": 0.1517,
+      "step": 1395
+    },
+    {
+      "epoch": 3.11804008908686,
+      "grad_norm": 1.0297425985336304,
+      "learning_rate": 1.0222035103237326e-05,
+      "loss": 0.1675,
+      "step": 1400
+    },
+    {
+      "epoch": 3.129175946547884,
+      "grad_norm": 0.9318389892578125,
+      "learning_rate": 1.011740653483148e-05,
+      "loss": 0.1182,
+      "step": 1405
+    },
+    {
+      "epoch": 3.140311804008909,
+      "grad_norm": 0.9119312763214111,
+      "learning_rate": 1.0013043007970515e-05,
+      "loss": 0.1321,
+      "step": 1410
+    },
+    {
+      "epoch": 3.151447661469933,
+      "grad_norm": 1.0865997076034546,
+      "learning_rate": 9.908950187813855e-06,
+      "loss": 0.1724,
+      "step": 1415
+    },
+    {
+      "epoch": 3.1625835189309575,
+      "grad_norm": 1.0325827598571777,
+      "learning_rate": 9.805133724826175e-06,
+      "loss": 0.1511,
+      "step": 1420
+    },
+    {
+      "epoch": 3.173719376391982,
+      "grad_norm": 0.9540656805038452,
+      "learning_rate": 9.701599254470653e-06,
+      "loss": 0.1184,
+      "step": 1425
+    },
+    {
+      "epoch": 3.1848552338530065,
+      "grad_norm": 1.1433629989624023,
+      "learning_rate": 9.598352396903091e-06,
+      "loss": 0.1786,
+      "step": 1430
+    },
+    {
+      "epoch": 3.1959910913140313,
+      "grad_norm": 1.1182628870010376,
+      "learning_rate": 9.495398756666824e-06,
+      "loss": 0.1278,
+      "step": 1435
+    },
+    {
+      "epoch": 3.2071269487750556,
+      "grad_norm": 1.02802574634552,
+      "learning_rate": 9.39274392238847e-06,
+      "loss": 0.1318,
+      "step": 1440
+    },
+    {
+      "epoch": 3.2182628062360803,
+      "grad_norm": 1.3277000188827515,
+      "learning_rate": 9.290393466474597e-06,
+      "loss": 0.1325,
+      "step": 1445
+    },
+    {
+      "epoch": 3.2293986636971046,
+      "grad_norm": 1.0262706279754639,
+      "learning_rate": 9.18835294480922e-06,
+      "loss": 0.1565,
+      "step": 1450
+    },
+    {
+      "epoch": 3.2405345211581293,
+      "grad_norm": 1.1255269050598145,
+      "learning_rate": 9.086627896452194e-06,
+      "loss": 0.149,
+      "step": 1455
+    },
+    {
+      "epoch": 3.2516703786191536,
+      "grad_norm": 1.312609314918518,
+      "learning_rate": 8.98522384333857e-06,
+      "loss": 0.1464,
+      "step": 1460
+    },
+    {
+      "epoch": 3.262806236080178,
+      "grad_norm": 1.094075322151184,
+      "learning_rate": 8.884146289978827e-06,
+      "loss": 0.1423,
+      "step": 1465
+    },
+    {
+      "epoch": 3.2739420935412027,
+      "grad_norm": 1.063836932182312,
+      "learning_rate": 8.783400723160082e-06,
+      "loss": 0.1521,
+      "step": 1470
+    },
+    {
+      "epoch": 3.285077951002227,
+      "grad_norm": 0.8800683617591858,
+      "learning_rate": 8.682992611648228e-06,
+      "loss": 0.1442,
+      "step": 1475
+    },
+    {
+      "epoch": 3.2962138084632517,
+      "grad_norm": 0.9332960247993469,
+      "learning_rate": 8.58292740589111e-06,
+      "loss": 0.1436,
+      "step": 1480
+    },
+    {
+      "epoch": 3.307349665924276,
+      "grad_norm": 0.982293963432312,
+      "learning_rate": 8.483210537722635e-06,
+      "loss": 0.1089,
+      "step": 1485
+    },
+    {
+      "epoch": 3.318485523385301,
+      "grad_norm": 1.0682051181793213,
+      "learning_rate": 8.3838474200679e-06,
+      "loss": 0.1461,
+      "step": 1490
+    },
+    {
+      "epoch": 3.329621380846325,
+      "grad_norm": 1.1447734832763672,
+      "learning_rate": 8.284843446649425e-06,
+      "loss": 0.1467,
+      "step": 1495
+    },
+    {
+      "epoch": 3.34075723830735,
+      "grad_norm": 1.1350065469741821,
+      "learning_rate": 8.186203991694284e-06,
+      "loss": 0.1356,
+      "step": 1500
+    },
+    {
+      "epoch": 3.351893095768374,
+      "grad_norm": 1.1445928812026978,
+      "learning_rate": 8.087934409642427e-06,
+      "loss": 0.138,
+      "step": 1505
+    },
+    {
+      "epoch": 3.3630289532293984,
+      "grad_norm": 0.9445508122444153,
+      "learning_rate": 7.99004003485603e-06,
+      "loss": 0.1314,
+      "step": 1510
+    },
+    {
+      "epoch": 3.374164810690423,
+      "grad_norm": 1.0288410186767578,
+      "learning_rate": 7.892526181329895e-06,
+      "loss": 0.1456,
+      "step": 1515
+    },
+    {
+      "epoch": 3.3853006681514475,
+      "grad_norm": 1.1071178913116455,
+      "learning_rate": 7.79539814240302e-06,
+      "loss": 0.1278,
+      "step": 1520
+    },
+    {
+      "epoch": 3.3964365256124722,
+      "grad_norm": 1.169844627380371,
+      "learning_rate": 7.698661190471245e-06,
+      "loss": 0.1372,
+      "step": 1525
+    },
+    {
+      "epoch": 3.4075723830734965,
+      "grad_norm": 0.9744820594787598,
+      "learning_rate": 7.602320576701054e-06,
+      "loss": 0.1157,
+      "step": 1530
+    },
+    {
+      "epoch": 3.4187082405345213,
+      "grad_norm": 0.9256933331489563,
+      "learning_rate": 7.506381530744549e-06,
+      "loss": 0.1391,
+      "step": 1535
+    },
+    {
+      "epoch": 3.4298440979955456,
+      "grad_norm": 1.0349007844924927,
+      "learning_rate": 7.4108492604555265e-06,
+      "loss": 0.1423,
+      "step": 1540
+    },
+    {
+      "epoch": 3.4409799554565703,
+      "grad_norm": 1.0828795433044434,
+      "learning_rate": 7.315728951606806e-06,
+      "loss": 0.1415,
+      "step": 1545
+    },
+    {
+      "epoch": 3.4521158129175946,
+      "grad_norm": 1.1426892280578613,
+      "learning_rate": 7.221025767608743e-06,
+      "loss": 0.1379,
+      "step": 1550
+    },
+    {
+      "epoch": 3.463251670378619,
+      "grad_norm": 0.8297178745269775,
+      "learning_rate": 7.126744849228912e-06,
+      "loss": 0.1087,
+      "step": 1555
+    },
+    {
+      "epoch": 3.4743875278396437,
+      "grad_norm": 1.0585066080093384,
+      "learning_rate": 7.032891314313066e-06,
+      "loss": 0.1298,
+      "step": 1560
+    },
+    {
+      "epoch": 3.485523385300668,
+      "grad_norm": 0.9125014543533325,
+      "learning_rate": 6.939470257507345e-06,
+      "loss": 0.121,
+      "step": 1565
+    },
+    {
+      "epoch": 3.4966592427616927,
+      "grad_norm": 1.1120779514312744,
+      "learning_rate": 6.846486749981685e-06,
+      "loss": 0.1288,
+      "step": 1570
+    },
+    {
+      "epoch": 3.507795100222717,
+      "grad_norm": 1.0022464990615845,
+      "learning_rate": 6.753945839154564e-06,
+      "loss": 0.1186,
+      "step": 1575
+    },
+    {
+      "epoch": 3.5189309576837418,
+      "grad_norm": 0.9147719144821167,
+      "learning_rate": 6.661852548419011e-06,
+      "loss": 0.1141,
+      "step": 1580
+    },
+    {
+      "epoch": 3.530066815144766,
+      "grad_norm": 0.9256666302680969,
+      "learning_rate": 6.57021187686991e-06,
+      "loss": 0.1074,
+      "step": 1585
+    },
+    {
+      "epoch": 3.541202672605791,
+      "grad_norm": 1.149868369102478,
+      "learning_rate": 6.479028799032664e-06,
+      "loss": 0.114,
+      "step": 1590
+    },
+    {
+      "epoch": 3.552338530066815,
+      "grad_norm": 1.0971661806106567,
+      "learning_rate": 6.388308264593126e-06,
+      "loss": 0.1229,
+      "step": 1595
+    },
+    {
+      "epoch": 3.5634743875278394,
+      "grad_norm": 1.0946404933929443,
+      "learning_rate": 6.298055198128925e-06,
+      "loss": 0.1194,
+      "step": 1600
+    },
+    {
+      "epoch": 3.574610244988864,
+      "grad_norm": 0.9242075085639954,
+      "learning_rate": 6.208274498842178e-06,
+      "loss": 0.1281,
+      "step": 1605
+    },
+    {
+      "epoch": 3.585746102449889,
+      "grad_norm": 0.9597280621528625,
+      "learning_rate": 6.118971040293496e-06,
+      "loss": 0.1278,
+      "step": 1610
+    },
+    {
+      "epoch": 3.596881959910913,
+      "grad_norm": 1.0481159687042236,
+      "learning_rate": 6.030149670137466e-06,
+      "loss": 0.1322,
+      "step": 1615
+    },
+    {
+      "epoch": 3.6080178173719375,
+      "grad_norm": 0.9799613356590271,
+      "learning_rate": 5.941815209859509e-06,
+      "loss": 0.1034,
+      "step": 1620
+    },
+    {
+      "epoch": 3.6191536748329622,
+      "grad_norm": 0.8920042514801025,
+      "learning_rate": 5.853972454514131e-06,
+      "loss": 0.1243,
+      "step": 1625
+    },
+    {
+      "epoch": 3.6302895322939865,
+      "grad_norm": 0.9567716717720032,
+      "learning_rate": 5.7666261724646506e-06,
+      "loss": 0.1427,
+      "step": 1630
+    },
+    {
+      "epoch": 3.6414253897550113,
+      "grad_norm": 1.0656696557998657,
+      "learning_rate": 5.679781105124357e-06,
+      "loss": 0.1267,
+      "step": 1635
+    },
+    {
+      "epoch": 3.6525612472160356,
+      "grad_norm": 0.8879715204238892,
+      "learning_rate": 5.593441966699122e-06,
+      "loss": 0.111,
+      "step": 1640
+    },
+    {
+      "epoch": 3.66369710467706,
+      "grad_norm": 0.7971992492675781,
+      "learning_rate": 5.507613443931524e-06,
+      "loss": 0.1136,
+      "step": 1645
+    },
+    {
+      "epoch": 3.6748329621380846,
+      "grad_norm": 1.1888115406036377,
+      "learning_rate": 5.422300195846402e-06,
+      "loss": 0.1221,
+      "step": 1650
+    },
+    {
+      "epoch": 3.6859688195991094,
+      "grad_norm": 0.9997703433036804,
+      "learning_rate": 5.337506853497973e-06,
+      "loss": 0.1112,
+      "step": 1655
+    },
+    {
+      "epoch": 3.6971046770601337,
+      "grad_norm": 0.9232822060585022,
+      "learning_rate": 5.253238019718451e-06,
+      "loss": 0.1276,
+      "step": 1660
+    },
+    {
+      "epoch": 3.708240534521158,
+      "grad_norm": 1.1803860664367676,
+      "learning_rate": 5.169498268868173e-06,
+      "loss": 0.1085,
+      "step": 1665
+    },
+    {
+      "epoch": 3.7193763919821827,
+      "grad_norm": 1.0455310344696045,
+      "learning_rate": 5.086292146587288e-06,
+      "loss": 0.1314,
+      "step": 1670
+    },
+    {
+      "epoch": 3.730512249443207,
+      "grad_norm": 1.1193212270736694,
+      "learning_rate": 5.003624169549036e-06,
+      "loss": 0.1107,
+      "step": 1675
+    },
+    {
+      "epoch": 3.7416481069042318,
+      "grad_norm": 1.0344520807266235,
+      "learning_rate": 4.921498825214535e-06,
+      "loss": 0.1137,
+      "step": 1680
+    },
+    {
+      "epoch": 3.752783964365256,
+      "grad_norm": 0.8843588829040527,
+      "learning_rate": 4.8399205715892065e-06,
+      "loss": 0.0955,
+      "step": 1685
+    },
+    {
+      "epoch": 3.7639198218262804,
+      "grad_norm": 0.9066224098205566,
+      "learning_rate": 4.758893836980775e-06,
+      "loss": 0.1129,
+      "step": 1690
+    },
+    {
+      "epoch": 3.775055679287305,
+      "grad_norm": 0.9448865056037903,
+      "learning_rate": 4.678423019758888e-06,
+      "loss": 0.1141,
+      "step": 1695
+    },
+    {
+      "epoch": 3.78619153674833,
+      "grad_norm": 0.983690619468689,
+      "learning_rate": 4.598512488116376e-06,
+      "loss": 0.126,
+      "step": 1700
+    },
+    {
+      "epoch": 3.797327394209354,
+      "grad_norm": 0.9622001647949219,
+      "learning_rate": 4.5191665798321005e-06,
+      "loss": 0.1126,
+      "step": 1705
+    },
+    {
+      "epoch": 3.8084632516703785,
+      "grad_norm": 0.9829707741737366,
+      "learning_rate": 4.44038960203551e-06,
+      "loss": 0.1164,
+      "step": 1710
+    },
+    {
+      "epoch": 3.819599109131403,
+      "grad_norm": 0.8512528538703918,
+      "learning_rate": 4.362185830972843e-06,
+      "loss": 0.1192,
+      "step": 1715
+    },
+    {
+      "epoch": 3.8307349665924275,
+      "grad_norm": 0.9649842381477356,
+      "learning_rate": 4.284559511774976e-06,
+      "loss": 0.1228,
+      "step": 1720
+    },
+    {
+      "epoch": 3.8418708240534523,
+      "grad_norm": 1.135290265083313,
+      "learning_rate": 4.207514858226997e-06,
+      "loss": 0.1154,
+      "step": 1725
+    },
+    {
+      "epoch": 3.8530066815144766,
+      "grad_norm": 0.8738131523132324,
+      "learning_rate": 4.131056052539492e-06,
+      "loss": 0.0901,
+      "step": 1730
+    },
+    {
+      "epoch": 3.864142538975501,
+      "grad_norm": 0.888911783695221,
+      "learning_rate": 4.0551872451214645e-06,
+      "loss": 0.0981,
+      "step": 1735
+    },
+    {
+      "epoch": 3.8752783964365256,
+      "grad_norm": 1.0050368309020996,
+      "learning_rate": 3.979912554355113e-06,
+      "loss": 0.1193,
+      "step": 1740
+    },
+    {
+      "epoch": 3.8864142538975504,
+      "grad_norm": 0.8284539580345154,
+      "learning_rate": 3.905236066372211e-06,
+      "loss": 0.0959,
+      "step": 1745
+    },
+    {
+      "epoch": 3.8975501113585747,
+      "grad_norm": 0.8358965516090393,
+      "learning_rate": 3.831161834832328e-06,
+      "loss": 0.1093,
+      "step": 1750
+    },
+    {
+      "epoch": 3.908685968819599,
+      "grad_norm": 1.0052320957183838,
+      "learning_rate": 3.757693880702795e-06,
+      "loss": 0.116,
+      "step": 1755
+    },
+    {
+      "epoch": 3.9198218262806237,
+      "grad_norm": 1.0203088521957397,
+      "learning_rate": 3.684836192040402e-06,
+      "loss": 0.1051,
+      "step": 1760
+    },
+    {
+      "epoch": 3.930957683741648,
+      "grad_norm": 0.936297595500946,
+      "learning_rate": 3.612592723774942e-06,
+      "loss": 0.1026,
+      "step": 1765
+    },
+    {
+      "epoch": 3.9420935412026727,
+      "grad_norm": 1.0695292949676514,
+      "learning_rate": 3.540967397494521e-06,
+      "loss": 0.103,
+      "step": 1770
+    },
+    {
+      "epoch": 3.953229398663697,
+      "grad_norm": 0.7689676880836487,
+      "learning_rate": 3.4699641012326723e-06,
+      "loss": 0.0903,
+      "step": 1775
+    },
+    {
+      "epoch": 3.9643652561247213,
+      "grad_norm": 0.7756615877151489,
+      "learning_rate": 3.399586689257305e-06,
+      "loss": 0.1067,
+      "step": 1780
+    },
+    {
+      "epoch": 3.975501113585746,
+      "grad_norm": 0.8232401013374329,
+      "learning_rate": 3.329838981861504e-06,
+      "loss": 0.0931,
+      "step": 1785
+    },
+    {
+      "epoch": 3.986636971046771,
+      "grad_norm": 1.06974458694458,
+      "learning_rate": 3.2607247651561094e-06,
+      "loss": 0.1117,
+      "step": 1790
+    },
+    {
+      "epoch": 3.997772828507795,
+      "grad_norm": 0.8997353315353394,
+      "learning_rate": 3.192247790864249e-06,
+      "loss": 0.1004,
+      "step": 1795
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2245,
+  "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": 2.485555820259967e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

38_128_e5_3e-5/checkpoint-1796/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28546b29d51193caf8c0ea9b7bbe68af481f6dc06b3ac3e27fa68da70bce3d29
+size 7736

38_128_e5_3e-5/checkpoint-1796/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)

38_128_e5_3e-5/checkpoint-2245/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

38_128_e5_3e-5/checkpoint-2245/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": [
+    "q_proj",
+    "v_proj",
+    "gate_proj",
+    "o_proj",
+    "down_proj",
+    "k_proj",
+    "up_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

38_128_e5_3e-5/checkpoint-2245/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8cb177dd934fb324095fc80d4b8259cfe6d7b80213536d7018c8ae5b43816769
+size 791751704

38_128_e5_3e-5/checkpoint-2245/latest

@@ -0,0 +1 @@
+global_step2245

38_128_e5_3e-5/checkpoint-2245/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df44e7d2b5f4009b65ef298ea1250a420a4695c737dbb7a43d3547b54fbf4c0a
+size 15920

38_128_e5_3e-5/checkpoint-2245/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2ab756b3992edb8b98888cea6457dc405d4e22c58f8222d16e594f90d56fafe5
+size 15920

38_128_e5_3e-5/checkpoint-2245/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4562ad81a3dda5f1cd840fef592a0b57b39b1a6833a50f1efff8cf31e8b5b011
+size 15920