Upload folder using huggingface_hub

25_128_e5_3e-5/checkpoint-1293/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

25_128_e5_3e-5/checkpoint-1293/adapter_config.json

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

25_128_e5_3e-5/checkpoint-1293/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b2e2947529e321d74526e36c78f9ab0998fcd416905cf127bc10b8587f90481
+size 791751704

25_128_e5_3e-5/checkpoint-1293/latest

@@ -0,0 +1 @@
+global_step1293

25_128_e5_3e-5/checkpoint-1293/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bee9d9f246c7f9bcc03732992f735f48a479ee662b3a9b6b747b35f086d0be6b
+size 15920

25_128_e5_3e-5/checkpoint-1293/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:841078892087194dd30f792e033ef277c1de77493c3f13eb38ad79bcab80adf8
+size 15920

25_128_e5_3e-5/checkpoint-1293/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dfa4b9526c78ff72c85b7578e10ae0ad94d7f532a6786b9bfadca6b522bcd61d
+size 15920

25_128_e5_3e-5/checkpoint-1293/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:355db91fc189b8d300af72f162aaffc8ceba72f22c1178400311b17af13bd680
+size 15920

25_128_e5_3e-5/checkpoint-1293/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80cc4975169668deffcb78de75b78709c2dee1a8a4609bc36cf2c4c557828099
+size 15920

25_128_e5_3e-5/checkpoint-1293/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65a59976a5a2718e3550d4013a7ebbebf5a99382ef5ad5588d2d18555292ec0e
+size 15920

25_128_e5_3e-5/checkpoint-1293/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10edf0729f5372d462ee74d399ec8a6951d49eb79b757540d5748b7c29889395
+size 15920

25_128_e5_3e-5/checkpoint-1293/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fab5483a10de71d1081cfb856329619130c0b0a31be892fb67a2489bff23105
+size 15920

25_128_e5_3e-5/checkpoint-1293/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:38fbdb165482ae85cb14120f7387ff84296e2c5a8919fe110c4fef6bd8d6ea11
+size 1064

25_128_e5_3e-5/checkpoint-1293/special_tokens_map.json

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

25_128_e5_3e-5/checkpoint-1293/tokenizer_config.json

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

25_128_e5_3e-5/checkpoint-1293/trainer_state.json

@@ -0,0 +1,1840 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1293,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011614401858304297,
+      "grad_norm": 1.2486553192138672,
+      "learning_rate": 1.111111111111111e-06,
+      "loss": 1.3051,
+      "step": 5
+    },
+    {
+      "epoch": 0.023228803716608595,
+      "grad_norm": 1.02664053440094,
+      "learning_rate": 2.4999999999999998e-06,
+      "loss": 1.3323,
+      "step": 10
+    },
+    {
+      "epoch": 0.03484320557491289,
+      "grad_norm": 0.8225151896476746,
+      "learning_rate": 3.888888888888889e-06,
+      "loss": 1.3047,
+      "step": 15
+    },
+    {
+      "epoch": 0.04645760743321719,
+      "grad_norm": 0.7819671630859375,
+      "learning_rate": 5.277777777777778e-06,
+      "loss": 1.2887,
+      "step": 20
+    },
+    {
+      "epoch": 0.05807200929152149,
+      "grad_norm": 0.6035093069076538,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 1.3248,
+      "step": 25
+    },
+    {
+      "epoch": 0.06968641114982578,
+      "grad_norm": 0.5274394750595093,
+      "learning_rate": 8.055555555555557e-06,
+      "loss": 1.2981,
+      "step": 30
+    },
+    {
+      "epoch": 0.08130081300813008,
+      "grad_norm": 0.4473659098148346,
+      "learning_rate": 9.444444444444445e-06,
+      "loss": 1.2253,
+      "step": 35
+    },
+    {
+      "epoch": 0.09291521486643438,
+      "grad_norm": 0.5344942808151245,
+      "learning_rate": 1.0833333333333334e-05,
+      "loss": 1.2101,
+      "step": 40
+    },
+    {
+      "epoch": 0.10452961672473868,
+      "grad_norm": 0.42955347895622253,
+      "learning_rate": 1.2222222222222222e-05,
+      "loss": 1.2136,
+      "step": 45
+    },
+    {
+      "epoch": 0.11614401858304298,
+      "grad_norm": 0.4606517553329468,
+      "learning_rate": 1.3611111111111111e-05,
+      "loss": 1.2501,
+      "step": 50
+    },
+    {
+      "epoch": 0.12775842044134728,
+      "grad_norm": 0.43243467807769775,
+      "learning_rate": 1.5e-05,
+      "loss": 1.2287,
+      "step": 55
+    },
+    {
+      "epoch": 0.13937282229965156,
+      "grad_norm": 0.45852982997894287,
+      "learning_rate": 1.638888888888889e-05,
+      "loss": 1.2272,
+      "step": 60
+    },
+    {
+      "epoch": 0.15098722415795587,
+      "grad_norm": 0.422735333442688,
+      "learning_rate": 1.7777777777777777e-05,
+      "loss": 1.1591,
+      "step": 65
+    },
+    {
+      "epoch": 0.16260162601626016,
+      "grad_norm": 0.48696285486221313,
+      "learning_rate": 1.9166666666666667e-05,
+      "loss": 1.2238,
+      "step": 70
+    },
+    {
+      "epoch": 0.17421602787456447,
+      "grad_norm": 0.6145470142364502,
+      "learning_rate": 2.0555555555555558e-05,
+      "loss": 1.2375,
+      "step": 75
+    },
+    {
+      "epoch": 0.18583042973286876,
+      "grad_norm": 0.5468384623527527,
+      "learning_rate": 2.1944444444444445e-05,
+      "loss": 1.1484,
+      "step": 80
+    },
+    {
+      "epoch": 0.19744483159117304,
+      "grad_norm": 0.4397794008255005,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 1.1567,
+      "step": 85
+    },
+    {
+      "epoch": 0.20905923344947736,
+      "grad_norm": 0.4396965801715851,
+      "learning_rate": 2.4722222222222223e-05,
+      "loss": 1.1665,
+      "step": 90
+    },
+    {
+      "epoch": 0.22067363530778164,
+      "grad_norm": 0.5894995927810669,
+      "learning_rate": 2.611111111111111e-05,
+      "loss": 1.1519,
+      "step": 95
+    },
+    {
+      "epoch": 0.23228803716608595,
+      "grad_norm": 0.4796619117259979,
+      "learning_rate": 2.75e-05,
+      "loss": 1.1242,
+      "step": 100
+    },
+    {
+      "epoch": 0.24390243902439024,
+      "grad_norm": 0.7274127006530762,
+      "learning_rate": 2.8888888888888888e-05,
+      "loss": 1.1197,
+      "step": 105
+    },
+    {
+      "epoch": 0.25551684088269455,
+      "grad_norm": 0.5538251996040344,
+      "learning_rate": 2.999998233452831e-05,
+      "loss": 1.09,
+      "step": 110
+    },
+    {
+      "epoch": 0.26713124274099886,
+      "grad_norm": 0.4963814616203308,
+      "learning_rate": 2.999936404738799e-05,
+      "loss": 1.14,
+      "step": 115
+    },
+    {
+      "epoch": 0.2787456445993031,
+      "grad_norm": 0.5045656561851501,
+      "learning_rate": 2.9997862528271754e-05,
+      "loss": 1.0846,
+      "step": 120
+    },
+    {
+      "epoch": 0.29036004645760743,
+      "grad_norm": 0.5694212913513184,
+      "learning_rate": 2.999547786559598e-05,
+      "loss": 1.0566,
+      "step": 125
+    },
+    {
+      "epoch": 0.30197444831591175,
+      "grad_norm": 0.5698084831237793,
+      "learning_rate": 2.9992210199780657e-05,
+      "loss": 1.0812,
+      "step": 130
+    },
+    {
+      "epoch": 0.313588850174216,
+      "grad_norm": 0.5170321464538574,
+      "learning_rate": 2.9988059723241064e-05,
+      "loss": 1.0414,
+      "step": 135
+    },
+    {
+      "epoch": 0.3252032520325203,
+      "grad_norm": 0.5222340226173401,
+      "learning_rate": 2.9983026680376472e-05,
+      "loss": 1.0631,
+      "step": 140
+    },
+    {
+      "epoch": 0.33681765389082463,
+      "grad_norm": 0.5456163883209229,
+      "learning_rate": 2.997711136755574e-05,
+      "loss": 1.0866,
+      "step": 145
+    },
+    {
+      "epoch": 0.34843205574912894,
+      "grad_norm": 0.5397904515266418,
+      "learning_rate": 2.9970314133099855e-05,
+      "loss": 1.1579,
+      "step": 150
+    },
+    {
+      "epoch": 0.3600464576074332,
+      "grad_norm": 0.6353335380554199,
+      "learning_rate": 2.9962635377261457e-05,
+      "loss": 0.9954,
+      "step": 155
+    },
+    {
+      "epoch": 0.3716608594657375,
+      "grad_norm": 0.6079632639884949,
+      "learning_rate": 2.9954075552201222e-05,
+      "loss": 1.0647,
+      "step": 160
+    },
+    {
+      "epoch": 0.3832752613240418,
+      "grad_norm": 0.6199746131896973,
+      "learning_rate": 2.994463516196126e-05,
+      "loss": 1.0637,
+      "step": 165
+    },
+    {
+      "epoch": 0.3948896631823461,
+      "grad_norm": 0.6312094926834106,
+      "learning_rate": 2.9934314762435444e-05,
+      "loss": 1.0242,
+      "step": 170
+    },
+    {
+      "epoch": 0.4065040650406504,
+      "grad_norm": 0.677851140499115,
+      "learning_rate": 2.9923114961336672e-05,
+      "loss": 0.9563,
+      "step": 175
+    },
+    {
+      "epoch": 0.4181184668989547,
+      "grad_norm": 0.6415532827377319,
+      "learning_rate": 2.9911036418161058e-05,
+      "loss": 0.9728,
+      "step": 180
+    },
+    {
+      "epoch": 0.429732868757259,
+      "grad_norm": 0.7620236873626709,
+      "learning_rate": 2.9898079844149132e-05,
+      "loss": 1.0066,
+      "step": 185
+    },
+    {
+      "epoch": 0.4413472706155633,
+      "grad_norm": 0.6720292568206787,
+      "learning_rate": 2.9884246002243936e-05,
+      "loss": 1.0059,
+      "step": 190
+    },
+    {
+      "epoch": 0.4529616724738676,
+      "grad_norm": 0.7387037873268127,
+      "learning_rate": 2.9869535707046104e-05,
+      "loss": 0.9314,
+      "step": 195
+    },
+    {
+      "epoch": 0.4645760743321719,
+      "grad_norm": 0.6071408987045288,
+      "learning_rate": 2.985394982476591e-05,
+      "loss": 0.9805,
+      "step": 200
+    },
+    {
+      "epoch": 0.47619047619047616,
+      "grad_norm": 0.7038552165031433,
+      "learning_rate": 2.9837489273172232e-05,
+      "loss": 0.9766,
+      "step": 205
+    },
+    {
+      "epoch": 0.4878048780487805,
+      "grad_norm": 0.7892246842384338,
+      "learning_rate": 2.9820155021538533e-05,
+      "loss": 0.9558,
+      "step": 210
+    },
+    {
+      "epoch": 0.4994192799070848,
+      "grad_norm": 0.7563413381576538,
+      "learning_rate": 2.980194809058577e-05,
+      "loss": 0.9455,
+      "step": 215
+    },
+    {
+      "epoch": 0.5110336817653891,
+      "grad_norm": 0.7255906462669373,
+      "learning_rate": 2.9782869552422316e-05,
+      "loss": 0.9183,
+      "step": 220
+    },
+    {
+      "epoch": 0.5226480836236934,
+      "grad_norm": 0.7455199956893921,
+      "learning_rate": 2.9762920530480788e-05,
+      "loss": 0.9363,
+      "step": 225
+    },
+    {
+      "epoch": 0.5342624854819977,
+      "grad_norm": 0.7163615822792053,
+      "learning_rate": 2.974210219945193e-05,
+      "loss": 0.9436,
+      "step": 230
+    },
+    {
+      "epoch": 0.5458768873403019,
+      "grad_norm": 0.6869027018547058,
+      "learning_rate": 2.9720415785215428e-05,
+      "loss": 0.8932,
+      "step": 235
+    },
+    {
+      "epoch": 0.5574912891986062,
+      "grad_norm": 0.776175320148468,
+      "learning_rate": 2.969786256476772e-05,
+      "loss": 0.9234,
+      "step": 240
+    },
+    {
+      "epoch": 0.5691056910569106,
+      "grad_norm": 0.687179446220398,
+      "learning_rate": 2.9674443866146807e-05,
+      "loss": 0.9165,
+      "step": 245
+    },
+    {
+      "epoch": 0.5807200929152149,
+      "grad_norm": 0.8298389911651611,
+      "learning_rate": 2.9650161068354054e-05,
+      "loss": 0.8863,
+      "step": 250
+    },
+    {
+      "epoch": 0.5923344947735192,
+      "grad_norm": 0.7793598175048828,
+      "learning_rate": 2.9625015601272974e-05,
+      "loss": 0.8693,
+      "step": 255
+    },
+    {
+      "epoch": 0.6039488966318235,
+      "grad_norm": 0.9237378835678101,
+      "learning_rate": 2.9599008945585066e-05,
+      "loss": 0.8435,
+      "step": 260
+    },
+    {
+      "epoch": 0.6155632984901278,
+      "grad_norm": 0.8277608752250671,
+      "learning_rate": 2.9572142632682562e-05,
+      "loss": 0.8683,
+      "step": 265
+    },
+    {
+      "epoch": 0.627177700348432,
+      "grad_norm": 0.7906745076179504,
+      "learning_rate": 2.954441824457832e-05,
+      "loss": 0.9144,
+      "step": 270
+    },
+    {
+      "epoch": 0.6387921022067363,
+      "grad_norm": 0.8107167482376099,
+      "learning_rate": 2.951583741381263e-05,
+      "loss": 0.8696,
+      "step": 275
+    },
+    {
+      "epoch": 0.6504065040650406,
+      "grad_norm": 0.734391987323761,
+      "learning_rate": 2.948640182335708e-05,
+      "loss": 0.871,
+      "step": 280
+    },
+    {
+      "epoch": 0.662020905923345,
+      "grad_norm": 0.8837676048278809,
+      "learning_rate": 2.9456113206515475e-05,
+      "loss": 0.8883,
+      "step": 285
+    },
+    {
+      "epoch": 0.6736353077816493,
+      "grad_norm": 0.8748791217803955,
+      "learning_rate": 2.942497334682176e-05,
+      "loss": 0.7801,
+      "step": 290
+    },
+    {
+      "epoch": 0.6852497096399536,
+      "grad_norm": 0.8546818494796753,
+      "learning_rate": 2.9392984077934987e-05,
+      "loss": 0.8724,
+      "step": 295
+    },
+    {
+      "epoch": 0.6968641114982579,
+      "grad_norm": 0.7864014506340027,
+      "learning_rate": 2.9360147283531373e-05,
+      "loss": 0.8214,
+      "step": 300
+    },
+    {
+      "epoch": 0.7084785133565621,
+      "grad_norm": 0.7984604835510254,
+      "learning_rate": 2.9326464897193343e-05,
+      "loss": 0.836,
+      "step": 305
+    },
+    {
+      "epoch": 0.7200929152148664,
+      "grad_norm": 0.9750844836235046,
+      "learning_rate": 2.9291938902295695e-05,
+      "loss": 0.7873,
+      "step": 310
+    },
+    {
+      "epoch": 0.7317073170731707,
+      "grad_norm": 0.8499948382377625,
+      "learning_rate": 2.925657133188881e-05,
+      "loss": 0.8412,
+      "step": 315
+    },
+    {
+      "epoch": 0.743321718931475,
+      "grad_norm": 0.86067134141922,
+      "learning_rate": 2.9220364268578922e-05,
+      "loss": 0.8256,
+      "step": 320
+    },
+    {
+      "epoch": 0.7549361207897793,
+      "grad_norm": 0.8807356357574463,
+      "learning_rate": 2.918331984440549e-05,
+      "loss": 0.8364,
+      "step": 325
+    },
+    {
+      "epoch": 0.7665505226480837,
+      "grad_norm": 0.8951946496963501,
+      "learning_rate": 2.9145440240715657e-05,
+      "loss": 0.8415,
+      "step": 330
+    },
+    {
+      "epoch": 0.778164924506388,
+      "grad_norm": 0.8705118894577026,
+      "learning_rate": 2.9106727688035814e-05,
+      "loss": 0.7762,
+      "step": 335
+    },
+    {
+      "epoch": 0.7897793263646922,
+      "grad_norm": 0.9147759675979614,
+      "learning_rate": 2.9067184465940225e-05,
+      "loss": 0.7983,
+      "step": 340
+    },
+    {
+      "epoch": 0.8013937282229965,
+      "grad_norm": 0.8624527454376221,
+      "learning_rate": 2.9026812902916834e-05,
+      "loss": 0.7434,
+      "step": 345
+    },
+    {
+      "epoch": 0.8130081300813008,
+      "grad_norm": 0.9535494446754456,
+      "learning_rate": 2.898561537623011e-05,
+      "loss": 0.8049,
+      "step": 350
+    },
+    {
+      "epoch": 0.8246225319396051,
+      "grad_norm": 0.836469292640686,
+      "learning_rate": 2.8943594311781104e-05,
+      "loss": 0.777,
+      "step": 355
+    },
+    {
+      "epoch": 0.8362369337979094,
+      "grad_norm": 0.9156211018562317,
+      "learning_rate": 2.8900752183964573e-05,
+      "loss": 0.7746,
+      "step": 360
+    },
+    {
+      "epoch": 0.8478513356562137,
+      "grad_norm": 0.9176764488220215,
+      "learning_rate": 2.8857091515523287e-05,
+      "loss": 0.7827,
+      "step": 365
+    },
+    {
+      "epoch": 0.859465737514518,
+      "grad_norm": 0.8853964805603027,
+      "learning_rate": 2.8812614877399476e-05,
+      "loss": 0.754,
+      "step": 370
+    },
+    {
+      "epoch": 0.8710801393728222,
+      "grad_norm": 0.8558771014213562,
+      "learning_rate": 2.876732488858344e-05,
+      "loss": 0.7039,
+      "step": 375
+    },
+    {
+      "epoch": 0.8826945412311266,
+      "grad_norm": 0.8176305294036865,
+      "learning_rate": 2.8721224215959335e-05,
+      "loss": 0.7892,
+      "step": 380
+    },
+    {
+      "epoch": 0.8943089430894309,
+      "grad_norm": 0.9530948996543884,
+      "learning_rate": 2.8674315574148126e-05,
+      "loss": 0.7049,
+      "step": 385
+    },
+    {
+      "epoch": 0.9059233449477352,
+      "grad_norm": 0.9136149287223816,
+      "learning_rate": 2.862660172534776e-05,
+      "loss": 0.7614,
+      "step": 390
+    },
+    {
+      "epoch": 0.9175377468060395,
+      "grad_norm": 0.9666270017623901,
+      "learning_rate": 2.8578085479170478e-05,
+      "loss": 0.7265,
+      "step": 395
+    },
+    {
+      "epoch": 0.9291521486643438,
+      "grad_norm": 0.9841073751449585,
+      "learning_rate": 2.85287696924774e-05,
+      "loss": 0.6903,
+      "step": 400
+    },
+    {
+      "epoch": 0.9407665505226481,
+      "grad_norm": 1.1007276773452759,
+      "learning_rate": 2.8478657269210294e-05,
+      "loss": 0.7114,
+      "step": 405
+    },
+    {
+      "epoch": 0.9523809523809523,
+      "grad_norm": 0.9525352120399475,
+      "learning_rate": 2.8427751160220573e-05,
+      "loss": 0.7115,
+      "step": 410
+    },
+    {
+      "epoch": 0.9639953542392566,
+      "grad_norm": 0.8143008351325989,
+      "learning_rate": 2.8376054363095545e-05,
+      "loss": 0.6784,
+      "step": 415
+    },
+    {
+      "epoch": 0.975609756097561,
+      "grad_norm": 0.9916719198226929,
+      "learning_rate": 2.8323569921981885e-05,
+      "loss": 0.7366,
+      "step": 420
+    },
+    {
+      "epoch": 0.9872241579558653,
+      "grad_norm": 0.9247941374778748,
+      "learning_rate": 2.82703009274064e-05,
+      "loss": 0.6974,
+      "step": 425
+    },
+    {
+      "epoch": 0.9988385598141696,
+      "grad_norm": 0.9906225800514221,
+      "learning_rate": 2.8216250516094027e-05,
+      "loss": 0.6673,
+      "step": 430
+    },
+    {
+      "epoch": 1.0092915214866434,
+      "grad_norm": 0.9208662509918213,
+      "learning_rate": 2.816142187078315e-05,
+      "loss": 0.603,
+      "step": 435
+    },
+    {
+      "epoch": 1.0209059233449478,
+      "grad_norm": 1.0540852546691895,
+      "learning_rate": 2.8105818220038167e-05,
+      "loss": 0.6062,
+      "step": 440
+    },
+    {
+      "epoch": 1.032520325203252,
+      "grad_norm": 0.9971239566802979,
+      "learning_rate": 2.804944283805938e-05,
+      "loss": 0.5986,
+      "step": 445
+    },
+    {
+      "epoch": 1.0441347270615564,
+      "grad_norm": 1.0977699756622314,
+      "learning_rate": 2.7992299044490192e-05,
+      "loss": 0.5556,
+      "step": 450
+    },
+    {
+      "epoch": 1.0557491289198606,
+      "grad_norm": 1.0428180694580078,
+      "learning_rate": 2.793439020422165e-05,
+      "loss": 0.599,
+      "step": 455
+    },
+    {
+      "epoch": 1.0673635307781648,
+      "grad_norm": 1.2783362865447998,
+      "learning_rate": 2.787571972719429e-05,
+      "loss": 0.5589,
+      "step": 460
+    },
+    {
+      "epoch": 1.0789779326364692,
+      "grad_norm": 1.0532207489013672,
+      "learning_rate": 2.781629106819733e-05,
+      "loss": 0.5534,
+      "step": 465
+    },
+    {
+      "epoch": 1.0905923344947734,
+      "grad_norm": 1.0525319576263428,
+      "learning_rate": 2.775610772666527e-05,
+      "loss": 0.577,
+      "step": 470
+    },
+    {
+      "epoch": 1.1022067363530779,
+      "grad_norm": 0.9602555632591248,
+      "learning_rate": 2.7695173246471803e-05,
+      "loss": 0.5625,
+      "step": 475
+    },
+    {
+      "epoch": 1.113821138211382,
+      "grad_norm": 1.0547131299972534,
+      "learning_rate": 2.763349121572114e-05,
+      "loss": 0.6173,
+      "step": 480
+    },
+    {
+      "epoch": 1.1254355400696865,
+      "grad_norm": 0.963524341583252,
+      "learning_rate": 2.7571065266536737e-05,
+      "loss": 0.62,
+      "step": 485
+    },
+    {
+      "epoch": 1.1370499419279907,
+      "grad_norm": 1.066751480102539,
+      "learning_rate": 2.7507899074847394e-05,
+      "loss": 0.5829,
+      "step": 490
+    },
+    {
+      "epoch": 1.1486643437862951,
+      "grad_norm": 0.9866227507591248,
+      "learning_rate": 2.7443996360170836e-05,
+      "loss": 0.5729,
+      "step": 495
+    },
+    {
+      "epoch": 1.1602787456445993,
+      "grad_norm": 1.037983775138855,
+      "learning_rate": 2.7379360885394664e-05,
+      "loss": 0.5955,
+      "step": 500
+    },
+    {
+      "epoch": 1.1718931475029035,
+      "grad_norm": 1.2030766010284424,
+      "learning_rate": 2.731399645655477e-05,
+      "loss": 0.6062,
+      "step": 505
+    },
+    {
+      "epoch": 1.183507549361208,
+      "grad_norm": 1.0329967737197876,
+      "learning_rate": 2.7247906922611254e-05,
+      "loss": 0.5871,
+      "step": 510
+    },
+    {
+      "epoch": 1.1951219512195121,
+      "grad_norm": 1.1776123046875,
+      "learning_rate": 2.7181096175221757e-05,
+      "loss": 0.5755,
+      "step": 515
+    },
+    {
+      "epoch": 1.2067363530778166,
+      "grad_norm": 1.0537413358688354,
+      "learning_rate": 2.7113568148512296e-05,
+      "loss": 0.545,
+      "step": 520
+    },
+    {
+      "epoch": 1.2183507549361208,
+      "grad_norm": 0.9902351498603821,
+      "learning_rate": 2.704532681884562e-05,
+      "loss": 0.5564,
+      "step": 525
+    },
+    {
+      "epoch": 1.229965156794425,
+      "grad_norm": 1.0097850561141968,
+      "learning_rate": 2.697637620458706e-05,
+      "loss": 0.5056,
+      "step": 530
+    },
+    {
+      "epoch": 1.2415795586527294,
+      "grad_norm": 0.9843356013298035,
+      "learning_rate": 2.690672036586791e-05,
+      "loss": 0.5348,
+      "step": 535
+    },
+    {
+      "epoch": 1.2531939605110336,
+      "grad_norm": 0.9540740251541138,
+      "learning_rate": 2.6836363404346324e-05,
+      "loss": 0.5411,
+      "step": 540
+    },
+    {
+      "epoch": 1.264808362369338,
+      "grad_norm": 1.0271143913269043,
+      "learning_rate": 2.6765309462965845e-05,
+      "loss": 0.5065,
+      "step": 545
+    },
+    {
+      "epoch": 1.2764227642276422,
+      "grad_norm": 1.077792763710022,
+      "learning_rate": 2.669356272571138e-05,
+      "loss": 0.5307,
+      "step": 550
+    },
+    {
+      "epoch": 1.2880371660859466,
+      "grad_norm": 1.0701345205307007,
+      "learning_rate": 2.6621127417362886e-05,
+      "loss": 0.5587,
+      "step": 555
+    },
+    {
+      "epoch": 1.2996515679442509,
+      "grad_norm": 1.0792369842529297,
+      "learning_rate": 2.6548007803246575e-05,
+      "loss": 0.5264,
+      "step": 560
+    },
+    {
+      "epoch": 1.3112659698025553,
+      "grad_norm": 0.8799574375152588,
+      "learning_rate": 2.647420818898373e-05,
+      "loss": 0.4964,
+      "step": 565
+    },
+    {
+      "epoch": 1.3228803716608595,
+      "grad_norm": 1.1636409759521484,
+      "learning_rate": 2.6399732920237212e-05,
+      "loss": 0.5576,
+      "step": 570
+    },
+    {
+      "epoch": 1.3344947735191637,
+      "grad_norm": 1.055418610572815,
+      "learning_rate": 2.6324586382455525e-05,
+      "loss": 0.5149,
+      "step": 575
+    },
+    {
+      "epoch": 1.346109175377468,
+      "grad_norm": 1.0145419836044312,
+      "learning_rate": 2.624877300061462e-05,
+      "loss": 0.4917,
+      "step": 580
+    },
+    {
+      "epoch": 1.3577235772357723,
+      "grad_norm": 0.9634491801261902,
+      "learning_rate": 2.6172297238957297e-05,
+      "loss": 0.515,
+      "step": 585
+    },
+    {
+      "epoch": 1.3693379790940767,
+      "grad_norm": 0.9772712588310242,
+      "learning_rate": 2.6095163600730355e-05,
+      "loss": 0.5012,
+      "step": 590
+    },
+    {
+      "epoch": 1.380952380952381,
+      "grad_norm": 0.9910968542098999,
+      "learning_rate": 2.6017376627919405e-05,
+      "loss": 0.4839,
+      "step": 595
+    },
+    {
+      "epoch": 1.3925667828106851,
+      "grad_norm": 1.103830099105835,
+      "learning_rate": 2.5938940900981424e-05,
+      "loss": 0.562,
+      "step": 600
+    },
+    {
+      "epoch": 1.4041811846689896,
+      "grad_norm": 1.1236923933029175,
+      "learning_rate": 2.5859861038575035e-05,
+      "loss": 0.493,
+      "step": 605
+    },
+    {
+      "epoch": 1.415795586527294,
+      "grad_norm": 0.9853049516677856,
+      "learning_rate": 2.5780141697288537e-05,
+      "loss": 0.4578,
+      "step": 610
+    },
+    {
+      "epoch": 1.4274099883855982,
+      "grad_norm": 1.02285635471344,
+      "learning_rate": 2.5699787571365704e-05,
+      "loss": 0.5141,
+      "step": 615
+    },
+    {
+      "epoch": 1.4390243902439024,
+      "grad_norm": 1.1389844417572021,
+      "learning_rate": 2.5618803392429373e-05,
+      "loss": 0.4684,
+      "step": 620
+    },
+    {
+      "epoch": 1.4506387921022068,
+      "grad_norm": 1.1375923156738281,
+      "learning_rate": 2.5537193929202815e-05,
+      "loss": 0.4761,
+      "step": 625
+    },
+    {
+      "epoch": 1.462253193960511,
+      "grad_norm": 1.046151876449585,
+      "learning_rate": 2.5454963987228926e-05,
+      "loss": 0.5179,
+      "step": 630
+    },
+    {
+      "epoch": 1.4738675958188154,
+      "grad_norm": 1.1057745218276978,
+      "learning_rate": 2.5372118408587284e-05,
+      "loss": 0.489,
+      "step": 635
+    },
+    {
+      "epoch": 1.4854819976771196,
+      "grad_norm": 1.032041311264038,
+      "learning_rate": 2.5288662071608975e-05,
+      "loss": 0.4871,
+      "step": 640
+    },
+    {
+      "epoch": 1.4970963995354238,
+      "grad_norm": 1.0814687013626099,
+      "learning_rate": 2.520459989058939e-05,
+      "loss": 0.5172,
+      "step": 645
+    },
+    {
+      "epoch": 1.5087108013937283,
+      "grad_norm": 1.0195306539535522,
+      "learning_rate": 2.5119936815498797e-05,
+      "loss": 0.5218,
+      "step": 650
+    },
+    {
+      "epoch": 1.5203252032520327,
+      "grad_norm": 0.9469033479690552,
+      "learning_rate": 2.503467783169091e-05,
+      "loss": 0.4883,
+      "step": 655
+    },
+    {
+      "epoch": 1.5319396051103369,
+      "grad_norm": 1.1030808687210083,
+      "learning_rate": 2.4948827959609285e-05,
+      "loss": 0.501,
+      "step": 660
+    },
+    {
+      "epoch": 1.543554006968641,
+      "grad_norm": 1.0832819938659668,
+      "learning_rate": 2.4862392254491736e-05,
+      "loss": 0.4813,
+      "step": 665
+    },
+    {
+      "epoch": 1.5551684088269453,
+      "grad_norm": 1.1702861785888672,
+      "learning_rate": 2.477537580607261e-05,
+      "loss": 0.4737,
+      "step": 670
+    },
+    {
+      "epoch": 1.5667828106852497,
+      "grad_norm": 1.0441266298294067,
+      "learning_rate": 2.4687783738283144e-05,
+      "loss": 0.454,
+      "step": 675
+    },
+    {
+      "epoch": 1.5783972125435541,
+      "grad_norm": 1.1927803754806519,
+      "learning_rate": 2.4599621208949674e-05,
+      "loss": 0.4328,
+      "step": 680
+    },
+    {
+      "epoch": 1.5900116144018583,
+      "grad_norm": 1.1129752397537231,
+      "learning_rate": 2.4510893409489967e-05,
+      "loss": 0.4024,
+      "step": 685
+    },
+    {
+      "epoch": 1.6016260162601625,
+      "grad_norm": 1.0736514329910278,
+      "learning_rate": 2.4421605564607514e-05,
+      "loss": 0.4499,
+      "step": 690
+    },
+    {
+      "epoch": 1.6132404181184667,
+      "grad_norm": 1.0657414197921753,
+      "learning_rate": 2.4331762931983866e-05,
+      "loss": 0.4649,
+      "step": 695
+    },
+    {
+      "epoch": 1.6248548199767712,
+      "grad_norm": 1.0618281364440918,
+      "learning_rate": 2.4241370801969045e-05,
+      "loss": 0.4337,
+      "step": 700
+    },
+    {
+      "epoch": 1.6364692218350756,
+      "grad_norm": 1.1562613248825073,
+      "learning_rate": 2.415043449727003e-05,
+      "loss": 0.4429,
+      "step": 705
+    },
+    {
+      "epoch": 1.6480836236933798,
+      "grad_norm": 1.1286054849624634,
+      "learning_rate": 2.4058959372637304e-05,
+      "loss": 0.4714,
+      "step": 710
+    },
+    {
+      "epoch": 1.659698025551684,
+      "grad_norm": 1.043707251548767,
+      "learning_rate": 2.396695081454959e-05,
+      "loss": 0.4549,
+      "step": 715
+    },
+    {
+      "epoch": 1.6713124274099884,
+      "grad_norm": 1.001025676727295,
+      "learning_rate": 2.387441424089662e-05,
+      "loss": 0.4843,
+      "step": 720
+    },
+    {
+      "epoch": 1.6829268292682928,
+      "grad_norm": 1.0033694505691528,
+      "learning_rate": 2.378135510066013e-05,
+      "loss": 0.4203,
+      "step": 725
+    },
+    {
+      "epoch": 1.694541231126597,
+      "grad_norm": 1.041979432106018,
+      "learning_rate": 2.3687778873593e-05,
+      "loss": 0.4609,
+      "step": 730
+    },
+    {
+      "epoch": 1.7061556329849012,
+      "grad_norm": 1.0614110231399536,
+      "learning_rate": 2.3593691069896582e-05,
+      "loss": 0.4328,
+      "step": 735
+    },
+    {
+      "epoch": 1.7177700348432055,
+      "grad_norm": 1.1184892654418945,
+      "learning_rate": 2.3499097229896213e-05,
+      "loss": 0.4323,
+      "step": 740
+    },
+    {
+      "epoch": 1.7293844367015099,
+      "grad_norm": 1.0571225881576538,
+      "learning_rate": 2.340400292371499e-05,
+      "loss": 0.3688,
+      "step": 745
+    },
+    {
+      "epoch": 1.7409988385598143,
+      "grad_norm": 0.9831321239471436,
+      "learning_rate": 2.3308413750945788e-05,
+      "loss": 0.4091,
+      "step": 750
+    },
+    {
+      "epoch": 1.7526132404181185,
+      "grad_norm": 1.334681749343872,
+      "learning_rate": 2.3212335340321518e-05,
+      "loss": 0.4363,
+      "step": 755
+    },
+    {
+      "epoch": 1.7642276422764227,
+      "grad_norm": 1.0677987337112427,
+      "learning_rate": 2.3115773349383658e-05,
+      "loss": 0.4202,
+      "step": 760
+    },
+    {
+      "epoch": 1.775842044134727,
+      "grad_norm": 1.1645525693893433,
+      "learning_rate": 2.3018733464149156e-05,
+      "loss": 0.4515,
+      "step": 765
+    },
+    {
+      "epoch": 1.7874564459930313,
+      "grad_norm": 1.0856503248214722,
+      "learning_rate": 2.292122139877558e-05,
+      "loss": 0.401,
+      "step": 770
+    },
+    {
+      "epoch": 1.7990708478513358,
+      "grad_norm": 1.0238492488861084,
+      "learning_rate": 2.2823242895224643e-05,
+      "loss": 0.3955,
+      "step": 775
+    },
+    {
+      "epoch": 1.81068524970964,
+      "grad_norm": 1.1932063102722168,
+      "learning_rate": 2.2724803722924106e-05,
+      "loss": 0.4154,
+      "step": 780
+    },
+    {
+      "epoch": 1.8222996515679442,
+      "grad_norm": 1.0874030590057373,
+      "learning_rate": 2.2625909678428038e-05,
+      "loss": 0.4082,
+      "step": 785
+    },
+    {
+      "epoch": 1.8339140534262486,
+      "grad_norm": 1.109732747077942,
+      "learning_rate": 2.2526566585075485e-05,
+      "loss": 0.3826,
+      "step": 790
+    },
+    {
+      "epoch": 1.845528455284553,
+      "grad_norm": 1.0591766834259033,
+      "learning_rate": 2.2426780292647568e-05,
+      "loss": 0.388,
+      "step": 795
+    },
+    {
+      "epoch": 1.8571428571428572,
+      "grad_norm": 1.1858235597610474,
+      "learning_rate": 2.2326556677023017e-05,
+      "loss": 0.4026,
+      "step": 800
+    },
+    {
+      "epoch": 1.8687572590011614,
+      "grad_norm": 1.0293341875076294,
+      "learning_rate": 2.2225901639832188e-05,
+      "loss": 0.419,
+      "step": 805
+    },
+    {
+      "epoch": 1.8803716608594656,
+      "grad_norm": 1.1112502813339233,
+      "learning_rate": 2.2124821108109515e-05,
+      "loss": 0.4005,
+      "step": 810
+    },
+    {
+      "epoch": 1.89198606271777,
+      "grad_norm": 0.9960114359855652,
+      "learning_rate": 2.2023321033944544e-05,
+      "loss": 0.328,
+      "step": 815
+    },
+    {
+      "epoch": 1.9036004645760745,
+      "grad_norm": 1.1106208562850952,
+      "learning_rate": 2.1921407394131406e-05,
+      "loss": 0.3667,
+      "step": 820
+    },
+    {
+      "epoch": 1.9152148664343787,
+      "grad_norm": 1.0423858165740967,
+      "learning_rate": 2.1819086189816893e-05,
+      "loss": 0.3773,
+      "step": 825
+    },
+    {
+      "epoch": 1.9268292682926829,
+      "grad_norm": 1.13150155544281,
+      "learning_rate": 2.171636344614708e-05,
+      "loss": 0.3491,
+      "step": 830
+    },
+    {
+      "epoch": 1.938443670150987,
+      "grad_norm": 1.0611273050308228,
+      "learning_rate": 2.1613245211912554e-05,
+      "loss": 0.3323,
+      "step": 835
+    },
+    {
+      "epoch": 1.9500580720092915,
+      "grad_norm": 0.9943997263908386,
+      "learning_rate": 2.1509737559192188e-05,
+      "loss": 0.3741,
+      "step": 840
+    },
+    {
+      "epoch": 1.961672473867596,
+      "grad_norm": 0.9484174847602844,
+      "learning_rate": 2.140584658299564e-05,
+      "loss": 0.3415,
+      "step": 845
+    },
+    {
+      "epoch": 1.9732868757259001,
+      "grad_norm": 1.0280919075012207,
+      "learning_rate": 2.1301578400904424e-05,
+      "loss": 0.3724,
+      "step": 850
+    },
+    {
+      "epoch": 1.9849012775842043,
+      "grad_norm": 1.1084401607513428,
+      "learning_rate": 2.119693915271168e-05,
+      "loss": 0.3867,
+      "step": 855
+    },
+    {
+      "epoch": 1.9965156794425087,
+      "grad_norm": 1.1195417642593384,
+      "learning_rate": 2.1091935000060637e-05,
+      "loss": 0.3641,
+      "step": 860
+    },
+    {
+      "epoch": 2.0069686411149825,
+      "grad_norm": 0.877273440361023,
+      "learning_rate": 2.0986572126081793e-05,
+      "loss": 0.3401,
+      "step": 865
+    },
+    {
+      "epoch": 2.0185830429732867,
+      "grad_norm": 1.3777830600738525,
+      "learning_rate": 2.0880856735028803e-05,
+      "loss": 0.3079,
+      "step": 870
+    },
+    {
+      "epoch": 2.0301974448315914,
+      "grad_norm": 1.1466902494430542,
+      "learning_rate": 2.0774795051913178e-05,
+      "loss": 0.2913,
+      "step": 875
+    },
+    {
+      "epoch": 2.0418118466898956,
+      "grad_norm": 1.0693480968475342,
+      "learning_rate": 2.0668393322137703e-05,
+      "loss": 0.2945,
+      "step": 880
+    },
+    {
+      "epoch": 2.0534262485481998,
+      "grad_norm": 1.0168899297714233,
+      "learning_rate": 2.056165781112867e-05,
+      "loss": 0.2828,
+      "step": 885
+    },
+    {
+      "epoch": 2.065040650406504,
+      "grad_norm": 1.0432965755462646,
+      "learning_rate": 2.0454594803966972e-05,
+      "loss": 0.2915,
+      "step": 890
+    },
+    {
+      "epoch": 2.076655052264808,
+      "grad_norm": 1.100029468536377,
+      "learning_rate": 2.034721060501799e-05,
+      "loss": 0.297,
+      "step": 895
+    },
+    {
+      "epoch": 2.088269454123113,
+      "grad_norm": 1.1748476028442383,
+      "learning_rate": 2.023951153756035e-05,
+      "loss": 0.2892,
+      "step": 900
+    },
+    {
+      "epoch": 2.099883855981417,
+      "grad_norm": 1.0221283435821533,
+      "learning_rate": 2.013150394341361e-05,
+      "loss": 0.2831,
+      "step": 905
+    },
+    {
+      "epoch": 2.1114982578397212,
+      "grad_norm": 0.9806298017501831,
+      "learning_rate": 2.002319418256479e-05,
+      "loss": 0.2785,
+      "step": 910
+    },
+    {
+      "epoch": 2.1231126596980254,
+      "grad_norm": 1.292575716972351,
+      "learning_rate": 1.9914588632793908e-05,
+      "loss": 0.2829,
+      "step": 915
+    },
+    {
+      "epoch": 2.1347270615563296,
+      "grad_norm": 1.0625466108322144,
+      "learning_rate": 1.9805693689298374e-05,
+      "loss": 0.2867,
+      "step": 920
+    },
+    {
+      "epoch": 2.1463414634146343,
+      "grad_norm": 1.0215439796447754,
+      "learning_rate": 1.969651576431646e-05,
+      "loss": 0.3223,
+      "step": 925
+    },
+    {
+      "epoch": 2.1579558652729385,
+      "grad_norm": 1.1626147031784058,
+      "learning_rate": 1.9587061286749693e-05,
+      "loss": 0.3042,
+      "step": 930
+    },
+    {
+      "epoch": 2.1695702671312427,
+      "grad_norm": 0.9875515699386597,
+      "learning_rate": 1.9477336701784292e-05,
+      "loss": 0.2827,
+      "step": 935
+    },
+    {
+      "epoch": 2.181184668989547,
+      "grad_norm": 1.0031932592391968,
+      "learning_rate": 1.9367348470511652e-05,
+      "loss": 0.2712,
+      "step": 940
+    },
+    {
+      "epoch": 2.1927990708478515,
+      "grad_norm": 1.1925930976867676,
+      "learning_rate": 1.9257103069547894e-05,
+      "loss": 0.2895,
+      "step": 945
+    },
+    {
+      "epoch": 2.2044134727061557,
+      "grad_norm": 1.217810869216919,
+      "learning_rate": 1.9146606990652452e-05,
+      "loss": 0.2714,
+      "step": 950
+    },
+    {
+      "epoch": 2.21602787456446,
+      "grad_norm": 0.9798834323883057,
+      "learning_rate": 1.903586674034587e-05,
+      "loss": 0.2665,
+      "step": 955
+    },
+    {
+      "epoch": 2.227642276422764,
+      "grad_norm": 1.2851895093917847,
+      "learning_rate": 1.8924888839526612e-05,
+      "loss": 0.2795,
+      "step": 960
+    },
+    {
+      "epoch": 2.2392566782810683,
+      "grad_norm": 1.1440101861953735,
+      "learning_rate": 1.881367982308711e-05,
+      "loss": 0.2779,
+      "step": 965
+    },
+    {
+      "epoch": 2.250871080139373,
+      "grad_norm": 1.3127503395080566,
+      "learning_rate": 1.8702246239528965e-05,
+      "loss": 0.2638,
+      "step": 970
+    },
+    {
+      "epoch": 2.262485481997677,
+      "grad_norm": 1.0733845233917236,
+      "learning_rate": 1.8590594650577314e-05,
+      "loss": 0.2646,
+      "step": 975
+    },
+    {
+      "epoch": 2.2740998838559814,
+      "grad_norm": 1.0694085359573364,
+      "learning_rate": 1.847873163079446e-05,
+      "loss": 0.2741,
+      "step": 980
+    },
+    {
+      "epoch": 2.2857142857142856,
+      "grad_norm": 1.1703122854232788,
+      "learning_rate": 1.8366663767192747e-05,
+      "loss": 0.2374,
+      "step": 985
+    },
+    {
+      "epoch": 2.2973286875725902,
+      "grad_norm": 1.2135448455810547,
+      "learning_rate": 1.8254397658846656e-05,
+      "loss": 0.271,
+      "step": 990
+    },
+    {
+      "epoch": 2.3089430894308944,
+      "grad_norm": 1.1723593473434448,
+      "learning_rate": 1.8141939916504245e-05,
+      "loss": 0.2744,
+      "step": 995
+    },
+    {
+      "epoch": 2.3205574912891986,
+      "grad_norm": 1.2068791389465332,
+      "learning_rate": 1.8029297162197867e-05,
+      "loss": 0.2597,
+      "step": 1000
+    },
+    {
+      "epoch": 2.332171893147503,
+      "grad_norm": 1.1227343082427979,
+      "learning_rate": 1.791647602885423e-05,
+      "loss": 0.254,
+      "step": 1005
+    },
+    {
+      "epoch": 2.343786295005807,
+      "grad_norm": 1.1190464496612549,
+      "learning_rate": 1.7803483159903835e-05,
+      "loss": 0.232,
+      "step": 1010
+    },
+    {
+      "epoch": 2.3554006968641117,
+      "grad_norm": 1.0415700674057007,
+      "learning_rate": 1.769032520888976e-05,
+      "loss": 0.2613,
+      "step": 1015
+    },
+    {
+      "epoch": 2.367015098722416,
+      "grad_norm": 1.188915491104126,
+      "learning_rate": 1.7577008839075886e-05,
+      "loss": 0.2392,
+      "step": 1020
+    },
+    {
+      "epoch": 2.37862950058072,
+      "grad_norm": 1.1599727869033813,
+      "learning_rate": 1.7463540723054526e-05,
+      "loss": 0.2607,
+      "step": 1025
+    },
+    {
+      "epoch": 2.3902439024390243,
+      "grad_norm": 1.0613243579864502,
+      "learning_rate": 1.734992754235351e-05,
+      "loss": 0.2517,
+      "step": 1030
+    },
+    {
+      "epoch": 2.4018583042973285,
+      "grad_norm": 1.140293836593628,
+      "learning_rate": 1.7236175987042744e-05,
+      "loss": 0.2757,
+      "step": 1035
+    },
+    {
+      "epoch": 2.413472706155633,
+      "grad_norm": 1.217897891998291,
+      "learning_rate": 1.7122292755340284e-05,
+      "loss": 0.2604,
+      "step": 1040
+    },
+    {
+      "epoch": 2.4250871080139373,
+      "grad_norm": 1.2544645071029663,
+      "learning_rate": 1.700828455321788e-05,
+      "loss": 0.2539,
+      "step": 1045
+    },
+    {
+      "epoch": 2.4367015098722415,
+      "grad_norm": 1.5165183544158936,
+      "learning_rate": 1.6894158094006142e-05,
+      "loss": 0.2392,
+      "step": 1050
+    },
+    {
+      "epoch": 2.4483159117305457,
+      "grad_norm": 1.0610485076904297,
+      "learning_rate": 1.6779920097999194e-05,
+      "loss": 0.2678,
+      "step": 1055
+    },
+    {
+      "epoch": 2.45993031358885,
+      "grad_norm": 1.1240959167480469,
+      "learning_rate": 1.6665577292058955e-05,
+      "loss": 0.2425,
+      "step": 1060
+    },
+    {
+      "epoch": 2.4715447154471546,
+      "grad_norm": 1.2169679403305054,
+      "learning_rate": 1.655113640921905e-05,
+      "loss": 0.2425,
+      "step": 1065
+    },
+    {
+      "epoch": 2.483159117305459,
+      "grad_norm": 1.0992380380630493,
+      "learning_rate": 1.6436604188288325e-05,
+      "loss": 0.2537,
+      "step": 1070
+    },
+    {
+      "epoch": 2.494773519163763,
+      "grad_norm": 1.1832488775253296,
+      "learning_rate": 1.632198737345403e-05,
+      "loss": 0.2596,
+      "step": 1075
+    },
+    {
+      "epoch": 2.506387921022067,
+      "grad_norm": 0.9653327465057373,
+      "learning_rate": 1.6207292713884692e-05,
+      "loss": 0.2427,
+      "step": 1080
+    },
+    {
+      "epoch": 2.5180023228803714,
+      "grad_norm": 1.139378547668457,
+      "learning_rate": 1.609252696333271e-05,
+      "loss": 0.2432,
+      "step": 1085
+    },
+    {
+      "epoch": 2.529616724738676,
+      "grad_norm": 1.0014028549194336,
+      "learning_rate": 1.597769687973664e-05,
+      "loss": 0.2329,
+      "step": 1090
+    },
+    {
+      "epoch": 2.5412311265969802,
+      "grad_norm": 1.1616328954696655,
+      "learning_rate": 1.586280922482327e-05,
+      "loss": 0.2513,
+      "step": 1095
+    },
+    {
+      "epoch": 2.5528455284552845,
+      "grad_norm": 1.022898554801941,
+      "learning_rate": 1.5747870763709432e-05,
+      "loss": 0.2697,
+      "step": 1100
+    },
+    {
+      "epoch": 2.564459930313589,
+      "grad_norm": 1.102315902709961,
+      "learning_rate": 1.5632888264503705e-05,
+      "loss": 0.2368,
+      "step": 1105
+    },
+    {
+      "epoch": 2.5760743321718933,
+      "grad_norm": 1.0889406204223633,
+      "learning_rate": 1.5517868497907795e-05,
+      "loss": 0.2377,
+      "step": 1110
+    },
+    {
+      "epoch": 2.5876887340301975,
+      "grad_norm": 1.0581204891204834,
+      "learning_rate": 1.54028182368179e-05,
+      "loss": 0.2424,
+      "step": 1115
+    },
+    {
+      "epoch": 2.5993031358885017,
+      "grad_norm": 1.180722713470459,
+      "learning_rate": 1.5287744255925882e-05,
+      "loss": 0.2191,
+      "step": 1120
+    },
+    {
+      "epoch": 2.610917537746806,
+      "grad_norm": 1.0064940452575684,
+      "learning_rate": 1.5172653331320318e-05,
+      "loss": 0.2718,
+      "step": 1125
+    },
+    {
+      "epoch": 2.6225319396051106,
+      "grad_norm": 1.1252609491348267,
+      "learning_rate": 1.5057552240087526e-05,
+      "loss": 0.2248,
+      "step": 1130
+    },
+    {
+      "epoch": 2.6341463414634148,
+      "grad_norm": 1.0868943929672241,
+      "learning_rate": 1.4942447759912473e-05,
+      "loss": 0.2609,
+      "step": 1135
+    },
+    {
+      "epoch": 2.645760743321719,
+      "grad_norm": 1.127143144607544,
+      "learning_rate": 1.4827346668679683e-05,
+      "loss": 0.233,
+      "step": 1140
+    },
+    {
+      "epoch": 2.657375145180023,
+      "grad_norm": 1.2413113117218018,
+      "learning_rate": 1.471225574407412e-05,
+      "loss": 0.2043,
+      "step": 1145
+    },
+    {
+      "epoch": 2.6689895470383274,
+      "grad_norm": 1.1059684753417969,
+      "learning_rate": 1.4597181763182104e-05,
+      "loss": 0.2375,
+      "step": 1150
+    },
+    {
+      "epoch": 2.680603948896632,
+      "grad_norm": 1.1019221544265747,
+      "learning_rate": 1.4482131502092208e-05,
+      "loss": 0.1922,
+      "step": 1155
+    },
+    {
+      "epoch": 2.692218350754936,
+      "grad_norm": 1.1478207111358643,
+      "learning_rate": 1.4367111735496298e-05,
+      "loss": 0.2286,
+      "step": 1160
+    },
+    {
+      "epoch": 2.7038327526132404,
+      "grad_norm": 1.0406802892684937,
+      "learning_rate": 1.4252129236290569e-05,
+      "loss": 0.2078,
+      "step": 1165
+    },
+    {
+      "epoch": 2.7154471544715446,
+      "grad_norm": 1.2635127305984497,
+      "learning_rate": 1.4137190775176739e-05,
+      "loss": 0.2146,
+      "step": 1170
+    },
+    {
+      "epoch": 2.727061556329849,
+      "grad_norm": 1.0526198148727417,
+      "learning_rate": 1.402230312026336e-05,
+      "loss": 0.2103,
+      "step": 1175
+    },
+    {
+      "epoch": 2.7386759581881535,
+      "grad_norm": 1.297333836555481,
+      "learning_rate": 1.390747303666729e-05,
+      "loss": 0.2183,
+      "step": 1180
+    },
+    {
+      "epoch": 2.7502903600464577,
+      "grad_norm": 1.0738940238952637,
+      "learning_rate": 1.3792707286115309e-05,
+      "loss": 0.2077,
+      "step": 1185
+    },
+    {
+      "epoch": 2.761904761904762,
+      "grad_norm": 1.0152521133422852,
+      "learning_rate": 1.3678012626545978e-05,
+      "loss": 0.2368,
+      "step": 1190
+    },
+    {
+      "epoch": 2.773519163763066,
+      "grad_norm": 1.0068998336791992,
+      "learning_rate": 1.3563395811711678e-05,
+      "loss": 0.2152,
+      "step": 1195
+    },
+    {
+      "epoch": 2.7851335656213703,
+      "grad_norm": 1.1199923753738403,
+      "learning_rate": 1.344886359078095e-05,
+      "loss": 0.2188,
+      "step": 1200
+    },
+    {
+      "epoch": 2.796747967479675,
+      "grad_norm": 1.037109613418579,
+      "learning_rate": 1.3334422707941051e-05,
+      "loss": 0.2286,
+      "step": 1205
+    },
+    {
+      "epoch": 2.808362369337979,
+      "grad_norm": 0.9589251279830933,
+      "learning_rate": 1.3220079902000811e-05,
+      "loss": 0.2025,
+      "step": 1210
+    },
+    {
+      "epoch": 2.8199767711962833,
+      "grad_norm": 1.2464582920074463,
+      "learning_rate": 1.3105841905993857e-05,
+      "loss": 0.2131,
+      "step": 1215
+    },
+    {
+      "epoch": 2.831591173054588,
+      "grad_norm": 1.091057538986206,
+      "learning_rate": 1.2991715446782121e-05,
+      "loss": 0.2067,
+      "step": 1220
+    },
+    {
+      "epoch": 2.8432055749128917,
+      "grad_norm": 1.278668761253357,
+      "learning_rate": 1.2877707244659722e-05,
+      "loss": 0.2209,
+      "step": 1225
+    },
+    {
+      "epoch": 2.8548199767711964,
+      "grad_norm": 1.2127578258514404,
+      "learning_rate": 1.2763824012957258e-05,
+      "loss": 0.1833,
+      "step": 1230
+    },
+    {
+      "epoch": 2.8664343786295006,
+      "grad_norm": 1.146657109260559,
+      "learning_rate": 1.2650072457646492e-05,
+      "loss": 0.1814,
+      "step": 1235
+    },
+    {
+      "epoch": 2.8780487804878048,
+      "grad_norm": 1.0890151262283325,
+      "learning_rate": 1.2536459276945473e-05,
+      "loss": 0.1801,
+      "step": 1240
+    },
+    {
+      "epoch": 2.8896631823461094,
+      "grad_norm": 0.9708114266395569,
+      "learning_rate": 1.2422991160924117e-05,
+      "loss": 0.1946,
+      "step": 1245
+    },
+    {
+      "epoch": 2.9012775842044136,
+      "grad_norm": 1.1722114086151123,
+      "learning_rate": 1.2309674791110244e-05,
+      "loss": 0.2438,
+      "step": 1250
+    },
+    {
+      "epoch": 2.912891986062718,
+      "grad_norm": 1.015057921409607,
+      "learning_rate": 1.2196516840096167e-05,
+      "loss": 0.2035,
+      "step": 1255
+    },
+    {
+      "epoch": 2.924506387921022,
+      "grad_norm": 0.9809845685958862,
+      "learning_rate": 1.2083523971145773e-05,
+      "loss": 0.189,
+      "step": 1260
+    },
+    {
+      "epoch": 2.9361207897793262,
+      "grad_norm": 1.2840055227279663,
+      "learning_rate": 1.1970702837802137e-05,
+      "loss": 0.2055,
+      "step": 1265
+    },
+    {
+      "epoch": 2.947735191637631,
+      "grad_norm": 0.9905469417572021,
+      "learning_rate": 1.1858060083495761e-05,
+      "loss": 0.1822,
+      "step": 1270
+    },
+    {
+      "epoch": 2.959349593495935,
+      "grad_norm": 1.1354817152023315,
+      "learning_rate": 1.1745602341153347e-05,
+      "loss": 0.1883,
+      "step": 1275
+    },
+    {
+      "epoch": 2.9709639953542393,
+      "grad_norm": 1.3380357027053833,
+      "learning_rate": 1.1633336232807256e-05,
+      "loss": 0.1874,
+      "step": 1280
+    },
+    {
+      "epoch": 2.9825783972125435,
+      "grad_norm": 1.232898235321045,
+      "learning_rate": 1.1521268369205546e-05,
+      "loss": 0.177,
+      "step": 1285
+    },
+    {
+      "epoch": 2.9941927990708477,
+      "grad_norm": 1.0560810565948486,
+      "learning_rate": 1.1409405349422693e-05,
+      "loss": 0.1929,
+      "step": 1290
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2155,
+  "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": 2.035566742712877e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

25_128_e5_3e-5/checkpoint-1293/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8826cd3cb0315b085e0bee4ad616aa4a2c50addf0b6c688836791cb1c313dff8
+size 7736

25_128_e5_3e-5/checkpoint-1293/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)

25_128_e5_3e-5/checkpoint-1724/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

25_128_e5_3e-5/checkpoint-1724/adapter_config.json

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

25_128_e5_3e-5/checkpoint-1724/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a77585e0aa773ae4495e472c3258e769d333e0bd1cbc39afcb4acb82fd3512f
+size 791751704

25_128_e5_3e-5/checkpoint-1724/latest

@@ -0,0 +1 @@
+global_step1724

25_128_e5_3e-5/checkpoint-1724/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5439a6035718d96ac1737e638df2e77cf2ce17258a4bc6c636e9a2d82e6afd15
+size 15920

25_128_e5_3e-5/checkpoint-1724/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b563e9b7b0bb1fe54736f943ebf5a964c7692be9281dc73be887b201014e19fd
+size 15920

25_128_e5_3e-5/checkpoint-1724/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4f6008fee9f92770297d74125d890c9249624b41d2c5920ce0dfa41e7c935ae
+size 15920

25_128_e5_3e-5/checkpoint-1724/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec62a0cd71615ee11bfed472cf444a654d12ac786aa89f7bf59e60e5b0a57042
+size 15920

25_128_e5_3e-5/checkpoint-1724/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f2aca6c276107ece231276a36efd32f1a702bf656c576fa3d1a889e9bae6a51
+size 15920

25_128_e5_3e-5/checkpoint-1724/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e059f01b8b3afb064aa8c321ad686eba584fb81806f56be1099b64a0bdc9396
+size 15920

25_128_e5_3e-5/checkpoint-1724/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59b6e030ba6fa792b3615f0e5cd9ccfbed7c5bf1607fb0d35661a956cfdef75e
+size 15920

25_128_e5_3e-5/checkpoint-1724/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6fa05efae1e71db67fc4ab3d46a75c17f580a10cc9cf1baaa2dbaedc33b365a
+size 15920

25_128_e5_3e-5/checkpoint-1724/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ef8c305b0b7681d6d6c9e3c5388f9c97dcf6bb0d2f66802eb0e6eabbbec7a5d
+size 1064

25_128_e5_3e-5/checkpoint-1724/special_tokens_map.json

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

25_128_e5_3e-5/checkpoint-1724/tokenizer_config.json

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

25_128_e5_3e-5/checkpoint-1724/trainer_state.json

@@ -0,0 +1,2442 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1724,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011614401858304297,
+      "grad_norm": 1.2486553192138672,
+      "learning_rate": 1.111111111111111e-06,
+      "loss": 1.3051,
+      "step": 5
+    },
+    {
+      "epoch": 0.023228803716608595,
+      "grad_norm": 1.02664053440094,
+      "learning_rate": 2.4999999999999998e-06,
+      "loss": 1.3323,
+      "step": 10
+    },
+    {
+      "epoch": 0.03484320557491289,
+      "grad_norm": 0.8225151896476746,
+      "learning_rate": 3.888888888888889e-06,
+      "loss": 1.3047,
+      "step": 15
+    },
+    {
+      "epoch": 0.04645760743321719,
+      "grad_norm": 0.7819671630859375,
+      "learning_rate": 5.277777777777778e-06,
+      "loss": 1.2887,
+      "step": 20
+    },
+    {
+      "epoch": 0.05807200929152149,
+      "grad_norm": 0.6035093069076538,
+      "learning_rate": 6.666666666666667e-06,
+      "loss": 1.3248,
+      "step": 25
+    },
+    {
+      "epoch": 0.06968641114982578,
+      "grad_norm": 0.5274394750595093,
+      "learning_rate": 8.055555555555557e-06,
+      "loss": 1.2981,
+      "step": 30
+    },
+    {
+      "epoch": 0.08130081300813008,
+      "grad_norm": 0.4473659098148346,
+      "learning_rate": 9.444444444444445e-06,
+      "loss": 1.2253,
+      "step": 35
+    },
+    {
+      "epoch": 0.09291521486643438,
+      "grad_norm": 0.5344942808151245,
+      "learning_rate": 1.0833333333333334e-05,
+      "loss": 1.2101,
+      "step": 40
+    },
+    {
+      "epoch": 0.10452961672473868,
+      "grad_norm": 0.42955347895622253,
+      "learning_rate": 1.2222222222222222e-05,
+      "loss": 1.2136,
+      "step": 45
+    },
+    {
+      "epoch": 0.11614401858304298,
+      "grad_norm": 0.4606517553329468,
+      "learning_rate": 1.3611111111111111e-05,
+      "loss": 1.2501,
+      "step": 50
+    },
+    {
+      "epoch": 0.12775842044134728,
+      "grad_norm": 0.43243467807769775,
+      "learning_rate": 1.5e-05,
+      "loss": 1.2287,
+      "step": 55
+    },
+    {
+      "epoch": 0.13937282229965156,
+      "grad_norm": 0.45852982997894287,
+      "learning_rate": 1.638888888888889e-05,
+      "loss": 1.2272,
+      "step": 60
+    },
+    {
+      "epoch": 0.15098722415795587,
+      "grad_norm": 0.422735333442688,
+      "learning_rate": 1.7777777777777777e-05,
+      "loss": 1.1591,
+      "step": 65
+    },
+    {
+      "epoch": 0.16260162601626016,
+      "grad_norm": 0.48696285486221313,
+      "learning_rate": 1.9166666666666667e-05,
+      "loss": 1.2238,
+      "step": 70
+    },
+    {
+      "epoch": 0.17421602787456447,
+      "grad_norm": 0.6145470142364502,
+      "learning_rate": 2.0555555555555558e-05,
+      "loss": 1.2375,
+      "step": 75
+    },
+    {
+      "epoch": 0.18583042973286876,
+      "grad_norm": 0.5468384623527527,
+      "learning_rate": 2.1944444444444445e-05,
+      "loss": 1.1484,
+      "step": 80
+    },
+    {
+      "epoch": 0.19744483159117304,
+      "grad_norm": 0.4397794008255005,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 1.1567,
+      "step": 85
+    },
+    {
+      "epoch": 0.20905923344947736,
+      "grad_norm": 0.4396965801715851,
+      "learning_rate": 2.4722222222222223e-05,
+      "loss": 1.1665,
+      "step": 90
+    },
+    {
+      "epoch": 0.22067363530778164,
+      "grad_norm": 0.5894995927810669,
+      "learning_rate": 2.611111111111111e-05,
+      "loss": 1.1519,
+      "step": 95
+    },
+    {
+      "epoch": 0.23228803716608595,
+      "grad_norm": 0.4796619117259979,
+      "learning_rate": 2.75e-05,
+      "loss": 1.1242,
+      "step": 100
+    },
+    {
+      "epoch": 0.24390243902439024,
+      "grad_norm": 0.7274127006530762,
+      "learning_rate": 2.8888888888888888e-05,
+      "loss": 1.1197,
+      "step": 105
+    },
+    {
+      "epoch": 0.25551684088269455,
+      "grad_norm": 0.5538251996040344,
+      "learning_rate": 2.999998233452831e-05,
+      "loss": 1.09,
+      "step": 110
+    },
+    {
+      "epoch": 0.26713124274099886,
+      "grad_norm": 0.4963814616203308,
+      "learning_rate": 2.999936404738799e-05,
+      "loss": 1.14,
+      "step": 115
+    },
+    {
+      "epoch": 0.2787456445993031,
+      "grad_norm": 0.5045656561851501,
+      "learning_rate": 2.9997862528271754e-05,
+      "loss": 1.0846,
+      "step": 120
+    },
+    {
+      "epoch": 0.29036004645760743,
+      "grad_norm": 0.5694212913513184,
+      "learning_rate": 2.999547786559598e-05,
+      "loss": 1.0566,
+      "step": 125
+    },
+    {
+      "epoch": 0.30197444831591175,
+      "grad_norm": 0.5698084831237793,
+      "learning_rate": 2.9992210199780657e-05,
+      "loss": 1.0812,
+      "step": 130
+    },
+    {
+      "epoch": 0.313588850174216,
+      "grad_norm": 0.5170321464538574,
+      "learning_rate": 2.9988059723241064e-05,
+      "loss": 1.0414,
+      "step": 135
+    },
+    {
+      "epoch": 0.3252032520325203,
+      "grad_norm": 0.5222340226173401,
+      "learning_rate": 2.9983026680376472e-05,
+      "loss": 1.0631,
+      "step": 140
+    },
+    {
+      "epoch": 0.33681765389082463,
+      "grad_norm": 0.5456163883209229,
+      "learning_rate": 2.997711136755574e-05,
+      "loss": 1.0866,
+      "step": 145
+    },
+    {
+      "epoch": 0.34843205574912894,
+      "grad_norm": 0.5397904515266418,
+      "learning_rate": 2.9970314133099855e-05,
+      "loss": 1.1579,
+      "step": 150
+    },
+    {
+      "epoch": 0.3600464576074332,
+      "grad_norm": 0.6353335380554199,
+      "learning_rate": 2.9962635377261457e-05,
+      "loss": 0.9954,
+      "step": 155
+    },
+    {
+      "epoch": 0.3716608594657375,
+      "grad_norm": 0.6079632639884949,
+      "learning_rate": 2.9954075552201222e-05,
+      "loss": 1.0647,
+      "step": 160
+    },
+    {
+      "epoch": 0.3832752613240418,
+      "grad_norm": 0.6199746131896973,
+      "learning_rate": 2.994463516196126e-05,
+      "loss": 1.0637,
+      "step": 165
+    },
+    {
+      "epoch": 0.3948896631823461,
+      "grad_norm": 0.6312094926834106,
+      "learning_rate": 2.9934314762435444e-05,
+      "loss": 1.0242,
+      "step": 170
+    },
+    {
+      "epoch": 0.4065040650406504,
+      "grad_norm": 0.677851140499115,
+      "learning_rate": 2.9923114961336672e-05,
+      "loss": 0.9563,
+      "step": 175
+    },
+    {
+      "epoch": 0.4181184668989547,
+      "grad_norm": 0.6415532827377319,
+      "learning_rate": 2.9911036418161058e-05,
+      "loss": 0.9728,
+      "step": 180
+    },
+    {
+      "epoch": 0.429732868757259,
+      "grad_norm": 0.7620236873626709,
+      "learning_rate": 2.9898079844149132e-05,
+      "loss": 1.0066,
+      "step": 185
+    },
+    {
+      "epoch": 0.4413472706155633,
+      "grad_norm": 0.6720292568206787,
+      "learning_rate": 2.9884246002243936e-05,
+      "loss": 1.0059,
+      "step": 190
+    },
+    {
+      "epoch": 0.4529616724738676,
+      "grad_norm": 0.7387037873268127,
+      "learning_rate": 2.9869535707046104e-05,
+      "loss": 0.9314,
+      "step": 195
+    },
+    {
+      "epoch": 0.4645760743321719,
+      "grad_norm": 0.6071408987045288,
+      "learning_rate": 2.985394982476591e-05,
+      "loss": 0.9805,
+      "step": 200
+    },
+    {
+      "epoch": 0.47619047619047616,
+      "grad_norm": 0.7038552165031433,
+      "learning_rate": 2.9837489273172232e-05,
+      "loss": 0.9766,
+      "step": 205
+    },
+    {
+      "epoch": 0.4878048780487805,
+      "grad_norm": 0.7892246842384338,
+      "learning_rate": 2.9820155021538533e-05,
+      "loss": 0.9558,
+      "step": 210
+    },
+    {
+      "epoch": 0.4994192799070848,
+      "grad_norm": 0.7563413381576538,
+      "learning_rate": 2.980194809058577e-05,
+      "loss": 0.9455,
+      "step": 215
+    },
+    {
+      "epoch": 0.5110336817653891,
+      "grad_norm": 0.7255906462669373,
+      "learning_rate": 2.9782869552422316e-05,
+      "loss": 0.9183,
+      "step": 220
+    },
+    {
+      "epoch": 0.5226480836236934,
+      "grad_norm": 0.7455199956893921,
+      "learning_rate": 2.9762920530480788e-05,
+      "loss": 0.9363,
+      "step": 225
+    },
+    {
+      "epoch": 0.5342624854819977,
+      "grad_norm": 0.7163615822792053,
+      "learning_rate": 2.974210219945193e-05,
+      "loss": 0.9436,
+      "step": 230
+    },
+    {
+      "epoch": 0.5458768873403019,
+      "grad_norm": 0.6869027018547058,
+      "learning_rate": 2.9720415785215428e-05,
+      "loss": 0.8932,
+      "step": 235
+    },
+    {
+      "epoch": 0.5574912891986062,
+      "grad_norm": 0.776175320148468,
+      "learning_rate": 2.969786256476772e-05,
+      "loss": 0.9234,
+      "step": 240
+    },
+    {
+      "epoch": 0.5691056910569106,
+      "grad_norm": 0.687179446220398,
+      "learning_rate": 2.9674443866146807e-05,
+      "loss": 0.9165,
+      "step": 245
+    },
+    {
+      "epoch": 0.5807200929152149,
+      "grad_norm": 0.8298389911651611,
+      "learning_rate": 2.9650161068354054e-05,
+      "loss": 0.8863,
+      "step": 250
+    },
+    {
+      "epoch": 0.5923344947735192,
+      "grad_norm": 0.7793598175048828,
+      "learning_rate": 2.9625015601272974e-05,
+      "loss": 0.8693,
+      "step": 255
+    },
+    {
+      "epoch": 0.6039488966318235,
+      "grad_norm": 0.9237378835678101,
+      "learning_rate": 2.9599008945585066e-05,
+      "loss": 0.8435,
+      "step": 260
+    },
+    {
+      "epoch": 0.6155632984901278,
+      "grad_norm": 0.8277608752250671,
+      "learning_rate": 2.9572142632682562e-05,
+      "loss": 0.8683,
+      "step": 265
+    },
+    {
+      "epoch": 0.627177700348432,
+      "grad_norm": 0.7906745076179504,
+      "learning_rate": 2.954441824457832e-05,
+      "loss": 0.9144,
+      "step": 270
+    },
+    {
+      "epoch": 0.6387921022067363,
+      "grad_norm": 0.8107167482376099,
+      "learning_rate": 2.951583741381263e-05,
+      "loss": 0.8696,
+      "step": 275
+    },
+    {
+      "epoch": 0.6504065040650406,
+      "grad_norm": 0.734391987323761,
+      "learning_rate": 2.948640182335708e-05,
+      "loss": 0.871,
+      "step": 280
+    },
+    {
+      "epoch": 0.662020905923345,
+      "grad_norm": 0.8837676048278809,
+      "learning_rate": 2.9456113206515475e-05,
+      "loss": 0.8883,
+      "step": 285
+    },
+    {
+      "epoch": 0.6736353077816493,
+      "grad_norm": 0.8748791217803955,
+      "learning_rate": 2.942497334682176e-05,
+      "loss": 0.7801,
+      "step": 290
+    },
+    {
+      "epoch": 0.6852497096399536,
+      "grad_norm": 0.8546818494796753,
+      "learning_rate": 2.9392984077934987e-05,
+      "loss": 0.8724,
+      "step": 295
+    },
+    {
+      "epoch": 0.6968641114982579,
+      "grad_norm": 0.7864014506340027,
+      "learning_rate": 2.9360147283531373e-05,
+      "loss": 0.8214,
+      "step": 300
+    },
+    {
+      "epoch": 0.7084785133565621,
+      "grad_norm": 0.7984604835510254,
+      "learning_rate": 2.9326464897193343e-05,
+      "loss": 0.836,
+      "step": 305
+    },
+    {
+      "epoch": 0.7200929152148664,
+      "grad_norm": 0.9750844836235046,
+      "learning_rate": 2.9291938902295695e-05,
+      "loss": 0.7873,
+      "step": 310
+    },
+    {
+      "epoch": 0.7317073170731707,
+      "grad_norm": 0.8499948382377625,
+      "learning_rate": 2.925657133188881e-05,
+      "loss": 0.8412,
+      "step": 315
+    },
+    {
+      "epoch": 0.743321718931475,
+      "grad_norm": 0.86067134141922,
+      "learning_rate": 2.9220364268578922e-05,
+      "loss": 0.8256,
+      "step": 320
+    },
+    {
+      "epoch": 0.7549361207897793,
+      "grad_norm": 0.8807356357574463,
+      "learning_rate": 2.918331984440549e-05,
+      "loss": 0.8364,
+      "step": 325
+    },
+    {
+      "epoch": 0.7665505226480837,
+      "grad_norm": 0.8951946496963501,
+      "learning_rate": 2.9145440240715657e-05,
+      "loss": 0.8415,
+      "step": 330
+    },
+    {
+      "epoch": 0.778164924506388,
+      "grad_norm": 0.8705118894577026,
+      "learning_rate": 2.9106727688035814e-05,
+      "loss": 0.7762,
+      "step": 335
+    },
+    {
+      "epoch": 0.7897793263646922,
+      "grad_norm": 0.9147759675979614,
+      "learning_rate": 2.9067184465940225e-05,
+      "loss": 0.7983,
+      "step": 340
+    },
+    {
+      "epoch": 0.8013937282229965,
+      "grad_norm": 0.8624527454376221,
+      "learning_rate": 2.9026812902916834e-05,
+      "loss": 0.7434,
+      "step": 345
+    },
+    {
+      "epoch": 0.8130081300813008,
+      "grad_norm": 0.9535494446754456,
+      "learning_rate": 2.898561537623011e-05,
+      "loss": 0.8049,
+      "step": 350
+    },
+    {
+      "epoch": 0.8246225319396051,
+      "grad_norm": 0.836469292640686,
+      "learning_rate": 2.8943594311781104e-05,
+      "loss": 0.777,
+      "step": 355
+    },
+    {
+      "epoch": 0.8362369337979094,
+      "grad_norm": 0.9156211018562317,
+      "learning_rate": 2.8900752183964573e-05,
+      "loss": 0.7746,
+      "step": 360
+    },
+    {
+      "epoch": 0.8478513356562137,
+      "grad_norm": 0.9176764488220215,
+      "learning_rate": 2.8857091515523287e-05,
+      "loss": 0.7827,
+      "step": 365
+    },
+    {
+      "epoch": 0.859465737514518,
+      "grad_norm": 0.8853964805603027,
+      "learning_rate": 2.8812614877399476e-05,
+      "loss": 0.754,
+      "step": 370
+    },
+    {
+      "epoch": 0.8710801393728222,
+      "grad_norm": 0.8558771014213562,
+      "learning_rate": 2.876732488858344e-05,
+      "loss": 0.7039,
+      "step": 375
+    },
+    {
+      "epoch": 0.8826945412311266,
+      "grad_norm": 0.8176305294036865,
+      "learning_rate": 2.8721224215959335e-05,
+      "loss": 0.7892,
+      "step": 380
+    },
+    {
+      "epoch": 0.8943089430894309,
+      "grad_norm": 0.9530948996543884,
+      "learning_rate": 2.8674315574148126e-05,
+      "loss": 0.7049,
+      "step": 385
+    },
+    {
+      "epoch": 0.9059233449477352,
+      "grad_norm": 0.9136149287223816,
+      "learning_rate": 2.862660172534776e-05,
+      "loss": 0.7614,
+      "step": 390
+    },
+    {
+      "epoch": 0.9175377468060395,
+      "grad_norm": 0.9666270017623901,
+      "learning_rate": 2.8578085479170478e-05,
+      "loss": 0.7265,
+      "step": 395
+    },
+    {
+      "epoch": 0.9291521486643438,
+      "grad_norm": 0.9841073751449585,
+      "learning_rate": 2.85287696924774e-05,
+      "loss": 0.6903,
+      "step": 400
+    },
+    {
+      "epoch": 0.9407665505226481,
+      "grad_norm": 1.1007276773452759,
+      "learning_rate": 2.8478657269210294e-05,
+      "loss": 0.7114,
+      "step": 405
+    },
+    {
+      "epoch": 0.9523809523809523,
+      "grad_norm": 0.9525352120399475,
+      "learning_rate": 2.8427751160220573e-05,
+      "loss": 0.7115,
+      "step": 410
+    },
+    {
+      "epoch": 0.9639953542392566,
+      "grad_norm": 0.8143008351325989,
+      "learning_rate": 2.8376054363095545e-05,
+      "loss": 0.6784,
+      "step": 415
+    },
+    {
+      "epoch": 0.975609756097561,
+      "grad_norm": 0.9916719198226929,
+      "learning_rate": 2.8323569921981885e-05,
+      "loss": 0.7366,
+      "step": 420
+    },
+    {
+      "epoch": 0.9872241579558653,
+      "grad_norm": 0.9247941374778748,
+      "learning_rate": 2.82703009274064e-05,
+      "loss": 0.6974,
+      "step": 425
+    },
+    {
+      "epoch": 0.9988385598141696,
+      "grad_norm": 0.9906225800514221,
+      "learning_rate": 2.8216250516094027e-05,
+      "loss": 0.6673,
+      "step": 430
+    },
+    {
+      "epoch": 1.0092915214866434,
+      "grad_norm": 0.9208662509918213,
+      "learning_rate": 2.816142187078315e-05,
+      "loss": 0.603,
+      "step": 435
+    },
+    {
+      "epoch": 1.0209059233449478,
+      "grad_norm": 1.0540852546691895,
+      "learning_rate": 2.8105818220038167e-05,
+      "loss": 0.6062,
+      "step": 440
+    },
+    {
+      "epoch": 1.032520325203252,
+      "grad_norm": 0.9971239566802979,
+      "learning_rate": 2.804944283805938e-05,
+      "loss": 0.5986,
+      "step": 445
+    },
+    {
+      "epoch": 1.0441347270615564,
+      "grad_norm": 1.0977699756622314,
+      "learning_rate": 2.7992299044490192e-05,
+      "loss": 0.5556,
+      "step": 450
+    },
+    {
+      "epoch": 1.0557491289198606,
+      "grad_norm": 1.0428180694580078,
+      "learning_rate": 2.793439020422165e-05,
+      "loss": 0.599,
+      "step": 455
+    },
+    {
+      "epoch": 1.0673635307781648,
+      "grad_norm": 1.2783362865447998,
+      "learning_rate": 2.787571972719429e-05,
+      "loss": 0.5589,
+      "step": 460
+    },
+    {
+      "epoch": 1.0789779326364692,
+      "grad_norm": 1.0532207489013672,
+      "learning_rate": 2.781629106819733e-05,
+      "loss": 0.5534,
+      "step": 465
+    },
+    {
+      "epoch": 1.0905923344947734,
+      "grad_norm": 1.0525319576263428,
+      "learning_rate": 2.775610772666527e-05,
+      "loss": 0.577,
+      "step": 470
+    },
+    {
+      "epoch": 1.1022067363530779,
+      "grad_norm": 0.9602555632591248,
+      "learning_rate": 2.7695173246471803e-05,
+      "loss": 0.5625,
+      "step": 475
+    },
+    {
+      "epoch": 1.113821138211382,
+      "grad_norm": 1.0547131299972534,
+      "learning_rate": 2.763349121572114e-05,
+      "loss": 0.6173,
+      "step": 480
+    },
+    {
+      "epoch": 1.1254355400696865,
+      "grad_norm": 0.963524341583252,
+      "learning_rate": 2.7571065266536737e-05,
+      "loss": 0.62,
+      "step": 485
+    },
+    {
+      "epoch": 1.1370499419279907,
+      "grad_norm": 1.066751480102539,
+      "learning_rate": 2.7507899074847394e-05,
+      "loss": 0.5829,
+      "step": 490
+    },
+    {
+      "epoch": 1.1486643437862951,
+      "grad_norm": 0.9866227507591248,
+      "learning_rate": 2.7443996360170836e-05,
+      "loss": 0.5729,
+      "step": 495
+    },
+    {
+      "epoch": 1.1602787456445993,
+      "grad_norm": 1.037983775138855,
+      "learning_rate": 2.7379360885394664e-05,
+      "loss": 0.5955,
+      "step": 500
+    },
+    {
+      "epoch": 1.1718931475029035,
+      "grad_norm": 1.2030766010284424,
+      "learning_rate": 2.731399645655477e-05,
+      "loss": 0.6062,
+      "step": 505
+    },
+    {
+      "epoch": 1.183507549361208,
+      "grad_norm": 1.0329967737197876,
+      "learning_rate": 2.7247906922611254e-05,
+      "loss": 0.5871,
+      "step": 510
+    },
+    {
+      "epoch": 1.1951219512195121,
+      "grad_norm": 1.1776123046875,
+      "learning_rate": 2.7181096175221757e-05,
+      "loss": 0.5755,
+      "step": 515
+    },
+    {
+      "epoch": 1.2067363530778166,
+      "grad_norm": 1.0537413358688354,
+      "learning_rate": 2.7113568148512296e-05,
+      "loss": 0.545,
+      "step": 520
+    },
+    {
+      "epoch": 1.2183507549361208,
+      "grad_norm": 0.9902351498603821,
+      "learning_rate": 2.704532681884562e-05,
+      "loss": 0.5564,
+      "step": 525
+    },
+    {
+      "epoch": 1.229965156794425,
+      "grad_norm": 1.0097850561141968,
+      "learning_rate": 2.697637620458706e-05,
+      "loss": 0.5056,
+      "step": 530
+    },
+    {
+      "epoch": 1.2415795586527294,
+      "grad_norm": 0.9843356013298035,
+      "learning_rate": 2.690672036586791e-05,
+      "loss": 0.5348,
+      "step": 535
+    },
+    {
+      "epoch": 1.2531939605110336,
+      "grad_norm": 0.9540740251541138,
+      "learning_rate": 2.6836363404346324e-05,
+      "loss": 0.5411,
+      "step": 540
+    },
+    {
+      "epoch": 1.264808362369338,
+      "grad_norm": 1.0271143913269043,
+      "learning_rate": 2.6765309462965845e-05,
+      "loss": 0.5065,
+      "step": 545
+    },
+    {
+      "epoch": 1.2764227642276422,
+      "grad_norm": 1.077792763710022,
+      "learning_rate": 2.669356272571138e-05,
+      "loss": 0.5307,
+      "step": 550
+    },
+    {
+      "epoch": 1.2880371660859466,
+      "grad_norm": 1.0701345205307007,
+      "learning_rate": 2.6621127417362886e-05,
+      "loss": 0.5587,
+      "step": 555
+    },
+    {
+      "epoch": 1.2996515679442509,
+      "grad_norm": 1.0792369842529297,
+      "learning_rate": 2.6548007803246575e-05,
+      "loss": 0.5264,
+      "step": 560
+    },
+    {
+      "epoch": 1.3112659698025553,
+      "grad_norm": 0.8799574375152588,
+      "learning_rate": 2.647420818898373e-05,
+      "loss": 0.4964,
+      "step": 565
+    },
+    {
+      "epoch": 1.3228803716608595,
+      "grad_norm": 1.1636409759521484,
+      "learning_rate": 2.6399732920237212e-05,
+      "loss": 0.5576,
+      "step": 570
+    },
+    {
+      "epoch": 1.3344947735191637,
+      "grad_norm": 1.055418610572815,
+      "learning_rate": 2.6324586382455525e-05,
+      "loss": 0.5149,
+      "step": 575
+    },
+    {
+      "epoch": 1.346109175377468,
+      "grad_norm": 1.0145419836044312,
+      "learning_rate": 2.624877300061462e-05,
+      "loss": 0.4917,
+      "step": 580
+    },
+    {
+      "epoch": 1.3577235772357723,
+      "grad_norm": 0.9634491801261902,
+      "learning_rate": 2.6172297238957297e-05,
+      "loss": 0.515,
+      "step": 585
+    },
+    {
+      "epoch": 1.3693379790940767,
+      "grad_norm": 0.9772712588310242,
+      "learning_rate": 2.6095163600730355e-05,
+      "loss": 0.5012,
+      "step": 590
+    },
+    {
+      "epoch": 1.380952380952381,
+      "grad_norm": 0.9910968542098999,
+      "learning_rate": 2.6017376627919405e-05,
+      "loss": 0.4839,
+      "step": 595
+    },
+    {
+      "epoch": 1.3925667828106851,
+      "grad_norm": 1.103830099105835,
+      "learning_rate": 2.5938940900981424e-05,
+      "loss": 0.562,
+      "step": 600
+    },
+    {
+      "epoch": 1.4041811846689896,
+      "grad_norm": 1.1236923933029175,
+      "learning_rate": 2.5859861038575035e-05,
+      "loss": 0.493,
+      "step": 605
+    },
+    {
+      "epoch": 1.415795586527294,
+      "grad_norm": 0.9853049516677856,
+      "learning_rate": 2.5780141697288537e-05,
+      "loss": 0.4578,
+      "step": 610
+    },
+    {
+      "epoch": 1.4274099883855982,
+      "grad_norm": 1.02285635471344,
+      "learning_rate": 2.5699787571365704e-05,
+      "loss": 0.5141,
+      "step": 615
+    },
+    {
+      "epoch": 1.4390243902439024,
+      "grad_norm": 1.1389844417572021,
+      "learning_rate": 2.5618803392429373e-05,
+      "loss": 0.4684,
+      "step": 620
+    },
+    {
+      "epoch": 1.4506387921022068,
+      "grad_norm": 1.1375923156738281,
+      "learning_rate": 2.5537193929202815e-05,
+      "loss": 0.4761,
+      "step": 625
+    },
+    {
+      "epoch": 1.462253193960511,
+      "grad_norm": 1.046151876449585,
+      "learning_rate": 2.5454963987228926e-05,
+      "loss": 0.5179,
+      "step": 630
+    },
+    {
+      "epoch": 1.4738675958188154,
+      "grad_norm": 1.1057745218276978,
+      "learning_rate": 2.5372118408587284e-05,
+      "loss": 0.489,
+      "step": 635
+    },
+    {
+      "epoch": 1.4854819976771196,
+      "grad_norm": 1.032041311264038,
+      "learning_rate": 2.5288662071608975e-05,
+      "loss": 0.4871,
+      "step": 640
+    },
+    {
+      "epoch": 1.4970963995354238,
+      "grad_norm": 1.0814687013626099,
+      "learning_rate": 2.520459989058939e-05,
+      "loss": 0.5172,
+      "step": 645
+    },
+    {
+      "epoch": 1.5087108013937283,
+      "grad_norm": 1.0195306539535522,
+      "learning_rate": 2.5119936815498797e-05,
+      "loss": 0.5218,
+      "step": 650
+    },
+    {
+      "epoch": 1.5203252032520327,
+      "grad_norm": 0.9469033479690552,
+      "learning_rate": 2.503467783169091e-05,
+      "loss": 0.4883,
+      "step": 655
+    },
+    {
+      "epoch": 1.5319396051103369,
+      "grad_norm": 1.1030808687210083,
+      "learning_rate": 2.4948827959609285e-05,
+      "loss": 0.501,
+      "step": 660
+    },
+    {
+      "epoch": 1.543554006968641,
+      "grad_norm": 1.0832819938659668,
+      "learning_rate": 2.4862392254491736e-05,
+      "loss": 0.4813,
+      "step": 665
+    },
+    {
+      "epoch": 1.5551684088269453,
+      "grad_norm": 1.1702861785888672,
+      "learning_rate": 2.477537580607261e-05,
+      "loss": 0.4737,
+      "step": 670
+    },
+    {
+      "epoch": 1.5667828106852497,
+      "grad_norm": 1.0441266298294067,
+      "learning_rate": 2.4687783738283144e-05,
+      "loss": 0.454,
+      "step": 675
+    },
+    {
+      "epoch": 1.5783972125435541,
+      "grad_norm": 1.1927803754806519,
+      "learning_rate": 2.4599621208949674e-05,
+      "loss": 0.4328,
+      "step": 680
+    },
+    {
+      "epoch": 1.5900116144018583,
+      "grad_norm": 1.1129752397537231,
+      "learning_rate": 2.4510893409489967e-05,
+      "loss": 0.4024,
+      "step": 685
+    },
+    {
+      "epoch": 1.6016260162601625,
+      "grad_norm": 1.0736514329910278,
+      "learning_rate": 2.4421605564607514e-05,
+      "loss": 0.4499,
+      "step": 690
+    },
+    {
+      "epoch": 1.6132404181184667,
+      "grad_norm": 1.0657414197921753,
+      "learning_rate": 2.4331762931983866e-05,
+      "loss": 0.4649,
+      "step": 695
+    },
+    {
+      "epoch": 1.6248548199767712,
+      "grad_norm": 1.0618281364440918,
+      "learning_rate": 2.4241370801969045e-05,
+      "loss": 0.4337,
+      "step": 700
+    },
+    {
+      "epoch": 1.6364692218350756,
+      "grad_norm": 1.1562613248825073,
+      "learning_rate": 2.415043449727003e-05,
+      "loss": 0.4429,
+      "step": 705
+    },
+    {
+      "epoch": 1.6480836236933798,
+      "grad_norm": 1.1286054849624634,
+      "learning_rate": 2.4058959372637304e-05,
+      "loss": 0.4714,
+      "step": 710
+    },
+    {
+      "epoch": 1.659698025551684,
+      "grad_norm": 1.043707251548767,
+      "learning_rate": 2.396695081454959e-05,
+      "loss": 0.4549,
+      "step": 715
+    },
+    {
+      "epoch": 1.6713124274099884,
+      "grad_norm": 1.001025676727295,
+      "learning_rate": 2.387441424089662e-05,
+      "loss": 0.4843,
+      "step": 720
+    },
+    {
+      "epoch": 1.6829268292682928,
+      "grad_norm": 1.0033694505691528,
+      "learning_rate": 2.378135510066013e-05,
+      "loss": 0.4203,
+      "step": 725
+    },
+    {
+      "epoch": 1.694541231126597,
+      "grad_norm": 1.041979432106018,
+      "learning_rate": 2.3687778873593e-05,
+      "loss": 0.4609,
+      "step": 730
+    },
+    {
+      "epoch": 1.7061556329849012,
+      "grad_norm": 1.0614110231399536,
+      "learning_rate": 2.3593691069896582e-05,
+      "loss": 0.4328,
+      "step": 735
+    },
+    {
+      "epoch": 1.7177700348432055,
+      "grad_norm": 1.1184892654418945,
+      "learning_rate": 2.3499097229896213e-05,
+      "loss": 0.4323,
+      "step": 740
+    },
+    {
+      "epoch": 1.7293844367015099,
+      "grad_norm": 1.0571225881576538,
+      "learning_rate": 2.340400292371499e-05,
+      "loss": 0.3688,
+      "step": 745
+    },
+    {
+      "epoch": 1.7409988385598143,
+      "grad_norm": 0.9831321239471436,
+      "learning_rate": 2.3308413750945788e-05,
+      "loss": 0.4091,
+      "step": 750
+    },
+    {
+      "epoch": 1.7526132404181185,
+      "grad_norm": 1.334681749343872,
+      "learning_rate": 2.3212335340321518e-05,
+      "loss": 0.4363,
+      "step": 755
+    },
+    {
+      "epoch": 1.7642276422764227,
+      "grad_norm": 1.0677987337112427,
+      "learning_rate": 2.3115773349383658e-05,
+      "loss": 0.4202,
+      "step": 760
+    },
+    {
+      "epoch": 1.775842044134727,
+      "grad_norm": 1.1645525693893433,
+      "learning_rate": 2.3018733464149156e-05,
+      "loss": 0.4515,
+      "step": 765
+    },
+    {
+      "epoch": 1.7874564459930313,
+      "grad_norm": 1.0856503248214722,
+      "learning_rate": 2.292122139877558e-05,
+      "loss": 0.401,
+      "step": 770
+    },
+    {
+      "epoch": 1.7990708478513358,
+      "grad_norm": 1.0238492488861084,
+      "learning_rate": 2.2823242895224643e-05,
+      "loss": 0.3955,
+      "step": 775
+    },
+    {
+      "epoch": 1.81068524970964,
+      "grad_norm": 1.1932063102722168,
+      "learning_rate": 2.2724803722924106e-05,
+      "loss": 0.4154,
+      "step": 780
+    },
+    {
+      "epoch": 1.8222996515679442,
+      "grad_norm": 1.0874030590057373,
+      "learning_rate": 2.2625909678428038e-05,
+      "loss": 0.4082,
+      "step": 785
+    },
+    {
+      "epoch": 1.8339140534262486,
+      "grad_norm": 1.109732747077942,
+      "learning_rate": 2.2526566585075485e-05,
+      "loss": 0.3826,
+      "step": 790
+    },
+    {
+      "epoch": 1.845528455284553,
+      "grad_norm": 1.0591766834259033,
+      "learning_rate": 2.2426780292647568e-05,
+      "loss": 0.388,
+      "step": 795
+    },
+    {
+      "epoch": 1.8571428571428572,
+      "grad_norm": 1.1858235597610474,
+      "learning_rate": 2.2326556677023017e-05,
+      "loss": 0.4026,
+      "step": 800
+    },
+    {
+      "epoch": 1.8687572590011614,
+      "grad_norm": 1.0293341875076294,
+      "learning_rate": 2.2225901639832188e-05,
+      "loss": 0.419,
+      "step": 805
+    },
+    {
+      "epoch": 1.8803716608594656,
+      "grad_norm": 1.1112502813339233,
+      "learning_rate": 2.2124821108109515e-05,
+      "loss": 0.4005,
+      "step": 810
+    },
+    {
+      "epoch": 1.89198606271777,
+      "grad_norm": 0.9960114359855652,
+      "learning_rate": 2.2023321033944544e-05,
+      "loss": 0.328,
+      "step": 815
+    },
+    {
+      "epoch": 1.9036004645760745,
+      "grad_norm": 1.1106208562850952,
+      "learning_rate": 2.1921407394131406e-05,
+      "loss": 0.3667,
+      "step": 820
+    },
+    {
+      "epoch": 1.9152148664343787,
+      "grad_norm": 1.0423858165740967,
+      "learning_rate": 2.1819086189816893e-05,
+      "loss": 0.3773,
+      "step": 825
+    },
+    {
+      "epoch": 1.9268292682926829,
+      "grad_norm": 1.13150155544281,
+      "learning_rate": 2.171636344614708e-05,
+      "loss": 0.3491,
+      "step": 830
+    },
+    {
+      "epoch": 1.938443670150987,
+      "grad_norm": 1.0611273050308228,
+      "learning_rate": 2.1613245211912554e-05,
+      "loss": 0.3323,
+      "step": 835
+    },
+    {
+      "epoch": 1.9500580720092915,
+      "grad_norm": 0.9943997263908386,
+      "learning_rate": 2.1509737559192188e-05,
+      "loss": 0.3741,
+      "step": 840
+    },
+    {
+      "epoch": 1.961672473867596,
+      "grad_norm": 0.9484174847602844,
+      "learning_rate": 2.140584658299564e-05,
+      "loss": 0.3415,
+      "step": 845
+    },
+    {
+      "epoch": 1.9732868757259001,
+      "grad_norm": 1.0280919075012207,
+      "learning_rate": 2.1301578400904424e-05,
+      "loss": 0.3724,
+      "step": 850
+    },
+    {
+      "epoch": 1.9849012775842043,
+      "grad_norm": 1.1084401607513428,
+      "learning_rate": 2.119693915271168e-05,
+      "loss": 0.3867,
+      "step": 855
+    },
+    {
+      "epoch": 1.9965156794425087,
+      "grad_norm": 1.1195417642593384,
+      "learning_rate": 2.1091935000060637e-05,
+      "loss": 0.3641,
+      "step": 860
+    },
+    {
+      "epoch": 2.0069686411149825,
+      "grad_norm": 0.877273440361023,
+      "learning_rate": 2.0986572126081793e-05,
+      "loss": 0.3401,
+      "step": 865
+    },
+    {
+      "epoch": 2.0185830429732867,
+      "grad_norm": 1.3777830600738525,
+      "learning_rate": 2.0880856735028803e-05,
+      "loss": 0.3079,
+      "step": 870
+    },
+    {
+      "epoch": 2.0301974448315914,
+      "grad_norm": 1.1466902494430542,
+      "learning_rate": 2.0774795051913178e-05,
+      "loss": 0.2913,
+      "step": 875
+    },
+    {
+      "epoch": 2.0418118466898956,
+      "grad_norm": 1.0693480968475342,
+      "learning_rate": 2.0668393322137703e-05,
+      "loss": 0.2945,
+      "step": 880
+    },
+    {
+      "epoch": 2.0534262485481998,
+      "grad_norm": 1.0168899297714233,
+      "learning_rate": 2.056165781112867e-05,
+      "loss": 0.2828,
+      "step": 885
+    },
+    {
+      "epoch": 2.065040650406504,
+      "grad_norm": 1.0432965755462646,
+      "learning_rate": 2.0454594803966972e-05,
+      "loss": 0.2915,
+      "step": 890
+    },
+    {
+      "epoch": 2.076655052264808,
+      "grad_norm": 1.100029468536377,
+      "learning_rate": 2.034721060501799e-05,
+      "loss": 0.297,
+      "step": 895
+    },
+    {
+      "epoch": 2.088269454123113,
+      "grad_norm": 1.1748476028442383,
+      "learning_rate": 2.023951153756035e-05,
+      "loss": 0.2892,
+      "step": 900
+    },
+    {
+      "epoch": 2.099883855981417,
+      "grad_norm": 1.0221283435821533,
+      "learning_rate": 2.013150394341361e-05,
+      "loss": 0.2831,
+      "step": 905
+    },
+    {
+      "epoch": 2.1114982578397212,
+      "grad_norm": 0.9806298017501831,
+      "learning_rate": 2.002319418256479e-05,
+      "loss": 0.2785,
+      "step": 910
+    },
+    {
+      "epoch": 2.1231126596980254,
+      "grad_norm": 1.292575716972351,
+      "learning_rate": 1.9914588632793908e-05,
+      "loss": 0.2829,
+      "step": 915
+    },
+    {
+      "epoch": 2.1347270615563296,
+      "grad_norm": 1.0625466108322144,
+      "learning_rate": 1.9805693689298374e-05,
+      "loss": 0.2867,
+      "step": 920
+    },
+    {
+      "epoch": 2.1463414634146343,
+      "grad_norm": 1.0215439796447754,
+      "learning_rate": 1.969651576431646e-05,
+      "loss": 0.3223,
+      "step": 925
+    },
+    {
+      "epoch": 2.1579558652729385,
+      "grad_norm": 1.1626147031784058,
+      "learning_rate": 1.9587061286749693e-05,
+      "loss": 0.3042,
+      "step": 930
+    },
+    {
+      "epoch": 2.1695702671312427,
+      "grad_norm": 0.9875515699386597,
+      "learning_rate": 1.9477336701784292e-05,
+      "loss": 0.2827,
+      "step": 935
+    },
+    {
+      "epoch": 2.181184668989547,
+      "grad_norm": 1.0031932592391968,
+      "learning_rate": 1.9367348470511652e-05,
+      "loss": 0.2712,
+      "step": 940
+    },
+    {
+      "epoch": 2.1927990708478515,
+      "grad_norm": 1.1925930976867676,
+      "learning_rate": 1.9257103069547894e-05,
+      "loss": 0.2895,
+      "step": 945
+    },
+    {
+      "epoch": 2.2044134727061557,
+      "grad_norm": 1.217810869216919,
+      "learning_rate": 1.9146606990652452e-05,
+      "loss": 0.2714,
+      "step": 950
+    },
+    {
+      "epoch": 2.21602787456446,
+      "grad_norm": 0.9798834323883057,
+      "learning_rate": 1.903586674034587e-05,
+      "loss": 0.2665,
+      "step": 955
+    },
+    {
+      "epoch": 2.227642276422764,
+      "grad_norm": 1.2851895093917847,
+      "learning_rate": 1.8924888839526612e-05,
+      "loss": 0.2795,
+      "step": 960
+    },
+    {
+      "epoch": 2.2392566782810683,
+      "grad_norm": 1.1440101861953735,
+      "learning_rate": 1.881367982308711e-05,
+      "loss": 0.2779,
+      "step": 965
+    },
+    {
+      "epoch": 2.250871080139373,
+      "grad_norm": 1.3127503395080566,
+      "learning_rate": 1.8702246239528965e-05,
+      "loss": 0.2638,
+      "step": 970
+    },
+    {
+      "epoch": 2.262485481997677,
+      "grad_norm": 1.0733845233917236,
+      "learning_rate": 1.8590594650577314e-05,
+      "loss": 0.2646,
+      "step": 975
+    },
+    {
+      "epoch": 2.2740998838559814,
+      "grad_norm": 1.0694085359573364,
+      "learning_rate": 1.847873163079446e-05,
+      "loss": 0.2741,
+      "step": 980
+    },
+    {
+      "epoch": 2.2857142857142856,
+      "grad_norm": 1.1703122854232788,
+      "learning_rate": 1.8366663767192747e-05,
+      "loss": 0.2374,
+      "step": 985
+    },
+    {
+      "epoch": 2.2973286875725902,
+      "grad_norm": 1.2135448455810547,
+      "learning_rate": 1.8254397658846656e-05,
+      "loss": 0.271,
+      "step": 990
+    },
+    {
+      "epoch": 2.3089430894308944,
+      "grad_norm": 1.1723593473434448,
+      "learning_rate": 1.8141939916504245e-05,
+      "loss": 0.2744,
+      "step": 995
+    },
+    {
+      "epoch": 2.3205574912891986,
+      "grad_norm": 1.2068791389465332,
+      "learning_rate": 1.8029297162197867e-05,
+      "loss": 0.2597,
+      "step": 1000
+    },
+    {
+      "epoch": 2.332171893147503,
+      "grad_norm": 1.1227343082427979,
+      "learning_rate": 1.791647602885423e-05,
+      "loss": 0.254,
+      "step": 1005
+    },
+    {
+      "epoch": 2.343786295005807,
+      "grad_norm": 1.1190464496612549,
+      "learning_rate": 1.7803483159903835e-05,
+      "loss": 0.232,
+      "step": 1010
+    },
+    {
+      "epoch": 2.3554006968641117,
+      "grad_norm": 1.0415700674057007,
+      "learning_rate": 1.769032520888976e-05,
+      "loss": 0.2613,
+      "step": 1015
+    },
+    {
+      "epoch": 2.367015098722416,
+      "grad_norm": 1.188915491104126,
+      "learning_rate": 1.7577008839075886e-05,
+      "loss": 0.2392,
+      "step": 1020
+    },
+    {
+      "epoch": 2.37862950058072,
+      "grad_norm": 1.1599727869033813,
+      "learning_rate": 1.7463540723054526e-05,
+      "loss": 0.2607,
+      "step": 1025
+    },
+    {
+      "epoch": 2.3902439024390243,
+      "grad_norm": 1.0613243579864502,
+      "learning_rate": 1.734992754235351e-05,
+      "loss": 0.2517,
+      "step": 1030
+    },
+    {
+      "epoch": 2.4018583042973285,
+      "grad_norm": 1.140293836593628,
+      "learning_rate": 1.7236175987042744e-05,
+      "loss": 0.2757,
+      "step": 1035
+    },
+    {
+      "epoch": 2.413472706155633,
+      "grad_norm": 1.217897891998291,
+      "learning_rate": 1.7122292755340284e-05,
+      "loss": 0.2604,
+      "step": 1040
+    },
+    {
+      "epoch": 2.4250871080139373,
+      "grad_norm": 1.2544645071029663,
+      "learning_rate": 1.700828455321788e-05,
+      "loss": 0.2539,
+      "step": 1045
+    },
+    {
+      "epoch": 2.4367015098722415,
+      "grad_norm": 1.5165183544158936,
+      "learning_rate": 1.6894158094006142e-05,
+      "loss": 0.2392,
+      "step": 1050
+    },
+    {
+      "epoch": 2.4483159117305457,
+      "grad_norm": 1.0610485076904297,
+      "learning_rate": 1.6779920097999194e-05,
+      "loss": 0.2678,
+      "step": 1055
+    },
+    {
+      "epoch": 2.45993031358885,
+      "grad_norm": 1.1240959167480469,
+      "learning_rate": 1.6665577292058955e-05,
+      "loss": 0.2425,
+      "step": 1060
+    },
+    {
+      "epoch": 2.4715447154471546,
+      "grad_norm": 1.2169679403305054,
+      "learning_rate": 1.655113640921905e-05,
+      "loss": 0.2425,
+      "step": 1065
+    },
+    {
+      "epoch": 2.483159117305459,
+      "grad_norm": 1.0992380380630493,
+      "learning_rate": 1.6436604188288325e-05,
+      "loss": 0.2537,
+      "step": 1070
+    },
+    {
+      "epoch": 2.494773519163763,
+      "grad_norm": 1.1832488775253296,
+      "learning_rate": 1.632198737345403e-05,
+      "loss": 0.2596,
+      "step": 1075
+    },
+    {
+      "epoch": 2.506387921022067,
+      "grad_norm": 0.9653327465057373,
+      "learning_rate": 1.6207292713884692e-05,
+      "loss": 0.2427,
+      "step": 1080
+    },
+    {
+      "epoch": 2.5180023228803714,
+      "grad_norm": 1.139378547668457,
+      "learning_rate": 1.609252696333271e-05,
+      "loss": 0.2432,
+      "step": 1085
+    },
+    {
+      "epoch": 2.529616724738676,
+      "grad_norm": 1.0014028549194336,
+      "learning_rate": 1.597769687973664e-05,
+      "loss": 0.2329,
+      "step": 1090
+    },
+    {
+      "epoch": 2.5412311265969802,
+      "grad_norm": 1.1616328954696655,
+      "learning_rate": 1.586280922482327e-05,
+      "loss": 0.2513,
+      "step": 1095
+    },
+    {
+      "epoch": 2.5528455284552845,
+      "grad_norm": 1.022898554801941,
+      "learning_rate": 1.5747870763709432e-05,
+      "loss": 0.2697,
+      "step": 1100
+    },
+    {
+      "epoch": 2.564459930313589,
+      "grad_norm": 1.102315902709961,
+      "learning_rate": 1.5632888264503705e-05,
+      "loss": 0.2368,
+      "step": 1105
+    },
+    {
+      "epoch": 2.5760743321718933,
+      "grad_norm": 1.0889406204223633,
+      "learning_rate": 1.5517868497907795e-05,
+      "loss": 0.2377,
+      "step": 1110
+    },
+    {
+      "epoch": 2.5876887340301975,
+      "grad_norm": 1.0581204891204834,
+      "learning_rate": 1.54028182368179e-05,
+      "loss": 0.2424,
+      "step": 1115
+    },
+    {
+      "epoch": 2.5993031358885017,
+      "grad_norm": 1.180722713470459,
+      "learning_rate": 1.5287744255925882e-05,
+      "loss": 0.2191,
+      "step": 1120
+    },
+    {
+      "epoch": 2.610917537746806,
+      "grad_norm": 1.0064940452575684,
+      "learning_rate": 1.5172653331320318e-05,
+      "loss": 0.2718,
+      "step": 1125
+    },
+    {
+      "epoch": 2.6225319396051106,
+      "grad_norm": 1.1252609491348267,
+      "learning_rate": 1.5057552240087526e-05,
+      "loss": 0.2248,
+      "step": 1130
+    },
+    {
+      "epoch": 2.6341463414634148,
+      "grad_norm": 1.0868943929672241,
+      "learning_rate": 1.4942447759912473e-05,
+      "loss": 0.2609,
+      "step": 1135
+    },
+    {
+      "epoch": 2.645760743321719,
+      "grad_norm": 1.127143144607544,
+      "learning_rate": 1.4827346668679683e-05,
+      "loss": 0.233,
+      "step": 1140
+    },
+    {
+      "epoch": 2.657375145180023,
+      "grad_norm": 1.2413113117218018,
+      "learning_rate": 1.471225574407412e-05,
+      "loss": 0.2043,
+      "step": 1145
+    },
+    {
+      "epoch": 2.6689895470383274,
+      "grad_norm": 1.1059684753417969,
+      "learning_rate": 1.4597181763182104e-05,
+      "loss": 0.2375,
+      "step": 1150
+    },
+    {
+      "epoch": 2.680603948896632,
+      "grad_norm": 1.1019221544265747,
+      "learning_rate": 1.4482131502092208e-05,
+      "loss": 0.1922,
+      "step": 1155
+    },
+    {
+      "epoch": 2.692218350754936,
+      "grad_norm": 1.1478207111358643,
+      "learning_rate": 1.4367111735496298e-05,
+      "loss": 0.2286,
+      "step": 1160
+    },
+    {
+      "epoch": 2.7038327526132404,
+      "grad_norm": 1.0406802892684937,
+      "learning_rate": 1.4252129236290569e-05,
+      "loss": 0.2078,
+      "step": 1165
+    },
+    {
+      "epoch": 2.7154471544715446,
+      "grad_norm": 1.2635127305984497,
+      "learning_rate": 1.4137190775176739e-05,
+      "loss": 0.2146,
+      "step": 1170
+    },
+    {
+      "epoch": 2.727061556329849,
+      "grad_norm": 1.0526198148727417,
+      "learning_rate": 1.402230312026336e-05,
+      "loss": 0.2103,
+      "step": 1175
+    },
+    {
+      "epoch": 2.7386759581881535,
+      "grad_norm": 1.297333836555481,
+      "learning_rate": 1.390747303666729e-05,
+      "loss": 0.2183,
+      "step": 1180
+    },
+    {
+      "epoch": 2.7502903600464577,
+      "grad_norm": 1.0738940238952637,
+      "learning_rate": 1.3792707286115309e-05,
+      "loss": 0.2077,
+      "step": 1185
+    },
+    {
+      "epoch": 2.761904761904762,
+      "grad_norm": 1.0152521133422852,
+      "learning_rate": 1.3678012626545978e-05,
+      "loss": 0.2368,
+      "step": 1190
+    },
+    {
+      "epoch": 2.773519163763066,
+      "grad_norm": 1.0068998336791992,
+      "learning_rate": 1.3563395811711678e-05,
+      "loss": 0.2152,
+      "step": 1195
+    },
+    {
+      "epoch": 2.7851335656213703,
+      "grad_norm": 1.1199923753738403,
+      "learning_rate": 1.344886359078095e-05,
+      "loss": 0.2188,
+      "step": 1200
+    },
+    {
+      "epoch": 2.796747967479675,
+      "grad_norm": 1.037109613418579,
+      "learning_rate": 1.3334422707941051e-05,
+      "loss": 0.2286,
+      "step": 1205
+    },
+    {
+      "epoch": 2.808362369337979,
+      "grad_norm": 0.9589251279830933,
+      "learning_rate": 1.3220079902000811e-05,
+      "loss": 0.2025,
+      "step": 1210
+    },
+    {
+      "epoch": 2.8199767711962833,
+      "grad_norm": 1.2464582920074463,
+      "learning_rate": 1.3105841905993857e-05,
+      "loss": 0.2131,
+      "step": 1215
+    },
+    {
+      "epoch": 2.831591173054588,
+      "grad_norm": 1.091057538986206,
+      "learning_rate": 1.2991715446782121e-05,
+      "loss": 0.2067,
+      "step": 1220
+    },
+    {
+      "epoch": 2.8432055749128917,
+      "grad_norm": 1.278668761253357,
+      "learning_rate": 1.2877707244659722e-05,
+      "loss": 0.2209,
+      "step": 1225
+    },
+    {
+      "epoch": 2.8548199767711964,
+      "grad_norm": 1.2127578258514404,
+      "learning_rate": 1.2763824012957258e-05,
+      "loss": 0.1833,
+      "step": 1230
+    },
+    {
+      "epoch": 2.8664343786295006,
+      "grad_norm": 1.146657109260559,
+      "learning_rate": 1.2650072457646492e-05,
+      "loss": 0.1814,
+      "step": 1235
+    },
+    {
+      "epoch": 2.8780487804878048,
+      "grad_norm": 1.0890151262283325,
+      "learning_rate": 1.2536459276945473e-05,
+      "loss": 0.1801,
+      "step": 1240
+    },
+    {
+      "epoch": 2.8896631823461094,
+      "grad_norm": 0.9708114266395569,
+      "learning_rate": 1.2422991160924117e-05,
+      "loss": 0.1946,
+      "step": 1245
+    },
+    {
+      "epoch": 2.9012775842044136,
+      "grad_norm": 1.1722114086151123,
+      "learning_rate": 1.2309674791110244e-05,
+      "loss": 0.2438,
+      "step": 1250
+    },
+    {
+      "epoch": 2.912891986062718,
+      "grad_norm": 1.015057921409607,
+      "learning_rate": 1.2196516840096167e-05,
+      "loss": 0.2035,
+      "step": 1255
+    },
+    {
+      "epoch": 2.924506387921022,
+      "grad_norm": 0.9809845685958862,
+      "learning_rate": 1.2083523971145773e-05,
+      "loss": 0.189,
+      "step": 1260
+    },
+    {
+      "epoch": 2.9361207897793262,
+      "grad_norm": 1.2840055227279663,
+      "learning_rate": 1.1970702837802137e-05,
+      "loss": 0.2055,
+      "step": 1265
+    },
+    {
+      "epoch": 2.947735191637631,
+      "grad_norm": 0.9905469417572021,
+      "learning_rate": 1.1858060083495761e-05,
+      "loss": 0.1822,
+      "step": 1270
+    },
+    {
+      "epoch": 2.959349593495935,
+      "grad_norm": 1.1354817152023315,
+      "learning_rate": 1.1745602341153347e-05,
+      "loss": 0.1883,
+      "step": 1275
+    },
+    {
+      "epoch": 2.9709639953542393,
+      "grad_norm": 1.3380357027053833,
+      "learning_rate": 1.1633336232807256e-05,
+      "loss": 0.1874,
+      "step": 1280
+    },
+    {
+      "epoch": 2.9825783972125435,
+      "grad_norm": 1.232898235321045,
+      "learning_rate": 1.1521268369205546e-05,
+      "loss": 0.177,
+      "step": 1285
+    },
+    {
+      "epoch": 2.9941927990708477,
+      "grad_norm": 1.0560810565948486,
+      "learning_rate": 1.1409405349422693e-05,
+      "loss": 0.1929,
+      "step": 1290
+    },
+    {
+      "epoch": 3.0046457607433217,
+      "grad_norm": 0.9177282452583313,
+      "learning_rate": 1.1297753760471034e-05,
+      "loss": 0.199,
+      "step": 1295
+    },
+    {
+      "epoch": 3.016260162601626,
+      "grad_norm": 1.1153042316436768,
+      "learning_rate": 1.118632017691289e-05,
+      "loss": 0.1459,
+      "step": 1300
+    },
+    {
+      "epoch": 3.0278745644599305,
+      "grad_norm": 0.9868292212486267,
+      "learning_rate": 1.1075111160473392e-05,
+      "loss": 0.149,
+      "step": 1305
+    },
+    {
+      "epoch": 3.0394889663182347,
+      "grad_norm": 0.9717047214508057,
+      "learning_rate": 1.0964133259654136e-05,
+      "loss": 0.154,
+      "step": 1310
+    },
+    {
+      "epoch": 3.051103368176539,
+      "grad_norm": 1.0850447416305542,
+      "learning_rate": 1.085339300934755e-05,
+      "loss": 0.164,
+      "step": 1315
+    },
+    {
+      "epoch": 3.062717770034843,
+      "grad_norm": 1.0128169059753418,
+      "learning_rate": 1.074289693045211e-05,
+      "loss": 0.1674,
+      "step": 1320
+    },
+    {
+      "epoch": 3.0743321718931473,
+      "grad_norm": 1.1913701295852661,
+      "learning_rate": 1.0632651529488349e-05,
+      "loss": 0.1756,
+      "step": 1325
+    },
+    {
+      "epoch": 3.085946573751452,
+      "grad_norm": 1.0130043029785156,
+      "learning_rate": 1.052266329821571e-05,
+      "loss": 0.1433,
+      "step": 1330
+    },
+    {
+      "epoch": 3.097560975609756,
+      "grad_norm": 1.0070334672927856,
+      "learning_rate": 1.0412938713250308e-05,
+      "loss": 0.1627,
+      "step": 1335
+    },
+    {
+      "epoch": 3.1091753774680604,
+      "grad_norm": 1.0455143451690674,
+      "learning_rate": 1.0303484235683543e-05,
+      "loss": 0.1506,
+      "step": 1340
+    },
+    {
+      "epoch": 3.1207897793263646,
+      "grad_norm": 1.0055582523345947,
+      "learning_rate": 1.0194306310701629e-05,
+      "loss": 0.1485,
+      "step": 1345
+    },
+    {
+      "epoch": 3.132404181184669,
+      "grad_norm": 1.1295661926269531,
+      "learning_rate": 1.0085411367206098e-05,
+      "loss": 0.1504,
+      "step": 1350
+    },
+    {
+      "epoch": 3.1440185830429734,
+      "grad_norm": 0.9493905305862427,
+      "learning_rate": 9.976805817435208e-06,
+      "loss": 0.1582,
+      "step": 1355
+    },
+    {
+      "epoch": 3.1556329849012776,
+      "grad_norm": 0.9460111856460571,
+      "learning_rate": 9.868496056586391e-06,
+      "loss": 0.1597,
+      "step": 1360
+    },
+    {
+      "epoch": 3.167247386759582,
+      "grad_norm": 0.9388046860694885,
+      "learning_rate": 9.760488462439653e-06,
+      "loss": 0.1433,
+      "step": 1365
+    },
+    {
+      "epoch": 3.178861788617886,
+      "grad_norm": 1.1539472341537476,
+      "learning_rate": 9.652789394982014e-06,
+      "loss": 0.142,
+      "step": 1370
+    },
+    {
+      "epoch": 3.1904761904761907,
+      "grad_norm": 1.1274142265319824,
+      "learning_rate": 9.545405196033025e-06,
+      "loss": 0.1488,
+      "step": 1375
+    },
+    {
+      "epoch": 3.202090592334495,
+      "grad_norm": 1.1115517616271973,
+      "learning_rate": 9.438342188871331e-06,
+      "loss": 0.1485,
+      "step": 1380
+    },
+    {
+      "epoch": 3.213704994192799,
+      "grad_norm": 0.985272228717804,
+      "learning_rate": 9.331606677862303e-06,
+      "loss": 0.13,
+      "step": 1385
+    },
+    {
+      "epoch": 3.2253193960511033,
+      "grad_norm": 0.8632730841636658,
+      "learning_rate": 9.225204948086824e-06,
+      "loss": 0.1538,
+      "step": 1390
+    },
+    {
+      "epoch": 3.2369337979094075,
+      "grad_norm": 0.9072132706642151,
+      "learning_rate": 9.119143264971198e-06,
+      "loss": 0.1362,
+      "step": 1395
+    },
+    {
+      "epoch": 3.248548199767712,
+      "grad_norm": 1.1985305547714233,
+      "learning_rate": 9.01342787391821e-06,
+      "loss": 0.1421,
+      "step": 1400
+    },
+    {
+      "epoch": 3.2601626016260163,
+      "grad_norm": 1.0423188209533691,
+      "learning_rate": 8.908064999939366e-06,
+      "loss": 0.1367,
+      "step": 1405
+    },
+    {
+      "epoch": 3.2717770034843205,
+      "grad_norm": 0.9370428919792175,
+      "learning_rate": 8.803060847288322e-06,
+      "loss": 0.1609,
+      "step": 1410
+    },
+    {
+      "epoch": 3.2833914053426247,
+      "grad_norm": 1.1630222797393799,
+      "learning_rate": 8.698421599095577e-06,
+      "loss": 0.1527,
+      "step": 1415
+    },
+    {
+      "epoch": 3.295005807200929,
+      "grad_norm": 1.0328259468078613,
+      "learning_rate": 8.594153417004359e-06,
+      "loss": 0.1376,
+      "step": 1420
+    },
+    {
+      "epoch": 3.3066202090592336,
+      "grad_norm": 1.0899609327316284,
+      "learning_rate": 8.490262440807816e-06,
+      "loss": 0.1298,
+      "step": 1425
+    },
+    {
+      "epoch": 3.318234610917538,
+      "grad_norm": 1.0752588510513306,
+      "learning_rate": 8.386754788087453e-06,
+      "loss": 0.1493,
+      "step": 1430
+    },
+    {
+      "epoch": 3.329849012775842,
+      "grad_norm": 1.199508786201477,
+      "learning_rate": 8.28363655385292e-06,
+      "loss": 0.1451,
+      "step": 1435
+    },
+    {
+      "epoch": 3.341463414634146,
+      "grad_norm": 1.1442402601242065,
+      "learning_rate": 8.180913810183111e-06,
+      "loss": 0.1361,
+      "step": 1440
+    },
+    {
+      "epoch": 3.3530778164924504,
+      "grad_norm": 0.8858070969581604,
+      "learning_rate": 8.078592605868597e-06,
+      "loss": 0.1267,
+      "step": 1445
+    },
+    {
+      "epoch": 3.364692218350755,
+      "grad_norm": 1.0057350397109985,
+      "learning_rate": 7.97667896605546e-06,
+      "loss": 0.1494,
+      "step": 1450
+    },
+    {
+      "epoch": 3.3763066202090593,
+      "grad_norm": 0.9687568545341492,
+      "learning_rate": 7.875178891890483e-06,
+      "loss": 0.1322,
+      "step": 1455
+    },
+    {
+      "epoch": 3.3879210220673635,
+      "grad_norm": 1.0430768728256226,
+      "learning_rate": 7.774098360167817e-06,
+      "loss": 0.1287,
+      "step": 1460
+    },
+    {
+      "epoch": 3.3995354239256677,
+      "grad_norm": 0.8199989199638367,
+      "learning_rate": 7.673443322976986e-06,
+      "loss": 0.1449,
+      "step": 1465
+    },
+    {
+      "epoch": 3.4111498257839723,
+      "grad_norm": 1.0182511806488037,
+      "learning_rate": 7.5732197073524404e-06,
+      "loss": 0.1623,
+      "step": 1470
+    },
+    {
+      "epoch": 3.4227642276422765,
+      "grad_norm": 0.9345471858978271,
+      "learning_rate": 7.473433414924512e-06,
+      "loss": 0.1396,
+      "step": 1475
+    },
+    {
+      "epoch": 3.4343786295005807,
+      "grad_norm": 1.0214126110076904,
+      "learning_rate": 7.374090321571961e-06,
+      "loss": 0.133,
+      "step": 1480
+    },
+    {
+      "epoch": 3.445993031358885,
+      "grad_norm": 1.0448100566864014,
+      "learning_rate": 7.275196277075896e-06,
+      "loss": 0.1594,
+      "step": 1485
+    },
+    {
+      "epoch": 3.4576074332171896,
+      "grad_norm": 0.9637423753738403,
+      "learning_rate": 7.176757104775364e-06,
+      "loss": 0.1389,
+      "step": 1490
+    },
+    {
+      "epoch": 3.4692218350754938,
+      "grad_norm": 1.011009693145752,
+      "learning_rate": 7.078778601224424e-06,
+      "loss": 0.1328,
+      "step": 1495
+    },
+    {
+      "epoch": 3.480836236933798,
+      "grad_norm": 0.9226102828979492,
+      "learning_rate": 6.981266535850848e-06,
+      "loss": 0.135,
+      "step": 1500
+    },
+    {
+      "epoch": 3.492450638792102,
+      "grad_norm": 0.9921573400497437,
+      "learning_rate": 6.8842266506163435e-06,
+      "loss": 0.1399,
+      "step": 1505
+    },
+    {
+      "epoch": 3.5040650406504064,
+      "grad_norm": 0.826085090637207,
+      "learning_rate": 6.787664659678488e-06,
+      "loss": 0.1338,
+      "step": 1510
+    },
+    {
+      "epoch": 3.515679442508711,
+      "grad_norm": 1.0425338745117188,
+      "learning_rate": 6.691586249054209e-06,
+      "loss": 0.1348,
+      "step": 1515
+    },
+    {
+      "epoch": 3.527293844367015,
+      "grad_norm": 0.9699033498764038,
+      "learning_rate": 6.595997076285013e-06,
+      "loss": 0.1321,
+      "step": 1520
+    },
+    {
+      "epoch": 3.5389082462253194,
+      "grad_norm": 1.0192960500717163,
+      "learning_rate": 6.500902770103796e-06,
+      "loss": 0.1428,
+      "step": 1525
+    },
+    {
+      "epoch": 3.5505226480836236,
+      "grad_norm": 1.0220447778701782,
+      "learning_rate": 6.406308930103421e-06,
+      "loss": 0.1358,
+      "step": 1530
+    },
+    {
+      "epoch": 3.562137049941928,
+      "grad_norm": 1.0450464487075806,
+      "learning_rate": 6.312221126406996e-06,
+      "loss": 0.1339,
+      "step": 1535
+    },
+    {
+      "epoch": 3.5737514518002325,
+      "grad_norm": 1.0781066417694092,
+      "learning_rate": 6.218644899339873e-06,
+      "loss": 0.119,
+      "step": 1540
+    },
+    {
+      "epoch": 3.5853658536585367,
+      "grad_norm": 1.0588756799697876,
+      "learning_rate": 6.125585759103387e-06,
+      "loss": 0.1278,
+      "step": 1545
+    },
+    {
+      "epoch": 3.596980255516841,
+      "grad_norm": 1.2598525285720825,
+      "learning_rate": 6.033049185450418e-06,
+      "loss": 0.1173,
+      "step": 1550
+    },
+    {
+      "epoch": 3.608594657375145,
+      "grad_norm": 0.8823047876358032,
+      "learning_rate": 5.9410406273626974e-06,
+      "loss": 0.1504,
+      "step": 1555
+    },
+    {
+      "epoch": 3.6202090592334493,
+      "grad_norm": 0.9176443815231323,
+      "learning_rate": 5.8495655027299724e-06,
+      "loss": 0.1351,
+      "step": 1560
+    },
+    {
+      "epoch": 3.631823461091754,
+      "grad_norm": 0.9507741332054138,
+      "learning_rate": 5.758629198030956e-06,
+      "loss": 0.136,
+      "step": 1565
+    },
+    {
+      "epoch": 3.643437862950058,
+      "grad_norm": 0.9544185400009155,
+      "learning_rate": 5.668237068016138e-06,
+      "loss": 0.1467,
+      "step": 1570
+    },
+    {
+      "epoch": 3.6550522648083623,
+      "grad_norm": 1.1011279821395874,
+      "learning_rate": 5.578394435392486e-06,
+      "loss": 0.1351,
+      "step": 1575
+    },
+    {
+      "epoch": 3.6666666666666665,
+      "grad_norm": 1.0757453441619873,
+      "learning_rate": 5.489106590510035e-06,
+      "loss": 0.1184,
+      "step": 1580
+    },
+    {
+      "epoch": 3.6782810685249707,
+      "grad_norm": 0.8664600253105164,
+      "learning_rate": 5.400378791050328e-06,
+      "loss": 0.1143,
+      "step": 1585
+    },
+    {
+      "epoch": 3.6898954703832754,
+      "grad_norm": 0.904226541519165,
+      "learning_rate": 5.312216261716861e-06,
+      "loss": 0.118,
+      "step": 1590
+    },
+    {
+      "epoch": 3.7015098722415796,
+      "grad_norm": 1.0200014114379883,
+      "learning_rate": 5.224624193927388e-06,
+      "loss": 0.1097,
+      "step": 1595
+    },
+    {
+      "epoch": 3.7131242740998838,
+      "grad_norm": 0.8909503817558289,
+      "learning_rate": 5.137607745508271e-06,
+      "loss": 0.114,
+      "step": 1600
+    },
+    {
+      "epoch": 3.7247386759581884,
+      "grad_norm": 1.0821064710617065,
+      "learning_rate": 5.0511720403907204e-06,
+      "loss": 0.1285,
+      "step": 1605
+    },
+    {
+      "epoch": 3.736353077816492,
+      "grad_norm": 0.9839633703231812,
+      "learning_rate": 4.965322168309093e-06,
+      "loss": 0.1119,
+      "step": 1610
+    },
+    {
+      "epoch": 3.747967479674797,
+      "grad_norm": 0.9328600764274597,
+      "learning_rate": 4.880063184501202e-06,
+      "loss": 0.1042,
+      "step": 1615
+    },
+    {
+      "epoch": 3.759581881533101,
+      "grad_norm": 1.0208909511566162,
+      "learning_rate": 4.795400109410613e-06,
+      "loss": 0.1139,
+      "step": 1620
+    },
+    {
+      "epoch": 3.7711962833914052,
+      "grad_norm": 0.9615679383277893,
+      "learning_rate": 4.711337928391027e-06,
+      "loss": 0.1244,
+      "step": 1625
+    },
+    {
+      "epoch": 3.78281068524971,
+      "grad_norm": 0.7882293462753296,
+      "learning_rate": 4.627881591412723e-06,
+      "loss": 0.1282,
+      "step": 1630
+    },
+    {
+      "epoch": 3.794425087108014,
+      "grad_norm": 0.8788831830024719,
+      "learning_rate": 4.545036012771075e-06,
+      "loss": 0.1071,
+      "step": 1635
+    },
+    {
+      "epoch": 3.8060394889663183,
+      "grad_norm": 0.9033803939819336,
+      "learning_rate": 4.462806070797187e-06,
+      "loss": 0.1248,
+      "step": 1640
+    },
+    {
+      "epoch": 3.8176538908246225,
+      "grad_norm": 0.9137513637542725,
+      "learning_rate": 4.38119660757063e-06,
+      "loss": 0.1155,
+      "step": 1645
+    },
+    {
+      "epoch": 3.8292682926829267,
+      "grad_norm": 0.8888696432113647,
+      "learning_rate": 4.300212428634301e-06,
+      "loss": 0.1122,
+      "step": 1650
+    },
+    {
+      "epoch": 3.8408826945412313,
+      "grad_norm": 1.0764269828796387,
+      "learning_rate": 4.219858302711466e-06,
+      "loss": 0.1172,
+      "step": 1655
+    },
+    {
+      "epoch": 3.8524970963995355,
+      "grad_norm": 0.8962306380271912,
+      "learning_rate": 4.140138961424969e-06,
+      "loss": 0.1187,
+      "step": 1660
+    },
+    {
+      "epoch": 3.8641114982578397,
+      "grad_norm": 0.9213681817054749,
+      "learning_rate": 4.061059099018576e-06,
+      "loss": 0.1039,
+      "step": 1665
+    },
+    {
+      "epoch": 3.875725900116144,
+      "grad_norm": 0.8205775618553162,
+      "learning_rate": 3.982623372080598e-06,
+      "loss": 0.0991,
+      "step": 1670
+    },
+    {
+      "epoch": 3.887340301974448,
+      "grad_norm": 1.0516890287399292,
+      "learning_rate": 3.904836399269645e-06,
+      "loss": 0.1253,
+      "step": 1675
+    },
+    {
+      "epoch": 3.8989547038327528,
+      "grad_norm": 0.9499752521514893,
+      "learning_rate": 3.827702761042704e-06,
+      "loss": 0.1278,
+      "step": 1680
+    },
+    {
+      "epoch": 3.910569105691057,
+      "grad_norm": 0.8349530696868896,
+      "learning_rate": 3.751226999385384e-06,
+      "loss": 0.1211,
+      "step": 1685
+    },
+    {
+      "epoch": 3.922183507549361,
+      "grad_norm": 0.8646232485771179,
+      "learning_rate": 3.6754136175444763e-06,
+      "loss": 0.1143,
+      "step": 1690
+    },
+    {
+      "epoch": 3.9337979094076654,
+      "grad_norm": 0.8931161761283875,
+      "learning_rate": 3.600267079762789e-06,
+      "loss": 0.1194,
+      "step": 1695
+    },
+    {
+      "epoch": 3.9454123112659696,
+      "grad_norm": 0.9849768280982971,
+      "learning_rate": 3.525791811016271e-06,
+      "loss": 0.1092,
+      "step": 1700
+    },
+    {
+      "epoch": 3.9570267131242742,
+      "grad_norm": 0.9731475114822388,
+      "learning_rate": 3.451992196753429e-06,
+      "loss": 0.117,
+      "step": 1705
+    },
+    {
+      "epoch": 3.9686411149825784,
+      "grad_norm": 0.7634958028793335,
+      "learning_rate": 3.378872582637117e-06,
+      "loss": 0.1107,
+      "step": 1710
+    },
+    {
+      "epoch": 3.9802555168408826,
+      "grad_norm": 0.8839854598045349,
+      "learning_rate": 3.306437274288624e-06,
+      "loss": 0.0999,
+      "step": 1715
+    },
+    {
+      "epoch": 3.991869918699187,
+      "grad_norm": 0.8653486371040344,
+      "learning_rate": 3.2346905370341575e-06,
+      "loss": 0.1106,
+      "step": 1720
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2155,
+  "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": 2.713126404853072e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

25_128_e5_3e-5/checkpoint-1724/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8826cd3cb0315b085e0bee4ad616aa4a2c50addf0b6c688836791cb1c313dff8
+size 7736

25_128_e5_3e-5/checkpoint-1724/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)

25_128_e5_3e-5/checkpoint-2155/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

25_128_e5_3e-5/checkpoint-2155/adapter_config.json

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

25_128_e5_3e-5/checkpoint-2155/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4658cb8a7bc4762d90cdaf09d89611499c1e3926ea404b293368fb0a838a2358
+size 791751704

25_128_e5_3e-5/checkpoint-2155/latest

@@ -0,0 +1 @@
+global_step2155

25_128_e5_3e-5/checkpoint-2155/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17f21a1b6500e0ff35b10a9c9785c7a966fe117e4b5ba284c1edee4c87fee24a
+size 15920

25_128_e5_3e-5/checkpoint-2155/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca3a3c0620a6ca1db565b5fd99e6eafc4ee7f9fddf53e6bd873b72d554ef454b
+size 15920

25_128_e5_3e-5/checkpoint-2155/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:539371ae490ba8ad974e60d6a6970ef1a269ec67453cdd4acd8034fb3c0f1d25
+size 15920