Upload folder using huggingface_hub

23_128_e5_3e-5/checkpoint-1083/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

23_128_e5_3e-5/checkpoint-1083/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": [
+    "up_proj",
+    "q_proj",
+    "down_proj",
+    "v_proj",
+    "gate_proj",
+    "o_proj",
+    "k_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

23_128_e5_3e-5/checkpoint-1083/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4accfeed6d83217a6237a74c968be99cf5c64efff6c41359bbfb681c8037c1d0
+size 791751704

23_128_e5_3e-5/checkpoint-1083/latest

@@ -0,0 +1 @@
+global_step1083

23_128_e5_3e-5/checkpoint-1083/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:583d4b9beabaf57c9a953807f4cc7872b3da37520c031fdeaa12c675a9799163
+size 15920

23_128_e5_3e-5/checkpoint-1083/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8cef2fac67b36c25d05c15a94a843be4421ddeccef315601eb98b50f65b79f34
+size 15920

23_128_e5_3e-5/checkpoint-1083/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6f54db6e48bab1cd7d37543298925620ece61293a1d0a84f66cd626e5f4ef64
+size 15920

23_128_e5_3e-5/checkpoint-1083/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:87fab82a63f40bb0a304cb4f1547bb5c977cf448cedd675bd20a4632427912d7
+size 15920

23_128_e5_3e-5/checkpoint-1083/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22c434d78fe58bd8b3547c6281b2b23c392e19daa41f3da683590eccc9018aad
+size 15920

23_128_e5_3e-5/checkpoint-1083/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94c157ed07883e25ddec8c9f7a126eacf04d068f2125d15e0b2a8e2cb6eb285c
+size 15920

23_128_e5_3e-5/checkpoint-1083/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83d7690fb52ac6dda754cc020efe672f137835d5f950b2b14dd42da7a6278926
+size 15920

23_128_e5_3e-5/checkpoint-1083/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e715d5d2ce3a95fe2b7d7c5122452f5ed957256913b1edc01dfd256399bc563c
+size 15920

23_128_e5_3e-5/checkpoint-1083/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a446880b41412857320d133adee10e2b83a504e79721f855dfcf02f65f4ad060
+size 1064

23_128_e5_3e-5/checkpoint-1083/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
+  }
+}

23_128_e5_3e-5/checkpoint-1083/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
+}

23_128_e5_3e-5/checkpoint-1083/trainer_state.json

@@ -0,0 +1,1546 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1083,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.013869625520110958,
+      "grad_norm": 1.2640684843063354,
+      "learning_rate": 1.3186813186813187e-06,
+      "loss": 1.3135,
+      "step": 5
+    },
+    {
+      "epoch": 0.027739251040221916,
+      "grad_norm": 1.024712085723877,
+      "learning_rate": 2.967032967032967e-06,
+      "loss": 1.3432,
+      "step": 10
+    },
+    {
+      "epoch": 0.04160887656033287,
+      "grad_norm": 0.7255855798721313,
+      "learning_rate": 4.615384615384616e-06,
+      "loss": 1.3293,
+      "step": 15
+    },
+    {
+      "epoch": 0.05547850208044383,
+      "grad_norm": 0.5718052983283997,
+      "learning_rate": 6.2637362637362645e-06,
+      "loss": 1.3092,
+      "step": 20
+    },
+    {
+      "epoch": 0.06934812760055478,
+      "grad_norm": 0.6330448985099792,
+      "learning_rate": 7.912087912087913e-06,
+      "loss": 1.3039,
+      "step": 25
+    },
+    {
+      "epoch": 0.08321775312066575,
+      "grad_norm": 0.7984022498130798,
+      "learning_rate": 9.56043956043956e-06,
+      "loss": 1.2686,
+      "step": 30
+    },
+    {
+      "epoch": 0.0970873786407767,
+      "grad_norm": 0.4936966300010681,
+      "learning_rate": 1.120879120879121e-05,
+      "loss": 1.2573,
+      "step": 35
+    },
+    {
+      "epoch": 0.11095700416088766,
+      "grad_norm": 0.5313364863395691,
+      "learning_rate": 1.2857142857142857e-05,
+      "loss": 1.2233,
+      "step": 40
+    },
+    {
+      "epoch": 0.12482662968099861,
+      "grad_norm": 0.4598558843135834,
+      "learning_rate": 1.4505494505494506e-05,
+      "loss": 1.2457,
+      "step": 45
+    },
+    {
+      "epoch": 0.13869625520110956,
+      "grad_norm": 0.4705202877521515,
+      "learning_rate": 1.6153846153846154e-05,
+      "loss": 1.2183,
+      "step": 50
+    },
+    {
+      "epoch": 0.15256588072122051,
+      "grad_norm": 0.6102734208106995,
+      "learning_rate": 1.78021978021978e-05,
+      "loss": 1.1374,
+      "step": 55
+    },
+    {
+      "epoch": 0.1664355062413315,
+      "grad_norm": 0.644282877445221,
+      "learning_rate": 1.9450549450549452e-05,
+      "loss": 1.1574,
+      "step": 60
+    },
+    {
+      "epoch": 0.18030513176144244,
+      "grad_norm": 0.5232361555099487,
+      "learning_rate": 2.10989010989011e-05,
+      "loss": 1.1972,
+      "step": 65
+    },
+    {
+      "epoch": 0.1941747572815534,
+      "grad_norm": 0.5031052827835083,
+      "learning_rate": 2.2747252747252748e-05,
+      "loss": 1.2067,
+      "step": 70
+    },
+    {
+      "epoch": 0.20804438280166435,
+      "grad_norm": 0.5804685950279236,
+      "learning_rate": 2.4395604395604395e-05,
+      "loss": 1.1819,
+      "step": 75
+    },
+    {
+      "epoch": 0.22191400832177532,
+      "grad_norm": 0.4593973159790039,
+      "learning_rate": 2.6043956043956046e-05,
+      "loss": 1.1541,
+      "step": 80
+    },
+    {
+      "epoch": 0.23578363384188628,
+      "grad_norm": 0.4562990367412567,
+      "learning_rate": 2.7692307692307694e-05,
+      "loss": 1.1592,
+      "step": 85
+    },
+    {
+      "epoch": 0.24965325936199723,
+      "grad_norm": 0.5135698318481445,
+      "learning_rate": 2.934065934065934e-05,
+      "loss": 1.1133,
+      "step": 90
+    },
+    {
+      "epoch": 0.2635228848821082,
+      "grad_norm": 0.5088622570037842,
+      "learning_rate": 2.9999773232512234e-05,
+      "loss": 1.121,
+      "step": 95
+    },
+    {
+      "epoch": 0.27739251040221913,
+      "grad_norm": 0.5037941336631775,
+      "learning_rate": 2.9998387456028022e-05,
+      "loss": 1.0824,
+      "step": 100
+    },
+    {
+      "epoch": 0.2912621359223301,
+      "grad_norm": 0.6273325085639954,
+      "learning_rate": 2.999574200124419e-05,
+      "loss": 1.1311,
+      "step": 105
+    },
+    {
+      "epoch": 0.30513176144244103,
+      "grad_norm": 0.6029002070426941,
+      "learning_rate": 2.999183709034608e-05,
+      "loss": 1.061,
+      "step": 110
+    },
+    {
+      "epoch": 0.31900138696255204,
+      "grad_norm": 0.49770888686180115,
+      "learning_rate": 2.998667305129772e-05,
+      "loss": 1.0573,
+      "step": 115
+    },
+    {
+      "epoch": 0.332871012482663,
+      "grad_norm": 0.6382117867469788,
+      "learning_rate": 2.9980250317814265e-05,
+      "loss": 1.0497,
+      "step": 120
+    },
+    {
+      "epoch": 0.34674063800277394,
+      "grad_norm": 0.5979148149490356,
+      "learning_rate": 2.9972569429325575e-05,
+      "loss": 1.0705,
+      "step": 125
+    },
+    {
+      "epoch": 0.3606102635228849,
+      "grad_norm": 0.5902369618415833,
+      "learning_rate": 2.99636310309309e-05,
+      "loss": 1.0601,
+      "step": 130
+    },
+    {
+      "epoch": 0.37447988904299584,
+      "grad_norm": 0.5844177007675171,
+      "learning_rate": 2.995343587334471e-05,
+      "loss": 0.9897,
+      "step": 135
+    },
+    {
+      "epoch": 0.3883495145631068,
+      "grad_norm": 0.5972667336463928,
+      "learning_rate": 2.994198481283364e-05,
+      "loss": 1.0212,
+      "step": 140
+    },
+    {
+      "epoch": 0.40221914008321774,
+      "grad_norm": 0.6355414390563965,
+      "learning_rate": 2.992927881114458e-05,
+      "loss": 1.0097,
+      "step": 145
+    },
+    {
+      "epoch": 0.4160887656033287,
+      "grad_norm": 0.6107580065727234,
+      "learning_rate": 2.991531893542389e-05,
+      "loss": 0.9994,
+      "step": 150
+    },
+    {
+      "epoch": 0.42995839112343964,
+      "grad_norm": 0.7091370224952698,
+      "learning_rate": 2.990010635812777e-05,
+      "loss": 0.9766,
+      "step": 155
+    },
+    {
+      "epoch": 0.44382801664355065,
+      "grad_norm": 0.6339324116706848,
+      "learning_rate": 2.9883642356923822e-05,
+      "loss": 1.0056,
+      "step": 160
+    },
+    {
+      "epoch": 0.4576976421636616,
+      "grad_norm": 0.6281903386116028,
+      "learning_rate": 2.986592831458369e-05,
+      "loss": 0.9624,
+      "step": 165
+    },
+    {
+      "epoch": 0.47156726768377255,
+      "grad_norm": 0.7146309018135071,
+      "learning_rate": 2.984696571886697e-05,
+      "loss": 0.9955,
+      "step": 170
+    },
+    {
+      "epoch": 0.4854368932038835,
+      "grad_norm": 0.852715015411377,
+      "learning_rate": 2.9826756162396226e-05,
+      "loss": 0.9626,
+      "step": 175
+    },
+    {
+      "epoch": 0.49930651872399445,
+      "grad_norm": 0.7176477909088135,
+      "learning_rate": 2.9805301342523238e-05,
+      "loss": 0.9451,
+      "step": 180
+    },
+    {
+      "epoch": 0.5131761442441054,
+      "grad_norm": 0.6825271248817444,
+      "learning_rate": 2.9782603061186458e-05,
+      "loss": 0.9969,
+      "step": 185
+    },
+    {
+      "epoch": 0.5270457697642164,
+      "grad_norm": 0.7406214475631714,
+      "learning_rate": 2.975866322475966e-05,
+      "loss": 0.945,
+      "step": 190
+    },
+    {
+      "epoch": 0.5409153952843273,
+      "grad_norm": 0.8366868495941162,
+      "learning_rate": 2.973348384389182e-05,
+      "loss": 0.9201,
+      "step": 195
+    },
+    {
+      "epoch": 0.5547850208044383,
+      "grad_norm": 0.8170830011367798,
+      "learning_rate": 2.9707067033338257e-05,
+      "loss": 0.9614,
+      "step": 200
+    },
+    {
+      "epoch": 0.5686546463245492,
+      "grad_norm": 0.7892953157424927,
+      "learning_rate": 2.967941501178302e-05,
+      "loss": 0.9041,
+      "step": 205
+    },
+    {
+      "epoch": 0.5825242718446602,
+      "grad_norm": 0.874885082244873,
+      "learning_rate": 2.9650530101652548e-05,
+      "loss": 0.9418,
+      "step": 210
+    },
+    {
+      "epoch": 0.5963938973647711,
+      "grad_norm": 0.8003681302070618,
+      "learning_rate": 2.9620414728920603e-05,
+      "loss": 0.8867,
+      "step": 215
+    },
+    {
+      "epoch": 0.6102635228848821,
+      "grad_norm": 0.9193400740623474,
+      "learning_rate": 2.9589071422904525e-05,
+      "loss": 0.9073,
+      "step": 220
+    },
+    {
+      "epoch": 0.624133148404993,
+      "grad_norm": 0.8814607858657837,
+      "learning_rate": 2.9556502816052805e-05,
+      "loss": 0.8449,
+      "step": 225
+    },
+    {
+      "epoch": 0.6380027739251041,
+      "grad_norm": 0.8270254135131836,
+      "learning_rate": 2.9522711643723997e-05,
+      "loss": 0.8744,
+      "step": 230
+    },
+    {
+      "epoch": 0.651872399445215,
+      "grad_norm": 0.8522059321403503,
+      "learning_rate": 2.9487700743956955e-05,
+      "loss": 0.8211,
+      "step": 235
+    },
+    {
+      "epoch": 0.665742024965326,
+      "grad_norm": 0.8157532215118408,
+      "learning_rate": 2.945147305723251e-05,
+      "loss": 0.8436,
+      "step": 240
+    },
+    {
+      "epoch": 0.6796116504854369,
+      "grad_norm": 0.9486557841300964,
+      "learning_rate": 2.9414031626226472e-05,
+      "loss": 0.7924,
+      "step": 245
+    },
+    {
+      "epoch": 0.6934812760055479,
+      "grad_norm": 0.7984339594841003,
+      "learning_rate": 2.9375379595554113e-05,
+      "loss": 0.8411,
+      "step": 250
+    },
+    {
+      "epoch": 0.7073509015256588,
+      "grad_norm": 0.9337722659111023,
+      "learning_rate": 2.9335520211506033e-05,
+      "loss": 0.8287,
+      "step": 255
+    },
+    {
+      "epoch": 0.7212205270457698,
+      "grad_norm": 0.8872073292732239,
+      "learning_rate": 2.9294456821775524e-05,
+      "loss": 0.794,
+      "step": 260
+    },
+    {
+      "epoch": 0.7350901525658807,
+      "grad_norm": 0.951157808303833,
+      "learning_rate": 2.9252192875177415e-05,
+      "loss": 0.8033,
+      "step": 265
+    },
+    {
+      "epoch": 0.7489597780859917,
+      "grad_norm": 0.9538988471031189,
+      "learning_rate": 2.920873192135839e-05,
+      "loss": 0.7984,
+      "step": 270
+    },
+    {
+      "epoch": 0.7628294036061026,
+      "grad_norm": 0.9138048887252808,
+      "learning_rate": 2.9164077610498886e-05,
+      "loss": 0.8138,
+      "step": 275
+    },
+    {
+      "epoch": 0.7766990291262136,
+      "grad_norm": 0.8404424786567688,
+      "learning_rate": 2.9118233693006503e-05,
+      "loss": 0.7732,
+      "step": 280
+    },
+    {
+      "epoch": 0.7905686546463245,
+      "grad_norm": 0.8813743591308594,
+      "learning_rate": 2.907120401920103e-05,
+      "loss": 0.7616,
+      "step": 285
+    },
+    {
+      "epoch": 0.8044382801664355,
+      "grad_norm": 0.9614547491073608,
+      "learning_rate": 2.9022992538991067e-05,
+      "loss": 0.8015,
+      "step": 290
+    },
+    {
+      "epoch": 0.8183079056865464,
+      "grad_norm": 0.8543416857719421,
+      "learning_rate": 2.897360330154227e-05,
+      "loss": 0.7915,
+      "step": 295
+    },
+    {
+      "epoch": 0.8321775312066574,
+      "grad_norm": 0.9641604423522949,
+      "learning_rate": 2.892304045493728e-05,
+      "loss": 0.708,
+      "step": 300
+    },
+    {
+      "epoch": 0.8460471567267683,
+      "grad_norm": 0.9229587912559509,
+      "learning_rate": 2.8871308245827336e-05,
+      "loss": 0.732,
+      "step": 305
+    },
+    {
+      "epoch": 0.8599167822468793,
+      "grad_norm": 0.9859731197357178,
+      "learning_rate": 2.88184110190756e-05,
+      "loss": 0.7759,
+      "step": 310
+    },
+    {
+      "epoch": 0.8737864077669902,
+      "grad_norm": 1.0025079250335693,
+      "learning_rate": 2.8764353217392253e-05,
+      "loss": 0.7113,
+      "step": 315
+    },
+    {
+      "epoch": 0.8876560332871013,
+      "grad_norm": 0.971370279788971,
+      "learning_rate": 2.870913938096136e-05,
+      "loss": 0.738,
+      "step": 320
+    },
+    {
+      "epoch": 0.9015256588072122,
+      "grad_norm": 1.0351135730743408,
+      "learning_rate": 2.865277414705955e-05,
+      "loss": 0.7727,
+      "step": 325
+    },
+    {
+      "epoch": 0.9153952843273232,
+      "grad_norm": 0.9907004833221436,
+      "learning_rate": 2.8595262249666536e-05,
+      "loss": 0.7487,
+      "step": 330
+    },
+    {
+      "epoch": 0.9292649098474342,
+      "grad_norm": 1.1555078029632568,
+      "learning_rate": 2.8536608519067532e-05,
+      "loss": 0.7369,
+      "step": 335
+    },
+    {
+      "epoch": 0.9431345353675451,
+      "grad_norm": 0.9639641642570496,
+      "learning_rate": 2.8476817881447555e-05,
+      "loss": 0.7045,
+      "step": 340
+    },
+    {
+      "epoch": 0.957004160887656,
+      "grad_norm": 0.9972969889640808,
+      "learning_rate": 2.8415895358477702e-05,
+      "loss": 0.6953,
+      "step": 345
+    },
+    {
+      "epoch": 0.970873786407767,
+      "grad_norm": 1.2426544427871704,
+      "learning_rate": 2.8353846066893382e-05,
+      "loss": 0.6732,
+      "step": 350
+    },
+    {
+      "epoch": 0.984743411927878,
+      "grad_norm": 1.1239923238754272,
+      "learning_rate": 2.8290675218064565e-05,
+      "loss": 0.6729,
+      "step": 355
+    },
+    {
+      "epoch": 0.9986130374479889,
+      "grad_norm": 1.099241852760315,
+      "learning_rate": 2.822638811755812e-05,
+      "loss": 0.6424,
+      "step": 360
+    },
+    {
+      "epoch": 1.0110957004160888,
+      "grad_norm": 0.9136450290679932,
+      "learning_rate": 2.8160990164692183e-05,
+      "loss": 0.5966,
+      "step": 365
+    },
+    {
+      "epoch": 1.0249653259361997,
+      "grad_norm": 1.128823161125183,
+      "learning_rate": 2.8094486852082692e-05,
+      "loss": 0.5859,
+      "step": 370
+    },
+    {
+      "epoch": 1.0388349514563107,
+      "grad_norm": 1.0320087671279907,
+      "learning_rate": 2.802688376518209e-05,
+      "loss": 0.6032,
+      "step": 375
+    },
+    {
+      "epoch": 1.0527045769764216,
+      "grad_norm": 0.9852442741394043,
+      "learning_rate": 2.7958186581810195e-05,
+      "loss": 0.5763,
+      "step": 380
+    },
+    {
+      "epoch": 1.0665742024965326,
+      "grad_norm": 1.0142955780029297,
+      "learning_rate": 2.7888401071677342e-05,
+      "loss": 0.615,
+      "step": 385
+    },
+    {
+      "epoch": 1.0804438280166435,
+      "grad_norm": 1.1618216037750244,
+      "learning_rate": 2.7817533095899806e-05,
+      "loss": 0.6023,
+      "step": 390
+    },
+    {
+      "epoch": 1.0943134535367545,
+      "grad_norm": 1.001564621925354,
+      "learning_rate": 2.774558860650752e-05,
+      "loss": 0.5843,
+      "step": 395
+    },
+    {
+      "epoch": 1.1081830790568654,
+      "grad_norm": 1.1653980016708374,
+      "learning_rate": 2.767257364594421e-05,
+      "loss": 0.5763,
+      "step": 400
+    },
+    {
+      "epoch": 1.1220527045769764,
+      "grad_norm": 1.349169373512268,
+      "learning_rate": 2.759849434655987e-05,
+      "loss": 0.6157,
+      "step": 405
+    },
+    {
+      "epoch": 1.1359223300970873,
+      "grad_norm": 1.1331448554992676,
+      "learning_rate": 2.7523356930095764e-05,
+      "loss": 0.5724,
+      "step": 410
+    },
+    {
+      "epoch": 1.1497919556171983,
+      "grad_norm": 1.133574366569519,
+      "learning_rate": 2.744716770716182e-05,
+      "loss": 0.5866,
+      "step": 415
+    },
+    {
+      "epoch": 1.1636615811373092,
+      "grad_norm": 1.164962649345398,
+      "learning_rate": 2.736993307670667e-05,
+      "loss": 0.5861,
+      "step": 420
+    },
+    {
+      "epoch": 1.1775312066574202,
+      "grad_norm": 1.0958647727966309,
+      "learning_rate": 2.7291659525480192e-05,
+      "loss": 0.5564,
+      "step": 425
+    },
+    {
+      "epoch": 1.1914008321775311,
+      "grad_norm": 1.6513663530349731,
+      "learning_rate": 2.72123536274887e-05,
+      "loss": 0.5444,
+      "step": 430
+    },
+    {
+      "epoch": 1.205270457697642,
+      "grad_norm": 1.0394660234451294,
+      "learning_rate": 2.713202204344282e-05,
+      "loss": 0.5366,
+      "step": 435
+    },
+    {
+      "epoch": 1.219140083217753,
+      "grad_norm": 1.290332555770874,
+      "learning_rate": 2.705067152019808e-05,
+      "loss": 0.5414,
+      "step": 440
+    },
+    {
+      "epoch": 1.233009708737864,
+      "grad_norm": 1.1973203420639038,
+      "learning_rate": 2.6968308890188235e-05,
+      "loss": 0.5563,
+      "step": 445
+    },
+    {
+      "epoch": 1.246879334257975,
+      "grad_norm": 1.1321580410003662,
+      "learning_rate": 2.6884941070851443e-05,
+      "loss": 0.5668,
+      "step": 450
+    },
+    {
+      "epoch": 1.2607489597780859,
+      "grad_norm": 1.1653034687042236,
+      "learning_rate": 2.680057506404929e-05,
+      "loss": 0.5898,
+      "step": 455
+    },
+    {
+      "epoch": 1.2746185852981968,
+      "grad_norm": 1.271990180015564,
+      "learning_rate": 2.6715217955478705e-05,
+      "loss": 0.4863,
+      "step": 460
+    },
+    {
+      "epoch": 1.2884882108183078,
+      "grad_norm": 1.083730697631836,
+      "learning_rate": 2.6628876914076865e-05,
+      "loss": 0.496,
+      "step": 465
+    },
+    {
+      "epoch": 1.3023578363384187,
+      "grad_norm": 1.3474218845367432,
+      "learning_rate": 2.654155919141908e-05,
+      "loss": 0.5484,
+      "step": 470
+    },
+    {
+      "epoch": 1.31622746185853,
+      "grad_norm": 1.100876808166504,
+      "learning_rate": 2.645327212110976e-05,
+      "loss": 0.4999,
+      "step": 475
+    },
+    {
+      "epoch": 1.3300970873786409,
+      "grad_norm": 1.1323593854904175,
+      "learning_rate": 2.636402311816649e-05,
+      "loss": 0.5368,
+      "step": 480
+    },
+    {
+      "epoch": 1.3439667128987518,
+      "grad_norm": 1.2535606622695923,
+      "learning_rate": 2.6273819678397233e-05,
+      "loss": 0.5537,
+      "step": 485
+    },
+    {
+      "epoch": 1.3578363384188628,
+      "grad_norm": 1.1354525089263916,
+      "learning_rate": 2.6182669377770807e-05,
+      "loss": 0.4727,
+      "step": 490
+    },
+    {
+      "epoch": 1.3717059639389737,
+      "grad_norm": 1.2694000005722046,
+      "learning_rate": 2.6090579871780583e-05,
+      "loss": 0.5171,
+      "step": 495
+    },
+    {
+      "epoch": 1.3855755894590847,
+      "grad_norm": 1.2506495714187622,
+      "learning_rate": 2.5997558894801514e-05,
+      "loss": 0.472,
+      "step": 500
+    },
+    {
+      "epoch": 1.3994452149791956,
+      "grad_norm": 1.1564306020736694,
+      "learning_rate": 2.5903614259440553e-05,
+      "loss": 0.5141,
+      "step": 505
+    },
+    {
+      "epoch": 1.4133148404993066,
+      "grad_norm": 1.1961454153060913,
+      "learning_rate": 2.580875385588048e-05,
+      "loss": 0.4358,
+      "step": 510
+    },
+    {
+      "epoch": 1.4271844660194175,
+      "grad_norm": 1.0629825592041016,
+      "learning_rate": 2.571298565121725e-05,
+      "loss": 0.4934,
+      "step": 515
+    },
+    {
+      "epoch": 1.4410540915395285,
+      "grad_norm": 1.1157474517822266,
+      "learning_rate": 2.561631768879082e-05,
+      "loss": 0.4666,
+      "step": 520
+    },
+    {
+      "epoch": 1.4549237170596394,
+      "grad_norm": 1.1023491621017456,
+      "learning_rate": 2.551875808750963e-05,
+      "loss": 0.4786,
+      "step": 525
+    },
+    {
+      "epoch": 1.4687933425797504,
+      "grad_norm": 1.3004133701324463,
+      "learning_rate": 2.5420315041168717e-05,
+      "loss": 0.5161,
+      "step": 530
+    },
+    {
+      "epoch": 1.4826629680998613,
+      "grad_norm": 1.2039952278137207,
+      "learning_rate": 2.5320996817761534e-05,
+      "loss": 0.4471,
+      "step": 535
+    },
+    {
+      "epoch": 1.4965325936199723,
+      "grad_norm": 1.3886206150054932,
+      "learning_rate": 2.5220811758785525e-05,
+      "loss": 0.4796,
+      "step": 540
+    },
+    {
+      "epoch": 1.5104022191400832,
+      "grad_norm": 1.13352370262146,
+      "learning_rate": 2.5119768278541576e-05,
+      "loss": 0.4782,
+      "step": 545
+    },
+    {
+      "epoch": 1.5242718446601942,
+      "grad_norm": 1.2641507387161255,
+      "learning_rate": 2.5017874863427292e-05,
+      "loss": 0.4516,
+      "step": 550
+    },
+    {
+      "epoch": 1.5381414701803051,
+      "grad_norm": 1.3237358331680298,
+      "learning_rate": 2.4915140071224247e-05,
+      "loss": 0.451,
+      "step": 555
+    },
+    {
+      "epoch": 1.552011095700416,
+      "grad_norm": 1.1909607648849487,
+      "learning_rate": 2.4811572530379253e-05,
+      "loss": 0.4389,
+      "step": 560
+    },
+    {
+      "epoch": 1.565880721220527,
+      "grad_norm": 1.2734519243240356,
+      "learning_rate": 2.4707180939279658e-05,
+      "loss": 0.4822,
+      "step": 565
+    },
+    {
+      "epoch": 1.579750346740638,
+      "grad_norm": 1.2859396934509277,
+      "learning_rate": 2.46019740655228e-05,
+      "loss": 0.4503,
+      "step": 570
+    },
+    {
+      "epoch": 1.593619972260749,
+      "grad_norm": 1.0753681659698486,
+      "learning_rate": 2.4495960745179646e-05,
+      "loss": 0.4571,
+      "step": 575
+    },
+    {
+      "epoch": 1.6074895977808599,
+      "grad_norm": 1.0502837896347046,
+      "learning_rate": 2.4389149882052654e-05,
+      "loss": 0.4532,
+      "step": 580
+    },
+    {
+      "epoch": 1.6213592233009708,
+      "grad_norm": 1.1627888679504395,
+      "learning_rate": 2.4281550446927968e-05,
+      "loss": 0.4613,
+      "step": 585
+    },
+    {
+      "epoch": 1.635228848821082,
+      "grad_norm": 1.1929353475570679,
+      "learning_rate": 2.4173171476821997e-05,
+      "loss": 0.4411,
+      "step": 590
+    },
+    {
+      "epoch": 1.649098474341193,
+      "grad_norm": 1.1515079736709595,
+      "learning_rate": 2.40640220742224e-05,
+      "loss": 0.427,
+      "step": 595
+    },
+    {
+      "epoch": 1.662968099861304,
+      "grad_norm": 1.0611435174942017,
+      "learning_rate": 2.3954111406323607e-05,
+      "loss": 0.4296,
+      "step": 600
+    },
+    {
+      "epoch": 1.6768377253814148,
+      "grad_norm": 1.2731926441192627,
+      "learning_rate": 2.3843448704256868e-05,
+      "loss": 0.3959,
+      "step": 605
+    },
+    {
+      "epoch": 1.6907073509015258,
+      "grad_norm": 1.104193091392517,
+      "learning_rate": 2.3732043262314974e-05,
+      "loss": 0.4378,
+      "step": 610
+    },
+    {
+      "epoch": 1.7045769764216367,
+      "grad_norm": 1.2957518100738525,
+      "learning_rate": 2.3619904437171642e-05,
+      "loss": 0.4088,
+      "step": 615
+    },
+    {
+      "epoch": 1.7184466019417477,
+      "grad_norm": 1.1752687692642212,
+      "learning_rate": 2.3507041647095668e-05,
+      "loss": 0.4095,
+      "step": 620
+    },
+    {
+      "epoch": 1.7323162274618586,
+      "grad_norm": 1.1501843929290771,
+      "learning_rate": 2.3393464371159914e-05,
+      "loss": 0.39,
+      "step": 625
+    },
+    {
+      "epoch": 1.7461858529819696,
+      "grad_norm": 1.2524619102478027,
+      "learning_rate": 2.327918214844518e-05,
+      "loss": 0.4312,
+      "step": 630
+    },
+    {
+      "epoch": 1.7600554785020806,
+      "grad_norm": 1.1194138526916504,
+      "learning_rate": 2.3164204577239037e-05,
+      "loss": 0.3986,
+      "step": 635
+    },
+    {
+      "epoch": 1.7739251040221915,
+      "grad_norm": 1.197754979133606,
+      "learning_rate": 2.304854131422971e-05,
+      "loss": 0.4333,
+      "step": 640
+    },
+    {
+      "epoch": 1.7877947295423025,
+      "grad_norm": 1.1752138137817383,
+      "learning_rate": 2.2932202073695003e-05,
+      "loss": 0.389,
+      "step": 645
+    },
+    {
+      "epoch": 1.8016643550624134,
+      "grad_norm": 1.1883151531219482,
+      "learning_rate": 2.281519662668645e-05,
+      "loss": 0.3669,
+      "step": 650
+    },
+    {
+      "epoch": 1.8155339805825244,
+      "grad_norm": 1.2165765762329102,
+      "learning_rate": 2.2697534800208655e-05,
+      "loss": 0.4625,
+      "step": 655
+    },
+    {
+      "epoch": 1.8294036061026353,
+      "grad_norm": 1.1365174055099487,
+      "learning_rate": 2.2579226476393932e-05,
+      "loss": 0.4195,
+      "step": 660
+    },
+    {
+      "epoch": 1.8432732316227463,
+      "grad_norm": 1.1491037607192993,
+      "learning_rate": 2.2460281591672365e-05,
+      "loss": 0.4132,
+      "step": 665
+    },
+    {
+      "epoch": 1.8571428571428572,
+      "grad_norm": 1.3076406717300415,
+      "learning_rate": 2.2340710135937233e-05,
+      "loss": 0.3854,
+      "step": 670
+    },
+    {
+      "epoch": 1.8710124826629682,
+      "grad_norm": 1.1297541856765747,
+      "learning_rate": 2.2220522151706012e-05,
+      "loss": 0.4352,
+      "step": 675
+    },
+    {
+      "epoch": 1.884882108183079,
+      "grad_norm": 1.168922781944275,
+      "learning_rate": 2.2099727733276905e-05,
+      "loss": 0.3606,
+      "step": 680
+    },
+    {
+      "epoch": 1.89875173370319,
+      "grad_norm": 1.1461069583892822,
+      "learning_rate": 2.1978337025881062e-05,
+      "loss": 0.3786,
+      "step": 685
+    },
+    {
+      "epoch": 1.912621359223301,
+      "grad_norm": 1.3558424711227417,
+      "learning_rate": 2.18563602248305e-05,
+      "loss": 0.3852,
+      "step": 690
+    },
+    {
+      "epoch": 1.926490984743412,
+      "grad_norm": 1.1597490310668945,
+      "learning_rate": 2.1733807574661823e-05,
+      "loss": 0.3721,
+      "step": 695
+    },
+    {
+      "epoch": 1.940360610263523,
+      "grad_norm": 1.5078226327896118,
+      "learning_rate": 2.1610689368275827e-05,
+      "loss": 0.4072,
+      "step": 700
+    },
+    {
+      "epoch": 1.9542302357836339,
+      "grad_norm": 1.1332188844680786,
+      "learning_rate": 2.148701594607298e-05,
+      "loss": 0.3278,
+      "step": 705
+    },
+    {
+      "epoch": 1.9680998613037448,
+      "grad_norm": 1.2428522109985352,
+      "learning_rate": 2.1362797695085e-05,
+      "loss": 0.3739,
+      "step": 710
+    },
+    {
+      "epoch": 1.9819694868238558,
+      "grad_norm": 1.3413153886795044,
+      "learning_rate": 2.123804504810246e-05,
+      "loss": 0.3625,
+      "step": 715
+    },
+    {
+      "epoch": 1.9958391123439667,
+      "grad_norm": 1.2263511419296265,
+      "learning_rate": 2.1112768482798553e-05,
+      "loss": 0.3571,
+      "step": 720
+    },
+    {
+      "epoch": 2.0083217753120666,
+      "grad_norm": 1.1522630453109741,
+      "learning_rate": 2.0986978520849108e-05,
+      "loss": 0.3112,
+      "step": 725
+    },
+    {
+      "epoch": 2.0221914008321775,
+      "grad_norm": 1.0950467586517334,
+      "learning_rate": 2.0860685727048895e-05,
+      "loss": 0.2632,
+      "step": 730
+    },
+    {
+      "epoch": 2.0360610263522885,
+      "grad_norm": 1.1049528121948242,
+      "learning_rate": 2.0733900708424318e-05,
+      "loss": 0.3054,
+      "step": 735
+    },
+    {
+      "epoch": 2.0499306518723994,
+      "grad_norm": 1.146390676498413,
+      "learning_rate": 2.060663411334255e-05,
+      "loss": 0.3114,
+      "step": 740
+    },
+    {
+      "epoch": 2.0638002773925104,
+      "grad_norm": 1.360620141029358,
+      "learning_rate": 2.0478896630617216e-05,
+      "loss": 0.2868,
+      "step": 745
+    },
+    {
+      "epoch": 2.0776699029126213,
+      "grad_norm": 1.2842508554458618,
+      "learning_rate": 2.0350698988610645e-05,
+      "loss": 0.311,
+      "step": 750
+    },
+    {
+      "epoch": 2.0915395284327323,
+      "grad_norm": 1.1546225547790527,
+      "learning_rate": 2.0222051954332835e-05,
+      "loss": 0.321,
+      "step": 755
+    },
+    {
+      "epoch": 2.1054091539528432,
+      "grad_norm": 1.112905502319336,
+      "learning_rate": 2.0092966332537152e-05,
+      "loss": 0.2572,
+      "step": 760
+    },
+    {
+      "epoch": 2.119278779472954,
+      "grad_norm": 1.122055172920227,
+      "learning_rate": 1.9963452964812873e-05,
+      "loss": 0.268,
+      "step": 765
+    },
+    {
+      "epoch": 2.133148404993065,
+      "grad_norm": 1.007519006729126,
+      "learning_rate": 1.9833522728674624e-05,
+      "loss": 0.2998,
+      "step": 770
+    },
+    {
+      "epoch": 2.147018030513176,
+      "grad_norm": 1.0647330284118652,
+      "learning_rate": 1.9703186536648786e-05,
+      "loss": 0.2406,
+      "step": 775
+    },
+    {
+      "epoch": 2.160887656033287,
+      "grad_norm": 1.1296517848968506,
+      "learning_rate": 1.957245533535701e-05,
+      "loss": 0.2856,
+      "step": 780
+    },
+    {
+      "epoch": 2.174757281553398,
+      "grad_norm": 1.148427963256836,
+      "learning_rate": 1.9441340104596808e-05,
+      "loss": 0.2616,
+      "step": 785
+    },
+    {
+      "epoch": 2.188626907073509,
+      "grad_norm": 1.1418274641036987,
+      "learning_rate": 1.9309851856419403e-05,
+      "loss": 0.2816,
+      "step": 790
+    },
+    {
+      "epoch": 2.20249653259362,
+      "grad_norm": 1.2052356004714966,
+      "learning_rate": 1.9178001634204858e-05,
+      "loss": 0.2679,
+      "step": 795
+    },
+    {
+      "epoch": 2.216366158113731,
+      "grad_norm": 1.0612205266952515,
+      "learning_rate": 1.904580051173454e-05,
+      "loss": 0.2781,
+      "step": 800
+    },
+    {
+      "epoch": 2.230235783633842,
+      "grad_norm": 1.198889136314392,
+      "learning_rate": 1.8913259592261105e-05,
+      "loss": 0.2583,
+      "step": 805
+    },
+    {
+      "epoch": 2.2441054091539527,
+      "grad_norm": 1.1233667135238647,
+      "learning_rate": 1.878039000757591e-05,
+      "loss": 0.2989,
+      "step": 810
+    },
+    {
+      "epoch": 2.2579750346740637,
+      "grad_norm": 1.131326675415039,
+      "learning_rate": 1.8647202917074138e-05,
+      "loss": 0.2489,
+      "step": 815
+    },
+    {
+      "epoch": 2.2718446601941746,
+      "grad_norm": 1.171873927116394,
+      "learning_rate": 1.85137095068175e-05,
+      "loss": 0.3085,
+      "step": 820
+    },
+    {
+      "epoch": 2.2857142857142856,
+      "grad_norm": 1.1070297956466675,
+      "learning_rate": 1.8379920988594756e-05,
+      "loss": 0.2358,
+      "step": 825
+    },
+    {
+      "epoch": 2.2995839112343965,
+      "grad_norm": 1.2584491968154907,
+      "learning_rate": 1.824584859898009e-05,
+      "loss": 0.2459,
+      "step": 830
+    },
+    {
+      "epoch": 2.3134535367545075,
+      "grad_norm": 1.1663209199905396,
+      "learning_rate": 1.811150359838934e-05,
+      "loss": 0.2666,
+      "step": 835
+    },
+    {
+      "epoch": 2.3273231622746184,
+      "grad_norm": 1.0747770071029663,
+      "learning_rate": 1.79768972701343e-05,
+      "loss": 0.2603,
+      "step": 840
+    },
+    {
+      "epoch": 2.3411927877947294,
+      "grad_norm": 1.1685748100280762,
+      "learning_rate": 1.784204091947502e-05,
+      "loss": 0.2675,
+      "step": 845
+    },
+    {
+      "epoch": 2.3550624133148403,
+      "grad_norm": 1.055817723274231,
+      "learning_rate": 1.770694587267034e-05,
+      "loss": 0.2473,
+      "step": 850
+    },
+    {
+      "epoch": 2.3689320388349513,
+      "grad_norm": 1.434633493423462,
+      "learning_rate": 1.7571623476026624e-05,
+      "loss": 0.2786,
+      "step": 855
+    },
+    {
+      "epoch": 2.3828016643550622,
+      "grad_norm": 1.1234263181686401,
+      "learning_rate": 1.743608509494477e-05,
+      "loss": 0.2524,
+      "step": 860
+    },
+    {
+      "epoch": 2.396671289875173,
+      "grad_norm": 1.1498136520385742,
+      "learning_rate": 1.7300342112965717e-05,
+      "loss": 0.2738,
+      "step": 865
+    },
+    {
+      "epoch": 2.410540915395284,
+      "grad_norm": 1.1265621185302734,
+      "learning_rate": 1.7164405930814313e-05,
+      "loss": 0.2519,
+      "step": 870
+    },
+    {
+      "epoch": 2.424410540915395,
+      "grad_norm": 1.1003201007843018,
+      "learning_rate": 1.7028287965441843e-05,
+      "loss": 0.2923,
+      "step": 875
+    },
+    {
+      "epoch": 2.438280166435506,
+      "grad_norm": 1.170314908027649,
+      "learning_rate": 1.6891999649067105e-05,
+      "loss": 0.2434,
+      "step": 880
+    },
+    {
+      "epoch": 2.452149791955617,
+      "grad_norm": 1.3272323608398438,
+      "learning_rate": 1.6755552428216278e-05,
+      "loss": 0.2573,
+      "step": 885
+    },
+    {
+      "epoch": 2.466019417475728,
+      "grad_norm": 1.2042258977890015,
+      "learning_rate": 1.6618957762761555e-05,
+      "loss": 0.2406,
+      "step": 890
+    },
+    {
+      "epoch": 2.479889042995839,
+      "grad_norm": 1.2370964288711548,
+      "learning_rate": 1.6482227124958617e-05,
+      "loss": 0.2438,
+      "step": 895
+    },
+    {
+      "epoch": 2.49375866851595,
+      "grad_norm": 1.3068962097167969,
+      "learning_rate": 1.6345371998483155e-05,
+      "loss": 0.2423,
+      "step": 900
+    },
+    {
+      "epoch": 2.507628294036061,
+      "grad_norm": 1.1322987079620361,
+      "learning_rate": 1.6208403877466354e-05,
+      "loss": 0.2355,
+      "step": 905
+    },
+    {
+      "epoch": 2.5214979195561718,
+      "grad_norm": 1.19632887840271,
+      "learning_rate": 1.6071334265529538e-05,
+      "loss": 0.2376,
+      "step": 910
+    },
+    {
+      "epoch": 2.5353675450762827,
+      "grad_norm": 1.1878324747085571,
+      "learning_rate": 1.5934174674818013e-05,
+      "loss": 0.2495,
+      "step": 915
+    },
+    {
+      "epoch": 2.5492371705963937,
+      "grad_norm": 1.1836864948272705,
+      "learning_rate": 1.5796936625034177e-05,
+      "loss": 0.2526,
+      "step": 920
+    },
+    {
+      "epoch": 2.5631067961165046,
+      "grad_norm": 1.1582130193710327,
+      "learning_rate": 1.5659631642470025e-05,
+      "loss": 0.2388,
+      "step": 925
+    },
+    {
+      "epoch": 2.5769764216366156,
+      "grad_norm": 1.140184760093689,
+      "learning_rate": 1.5522271259039063e-05,
+      "loss": 0.2468,
+      "step": 930
+    },
+    {
+      "epoch": 2.5908460471567265,
+      "grad_norm": 1.086246132850647,
+      "learning_rate": 1.5384867011307797e-05,
+      "loss": 0.2685,
+      "step": 935
+    },
+    {
+      "epoch": 2.6047156726768375,
+      "grad_norm": 1.1802036762237549,
+      "learning_rate": 1.5247430439526771e-05,
+      "loss": 0.2597,
+      "step": 940
+    },
+    {
+      "epoch": 2.6185852981969484,
+      "grad_norm": 1.1825424432754517,
+      "learning_rate": 1.5109973086661361e-05,
+      "loss": 0.2185,
+      "step": 945
+    },
+    {
+      "epoch": 2.63245492371706,
+      "grad_norm": 1.2922337055206299,
+      "learning_rate": 1.4972506497422298e-05,
+      "loss": 0.2154,
+      "step": 950
+    },
+    {
+      "epoch": 2.6463245492371708,
+      "grad_norm": 1.451962947845459,
+      "learning_rate": 1.4835042217296032e-05,
+      "loss": 0.2287,
+      "step": 955
+    },
+    {
+      "epoch": 2.6601941747572817,
+      "grad_norm": 1.0612127780914307,
+      "learning_rate": 1.4697591791575115e-05,
+      "loss": 0.2298,
+      "step": 960
+    },
+    {
+      "epoch": 2.6740638002773927,
+      "grad_norm": 1.1951512098312378,
+      "learning_rate": 1.4560166764388465e-05,
+      "loss": 0.2392,
+      "step": 965
+    },
+    {
+      "epoch": 2.6879334257975036,
+      "grad_norm": 1.1254812479019165,
+      "learning_rate": 1.442277867773185e-05,
+      "loss": 0.211,
+      "step": 970
+    },
+    {
+      "epoch": 2.7018030513176146,
+      "grad_norm": 1.0225093364715576,
+      "learning_rate": 1.4285439070498504e-05,
+      "loss": 0.2218,
+      "step": 975
+    },
+    {
+      "epoch": 2.7156726768377255,
+      "grad_norm": 1.0218101739883423,
+      "learning_rate": 1.4148159477509968e-05,
+      "loss": 0.2297,
+      "step": 980
+    },
+    {
+      "epoch": 2.7295423023578365,
+      "grad_norm": 1.3001481294631958,
+      "learning_rate": 1.4010951428547356e-05,
+      "loss": 0.1802,
+      "step": 985
+    },
+    {
+      "epoch": 2.7434119278779474,
+      "grad_norm": 1.085077166557312,
+      "learning_rate": 1.3873826447382957e-05,
+      "loss": 0.2022,
+      "step": 990
+    },
+    {
+      "epoch": 2.7572815533980584,
+      "grad_norm": 1.1113858222961426,
+      "learning_rate": 1.3736796050812406e-05,
+      "loss": 0.2163,
+      "step": 995
+    },
+    {
+      "epoch": 2.7711511789181693,
+      "grad_norm": 1.412807822227478,
+      "learning_rate": 1.3599871747687394e-05,
+      "loss": 0.2158,
+      "step": 1000
+    },
+    {
+      "epoch": 2.7850208044382803,
+      "grad_norm": 5.181802749633789,
+      "learning_rate": 1.346306503794909e-05,
+      "loss": 0.2101,
+      "step": 1005
+    },
+    {
+      "epoch": 2.798890429958391,
+      "grad_norm": 1.1314557790756226,
+      "learning_rate": 1.3326387411662279e-05,
+      "loss": 0.1977,
+      "step": 1010
+    },
+    {
+      "epoch": 2.812760055478502,
+      "grad_norm": 1.1890625953674316,
+      "learning_rate": 1.3189850348050336e-05,
+      "loss": 0.2135,
+      "step": 1015
+    },
+    {
+      "epoch": 2.826629680998613,
+      "grad_norm": 1.2581861019134521,
+      "learning_rate": 1.3053465314531118e-05,
+      "loss": 0.2091,
+      "step": 1020
+    },
+    {
+      "epoch": 2.840499306518724,
+      "grad_norm": 1.0491549968719482,
+      "learning_rate": 1.2917243765753836e-05,
+      "loss": 0.2041,
+      "step": 1025
+    },
+    {
+      "epoch": 2.854368932038835,
+      "grad_norm": 1.208863615989685,
+      "learning_rate": 1.2781197142637018e-05,
+      "loss": 0.1964,
+      "step": 1030
+    },
+    {
+      "epoch": 2.868238557558946,
+      "grad_norm": 1.2405157089233398,
+      "learning_rate": 1.2645336871407615e-05,
+      "loss": 0.2303,
+      "step": 1035
+    },
+    {
+      "epoch": 2.882108183079057,
+      "grad_norm": 1.1387736797332764,
+      "learning_rate": 1.2509674362641327e-05,
+      "loss": 0.1905,
+      "step": 1040
+    },
+    {
+      "epoch": 2.895977808599168,
+      "grad_norm": 1.1768152713775635,
+      "learning_rate": 1.237422101030426e-05,
+      "loss": 0.1963,
+      "step": 1045
+    },
+    {
+      "epoch": 2.909847434119279,
+      "grad_norm": 1.097781777381897,
+      "learning_rate": 1.223898819079598e-05,
+      "loss": 0.1991,
+      "step": 1050
+    },
+    {
+      "epoch": 2.9237170596393898,
+      "grad_norm": 1.2869900465011597,
+      "learning_rate": 1.2103987261994029e-05,
+      "loss": 0.1915,
+      "step": 1055
+    },
+    {
+      "epoch": 2.9375866851595007,
+      "grad_norm": 1.0657851696014404,
+      "learning_rate": 1.1969229562300025e-05,
+      "loss": 0.1841,
+      "step": 1060
+    },
+    {
+      "epoch": 2.9514563106796117,
+      "grad_norm": 1.0998252630233765,
+      "learning_rate": 1.1834726409687351e-05,
+      "loss": 0.1834,
+      "step": 1065
+    },
+    {
+      "epoch": 2.9653259361997226,
+      "grad_norm": 1.1832915544509888,
+      "learning_rate": 1.1700489100750585e-05,
+      "loss": 0.1848,
+      "step": 1070
+    },
+    {
+      "epoch": 2.9791955617198336,
+      "grad_norm": 1.1553837060928345,
+      "learning_rate": 1.1566528909756757e-05,
+      "loss": 0.1986,
+      "step": 1075
+    },
+    {
+      "epoch": 2.9930651872399445,
+      "grad_norm": 1.2271969318389893,
+      "learning_rate": 1.1432857087698414e-05,
+      "loss": 0.1971,
+      "step": 1080
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1805,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 2000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.5697919712228803e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

23_128_e5_3e-5/checkpoint-1083/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ebded260fea4e3a399f7970554139fd5ed218599097aac84526fbe4423490a7
+size 7736

23_128_e5_3e-5/checkpoint-1083/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)

23_128_e5_3e-5/checkpoint-1444/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

23_128_e5_3e-5/checkpoint-1444/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": [
+    "up_proj",
+    "q_proj",
+    "down_proj",
+    "v_proj",
+    "gate_proj",
+    "o_proj",
+    "k_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

23_128_e5_3e-5/checkpoint-1444/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f333a6eeab282409b134357f008783a72bdf18bd93ae1872dee2370f05afd259
+size 791751704

23_128_e5_3e-5/checkpoint-1444/latest

@@ -0,0 +1 @@
+global_step1444

23_128_e5_3e-5/checkpoint-1444/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74cb952aa009793c92237b4738723aa154261ad54ee6122cbc24a5bc7fd3fee7
+size 15920

23_128_e5_3e-5/checkpoint-1444/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d95b55a8c7435c17031db68b48c92c139c799d5711d31dc31563ebfb0a7cbe72
+size 15920

23_128_e5_3e-5/checkpoint-1444/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b93aaf4c759d13e6da47d274a4ef9b1677d038c9ba11556a6dae30b4c33919f
+size 15920

23_128_e5_3e-5/checkpoint-1444/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab38597bbf10965a50d52971a0af71e27c2bd06bbd975d523b1102124aa2a9dc
+size 15920

23_128_e5_3e-5/checkpoint-1444/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:572e1d914fcf96ea8143b91ff6868f9bd09e7b1ac01050beea363a4c6c932b15
+size 15920

23_128_e5_3e-5/checkpoint-1444/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0efded27eea23ef34be51271b93b034c1b19def7de59256b9eb6ae5a7e8c72a2
+size 15920

23_128_e5_3e-5/checkpoint-1444/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2407f54c284bffe6a4de7f2b67e469dc97742640f1f476d114db5c8842942d34
+size 15920

23_128_e5_3e-5/checkpoint-1444/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2537c76d4a7cdbc3174ee44a9304c8e125357738226b9d7a0fabd6828d12f1f7
+size 15920

23_128_e5_3e-5/checkpoint-1444/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0da2d793a3e9f3f0fa68631f126f20f1007707cd681a8b69a5a6a0ac87650a7
+size 1064

23_128_e5_3e-5/checkpoint-1444/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
+  }
+}

23_128_e5_3e-5/checkpoint-1444/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
+}

23_128_e5_3e-5/checkpoint-1444/trainer_state.json

@@ -0,0 +1,2050 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1444,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.013869625520110958,
+      "grad_norm": 1.2640684843063354,
+      "learning_rate": 1.3186813186813187e-06,
+      "loss": 1.3135,
+      "step": 5
+    },
+    {
+      "epoch": 0.027739251040221916,
+      "grad_norm": 1.024712085723877,
+      "learning_rate": 2.967032967032967e-06,
+      "loss": 1.3432,
+      "step": 10
+    },
+    {
+      "epoch": 0.04160887656033287,
+      "grad_norm": 0.7255855798721313,
+      "learning_rate": 4.615384615384616e-06,
+      "loss": 1.3293,
+      "step": 15
+    },
+    {
+      "epoch": 0.05547850208044383,
+      "grad_norm": 0.5718052983283997,
+      "learning_rate": 6.2637362637362645e-06,
+      "loss": 1.3092,
+      "step": 20
+    },
+    {
+      "epoch": 0.06934812760055478,
+      "grad_norm": 0.6330448985099792,
+      "learning_rate": 7.912087912087913e-06,
+      "loss": 1.3039,
+      "step": 25
+    },
+    {
+      "epoch": 0.08321775312066575,
+      "grad_norm": 0.7984022498130798,
+      "learning_rate": 9.56043956043956e-06,
+      "loss": 1.2686,
+      "step": 30
+    },
+    {
+      "epoch": 0.0970873786407767,
+      "grad_norm": 0.4936966300010681,
+      "learning_rate": 1.120879120879121e-05,
+      "loss": 1.2573,
+      "step": 35
+    },
+    {
+      "epoch": 0.11095700416088766,
+      "grad_norm": 0.5313364863395691,
+      "learning_rate": 1.2857142857142857e-05,
+      "loss": 1.2233,
+      "step": 40
+    },
+    {
+      "epoch": 0.12482662968099861,
+      "grad_norm": 0.4598558843135834,
+      "learning_rate": 1.4505494505494506e-05,
+      "loss": 1.2457,
+      "step": 45
+    },
+    {
+      "epoch": 0.13869625520110956,
+      "grad_norm": 0.4705202877521515,
+      "learning_rate": 1.6153846153846154e-05,
+      "loss": 1.2183,
+      "step": 50
+    },
+    {
+      "epoch": 0.15256588072122051,
+      "grad_norm": 0.6102734208106995,
+      "learning_rate": 1.78021978021978e-05,
+      "loss": 1.1374,
+      "step": 55
+    },
+    {
+      "epoch": 0.1664355062413315,
+      "grad_norm": 0.644282877445221,
+      "learning_rate": 1.9450549450549452e-05,
+      "loss": 1.1574,
+      "step": 60
+    },
+    {
+      "epoch": 0.18030513176144244,
+      "grad_norm": 0.5232361555099487,
+      "learning_rate": 2.10989010989011e-05,
+      "loss": 1.1972,
+      "step": 65
+    },
+    {
+      "epoch": 0.1941747572815534,
+      "grad_norm": 0.5031052827835083,
+      "learning_rate": 2.2747252747252748e-05,
+      "loss": 1.2067,
+      "step": 70
+    },
+    {
+      "epoch": 0.20804438280166435,
+      "grad_norm": 0.5804685950279236,
+      "learning_rate": 2.4395604395604395e-05,
+      "loss": 1.1819,
+      "step": 75
+    },
+    {
+      "epoch": 0.22191400832177532,
+      "grad_norm": 0.4593973159790039,
+      "learning_rate": 2.6043956043956046e-05,
+      "loss": 1.1541,
+      "step": 80
+    },
+    {
+      "epoch": 0.23578363384188628,
+      "grad_norm": 0.4562990367412567,
+      "learning_rate": 2.7692307692307694e-05,
+      "loss": 1.1592,
+      "step": 85
+    },
+    {
+      "epoch": 0.24965325936199723,
+      "grad_norm": 0.5135698318481445,
+      "learning_rate": 2.934065934065934e-05,
+      "loss": 1.1133,
+      "step": 90
+    },
+    {
+      "epoch": 0.2635228848821082,
+      "grad_norm": 0.5088622570037842,
+      "learning_rate": 2.9999773232512234e-05,
+      "loss": 1.121,
+      "step": 95
+    },
+    {
+      "epoch": 0.27739251040221913,
+      "grad_norm": 0.5037941336631775,
+      "learning_rate": 2.9998387456028022e-05,
+      "loss": 1.0824,
+      "step": 100
+    },
+    {
+      "epoch": 0.2912621359223301,
+      "grad_norm": 0.6273325085639954,
+      "learning_rate": 2.999574200124419e-05,
+      "loss": 1.1311,
+      "step": 105
+    },
+    {
+      "epoch": 0.30513176144244103,
+      "grad_norm": 0.6029002070426941,
+      "learning_rate": 2.999183709034608e-05,
+      "loss": 1.061,
+      "step": 110
+    },
+    {
+      "epoch": 0.31900138696255204,
+      "grad_norm": 0.49770888686180115,
+      "learning_rate": 2.998667305129772e-05,
+      "loss": 1.0573,
+      "step": 115
+    },
+    {
+      "epoch": 0.332871012482663,
+      "grad_norm": 0.6382117867469788,
+      "learning_rate": 2.9980250317814265e-05,
+      "loss": 1.0497,
+      "step": 120
+    },
+    {
+      "epoch": 0.34674063800277394,
+      "grad_norm": 0.5979148149490356,
+      "learning_rate": 2.9972569429325575e-05,
+      "loss": 1.0705,
+      "step": 125
+    },
+    {
+      "epoch": 0.3606102635228849,
+      "grad_norm": 0.5902369618415833,
+      "learning_rate": 2.99636310309309e-05,
+      "loss": 1.0601,
+      "step": 130
+    },
+    {
+      "epoch": 0.37447988904299584,
+      "grad_norm": 0.5844177007675171,
+      "learning_rate": 2.995343587334471e-05,
+      "loss": 0.9897,
+      "step": 135
+    },
+    {
+      "epoch": 0.3883495145631068,
+      "grad_norm": 0.5972667336463928,
+      "learning_rate": 2.994198481283364e-05,
+      "loss": 1.0212,
+      "step": 140
+    },
+    {
+      "epoch": 0.40221914008321774,
+      "grad_norm": 0.6355414390563965,
+      "learning_rate": 2.992927881114458e-05,
+      "loss": 1.0097,
+      "step": 145
+    },
+    {
+      "epoch": 0.4160887656033287,
+      "grad_norm": 0.6107580065727234,
+      "learning_rate": 2.991531893542389e-05,
+      "loss": 0.9994,
+      "step": 150
+    },
+    {
+      "epoch": 0.42995839112343964,
+      "grad_norm": 0.7091370224952698,
+      "learning_rate": 2.990010635812777e-05,
+      "loss": 0.9766,
+      "step": 155
+    },
+    {
+      "epoch": 0.44382801664355065,
+      "grad_norm": 0.6339324116706848,
+      "learning_rate": 2.9883642356923822e-05,
+      "loss": 1.0056,
+      "step": 160
+    },
+    {
+      "epoch": 0.4576976421636616,
+      "grad_norm": 0.6281903386116028,
+      "learning_rate": 2.986592831458369e-05,
+      "loss": 0.9624,
+      "step": 165
+    },
+    {
+      "epoch": 0.47156726768377255,
+      "grad_norm": 0.7146309018135071,
+      "learning_rate": 2.984696571886697e-05,
+      "loss": 0.9955,
+      "step": 170
+    },
+    {
+      "epoch": 0.4854368932038835,
+      "grad_norm": 0.852715015411377,
+      "learning_rate": 2.9826756162396226e-05,
+      "loss": 0.9626,
+      "step": 175
+    },
+    {
+      "epoch": 0.49930651872399445,
+      "grad_norm": 0.7176477909088135,
+      "learning_rate": 2.9805301342523238e-05,
+      "loss": 0.9451,
+      "step": 180
+    },
+    {
+      "epoch": 0.5131761442441054,
+      "grad_norm": 0.6825271248817444,
+      "learning_rate": 2.9782603061186458e-05,
+      "loss": 0.9969,
+      "step": 185
+    },
+    {
+      "epoch": 0.5270457697642164,
+      "grad_norm": 0.7406214475631714,
+      "learning_rate": 2.975866322475966e-05,
+      "loss": 0.945,
+      "step": 190
+    },
+    {
+      "epoch": 0.5409153952843273,
+      "grad_norm": 0.8366868495941162,
+      "learning_rate": 2.973348384389182e-05,
+      "loss": 0.9201,
+      "step": 195
+    },
+    {
+      "epoch": 0.5547850208044383,
+      "grad_norm": 0.8170830011367798,
+      "learning_rate": 2.9707067033338257e-05,
+      "loss": 0.9614,
+      "step": 200
+    },
+    {
+      "epoch": 0.5686546463245492,
+      "grad_norm": 0.7892953157424927,
+      "learning_rate": 2.967941501178302e-05,
+      "loss": 0.9041,
+      "step": 205
+    },
+    {
+      "epoch": 0.5825242718446602,
+      "grad_norm": 0.874885082244873,
+      "learning_rate": 2.9650530101652548e-05,
+      "loss": 0.9418,
+      "step": 210
+    },
+    {
+      "epoch": 0.5963938973647711,
+      "grad_norm": 0.8003681302070618,
+      "learning_rate": 2.9620414728920603e-05,
+      "loss": 0.8867,
+      "step": 215
+    },
+    {
+      "epoch": 0.6102635228848821,
+      "grad_norm": 0.9193400740623474,
+      "learning_rate": 2.9589071422904525e-05,
+      "loss": 0.9073,
+      "step": 220
+    },
+    {
+      "epoch": 0.624133148404993,
+      "grad_norm": 0.8814607858657837,
+      "learning_rate": 2.9556502816052805e-05,
+      "loss": 0.8449,
+      "step": 225
+    },
+    {
+      "epoch": 0.6380027739251041,
+      "grad_norm": 0.8270254135131836,
+      "learning_rate": 2.9522711643723997e-05,
+      "loss": 0.8744,
+      "step": 230
+    },
+    {
+      "epoch": 0.651872399445215,
+      "grad_norm": 0.8522059321403503,
+      "learning_rate": 2.9487700743956955e-05,
+      "loss": 0.8211,
+      "step": 235
+    },
+    {
+      "epoch": 0.665742024965326,
+      "grad_norm": 0.8157532215118408,
+      "learning_rate": 2.945147305723251e-05,
+      "loss": 0.8436,
+      "step": 240
+    },
+    {
+      "epoch": 0.6796116504854369,
+      "grad_norm": 0.9486557841300964,
+      "learning_rate": 2.9414031626226472e-05,
+      "loss": 0.7924,
+      "step": 245
+    },
+    {
+      "epoch": 0.6934812760055479,
+      "grad_norm": 0.7984339594841003,
+      "learning_rate": 2.9375379595554113e-05,
+      "loss": 0.8411,
+      "step": 250
+    },
+    {
+      "epoch": 0.7073509015256588,
+      "grad_norm": 0.9337722659111023,
+      "learning_rate": 2.9335520211506033e-05,
+      "loss": 0.8287,
+      "step": 255
+    },
+    {
+      "epoch": 0.7212205270457698,
+      "grad_norm": 0.8872073292732239,
+      "learning_rate": 2.9294456821775524e-05,
+      "loss": 0.794,
+      "step": 260
+    },
+    {
+      "epoch": 0.7350901525658807,
+      "grad_norm": 0.951157808303833,
+      "learning_rate": 2.9252192875177415e-05,
+      "loss": 0.8033,
+      "step": 265
+    },
+    {
+      "epoch": 0.7489597780859917,
+      "grad_norm": 0.9538988471031189,
+      "learning_rate": 2.920873192135839e-05,
+      "loss": 0.7984,
+      "step": 270
+    },
+    {
+      "epoch": 0.7628294036061026,
+      "grad_norm": 0.9138048887252808,
+      "learning_rate": 2.9164077610498886e-05,
+      "loss": 0.8138,
+      "step": 275
+    },
+    {
+      "epoch": 0.7766990291262136,
+      "grad_norm": 0.8404424786567688,
+      "learning_rate": 2.9118233693006503e-05,
+      "loss": 0.7732,
+      "step": 280
+    },
+    {
+      "epoch": 0.7905686546463245,
+      "grad_norm": 0.8813743591308594,
+      "learning_rate": 2.907120401920103e-05,
+      "loss": 0.7616,
+      "step": 285
+    },
+    {
+      "epoch": 0.8044382801664355,
+      "grad_norm": 0.9614547491073608,
+      "learning_rate": 2.9022992538991067e-05,
+      "loss": 0.8015,
+      "step": 290
+    },
+    {
+      "epoch": 0.8183079056865464,
+      "grad_norm": 0.8543416857719421,
+      "learning_rate": 2.897360330154227e-05,
+      "loss": 0.7915,
+      "step": 295
+    },
+    {
+      "epoch": 0.8321775312066574,
+      "grad_norm": 0.9641604423522949,
+      "learning_rate": 2.892304045493728e-05,
+      "loss": 0.708,
+      "step": 300
+    },
+    {
+      "epoch": 0.8460471567267683,
+      "grad_norm": 0.9229587912559509,
+      "learning_rate": 2.8871308245827336e-05,
+      "loss": 0.732,
+      "step": 305
+    },
+    {
+      "epoch": 0.8599167822468793,
+      "grad_norm": 0.9859731197357178,
+      "learning_rate": 2.88184110190756e-05,
+      "loss": 0.7759,
+      "step": 310
+    },
+    {
+      "epoch": 0.8737864077669902,
+      "grad_norm": 1.0025079250335693,
+      "learning_rate": 2.8764353217392253e-05,
+      "loss": 0.7113,
+      "step": 315
+    },
+    {
+      "epoch": 0.8876560332871013,
+      "grad_norm": 0.971370279788971,
+      "learning_rate": 2.870913938096136e-05,
+      "loss": 0.738,
+      "step": 320
+    },
+    {
+      "epoch": 0.9015256588072122,
+      "grad_norm": 1.0351135730743408,
+      "learning_rate": 2.865277414705955e-05,
+      "loss": 0.7727,
+      "step": 325
+    },
+    {
+      "epoch": 0.9153952843273232,
+      "grad_norm": 0.9907004833221436,
+      "learning_rate": 2.8595262249666536e-05,
+      "loss": 0.7487,
+      "step": 330
+    },
+    {
+      "epoch": 0.9292649098474342,
+      "grad_norm": 1.1555078029632568,
+      "learning_rate": 2.8536608519067532e-05,
+      "loss": 0.7369,
+      "step": 335
+    },
+    {
+      "epoch": 0.9431345353675451,
+      "grad_norm": 0.9639641642570496,
+      "learning_rate": 2.8476817881447555e-05,
+      "loss": 0.7045,
+      "step": 340
+    },
+    {
+      "epoch": 0.957004160887656,
+      "grad_norm": 0.9972969889640808,
+      "learning_rate": 2.8415895358477702e-05,
+      "loss": 0.6953,
+      "step": 345
+    },
+    {
+      "epoch": 0.970873786407767,
+      "grad_norm": 1.2426544427871704,
+      "learning_rate": 2.8353846066893382e-05,
+      "loss": 0.6732,
+      "step": 350
+    },
+    {
+      "epoch": 0.984743411927878,
+      "grad_norm": 1.1239923238754272,
+      "learning_rate": 2.8290675218064565e-05,
+      "loss": 0.6729,
+      "step": 355
+    },
+    {
+      "epoch": 0.9986130374479889,
+      "grad_norm": 1.099241852760315,
+      "learning_rate": 2.822638811755812e-05,
+      "loss": 0.6424,
+      "step": 360
+    },
+    {
+      "epoch": 1.0110957004160888,
+      "grad_norm": 0.9136450290679932,
+      "learning_rate": 2.8160990164692183e-05,
+      "loss": 0.5966,
+      "step": 365
+    },
+    {
+      "epoch": 1.0249653259361997,
+      "grad_norm": 1.128823161125183,
+      "learning_rate": 2.8094486852082692e-05,
+      "loss": 0.5859,
+      "step": 370
+    },
+    {
+      "epoch": 1.0388349514563107,
+      "grad_norm": 1.0320087671279907,
+      "learning_rate": 2.802688376518209e-05,
+      "loss": 0.6032,
+      "step": 375
+    },
+    {
+      "epoch": 1.0527045769764216,
+      "grad_norm": 0.9852442741394043,
+      "learning_rate": 2.7958186581810195e-05,
+      "loss": 0.5763,
+      "step": 380
+    },
+    {
+      "epoch": 1.0665742024965326,
+      "grad_norm": 1.0142955780029297,
+      "learning_rate": 2.7888401071677342e-05,
+      "loss": 0.615,
+      "step": 385
+    },
+    {
+      "epoch": 1.0804438280166435,
+      "grad_norm": 1.1618216037750244,
+      "learning_rate": 2.7817533095899806e-05,
+      "loss": 0.6023,
+      "step": 390
+    },
+    {
+      "epoch": 1.0943134535367545,
+      "grad_norm": 1.001564621925354,
+      "learning_rate": 2.774558860650752e-05,
+      "loss": 0.5843,
+      "step": 395
+    },
+    {
+      "epoch": 1.1081830790568654,
+      "grad_norm": 1.1653980016708374,
+      "learning_rate": 2.767257364594421e-05,
+      "loss": 0.5763,
+      "step": 400
+    },
+    {
+      "epoch": 1.1220527045769764,
+      "grad_norm": 1.349169373512268,
+      "learning_rate": 2.759849434655987e-05,
+      "loss": 0.6157,
+      "step": 405
+    },
+    {
+      "epoch": 1.1359223300970873,
+      "grad_norm": 1.1331448554992676,
+      "learning_rate": 2.7523356930095764e-05,
+      "loss": 0.5724,
+      "step": 410
+    },
+    {
+      "epoch": 1.1497919556171983,
+      "grad_norm": 1.133574366569519,
+      "learning_rate": 2.744716770716182e-05,
+      "loss": 0.5866,
+      "step": 415
+    },
+    {
+      "epoch": 1.1636615811373092,
+      "grad_norm": 1.164962649345398,
+      "learning_rate": 2.736993307670667e-05,
+      "loss": 0.5861,
+      "step": 420
+    },
+    {
+      "epoch": 1.1775312066574202,
+      "grad_norm": 1.0958647727966309,
+      "learning_rate": 2.7291659525480192e-05,
+      "loss": 0.5564,
+      "step": 425
+    },
+    {
+      "epoch": 1.1914008321775311,
+      "grad_norm": 1.6513663530349731,
+      "learning_rate": 2.72123536274887e-05,
+      "loss": 0.5444,
+      "step": 430
+    },
+    {
+      "epoch": 1.205270457697642,
+      "grad_norm": 1.0394660234451294,
+      "learning_rate": 2.713202204344282e-05,
+      "loss": 0.5366,
+      "step": 435
+    },
+    {
+      "epoch": 1.219140083217753,
+      "grad_norm": 1.290332555770874,
+      "learning_rate": 2.705067152019808e-05,
+      "loss": 0.5414,
+      "step": 440
+    },
+    {
+      "epoch": 1.233009708737864,
+      "grad_norm": 1.1973203420639038,
+      "learning_rate": 2.6968308890188235e-05,
+      "loss": 0.5563,
+      "step": 445
+    },
+    {
+      "epoch": 1.246879334257975,
+      "grad_norm": 1.1321580410003662,
+      "learning_rate": 2.6884941070851443e-05,
+      "loss": 0.5668,
+      "step": 450
+    },
+    {
+      "epoch": 1.2607489597780859,
+      "grad_norm": 1.1653034687042236,
+      "learning_rate": 2.680057506404929e-05,
+      "loss": 0.5898,
+      "step": 455
+    },
+    {
+      "epoch": 1.2746185852981968,
+      "grad_norm": 1.271990180015564,
+      "learning_rate": 2.6715217955478705e-05,
+      "loss": 0.4863,
+      "step": 460
+    },
+    {
+      "epoch": 1.2884882108183078,
+      "grad_norm": 1.083730697631836,
+      "learning_rate": 2.6628876914076865e-05,
+      "loss": 0.496,
+      "step": 465
+    },
+    {
+      "epoch": 1.3023578363384187,
+      "grad_norm": 1.3474218845367432,
+      "learning_rate": 2.654155919141908e-05,
+      "loss": 0.5484,
+      "step": 470
+    },
+    {
+      "epoch": 1.31622746185853,
+      "grad_norm": 1.100876808166504,
+      "learning_rate": 2.645327212110976e-05,
+      "loss": 0.4999,
+      "step": 475
+    },
+    {
+      "epoch": 1.3300970873786409,
+      "grad_norm": 1.1323593854904175,
+      "learning_rate": 2.636402311816649e-05,
+      "loss": 0.5368,
+      "step": 480
+    },
+    {
+      "epoch": 1.3439667128987518,
+      "grad_norm": 1.2535606622695923,
+      "learning_rate": 2.6273819678397233e-05,
+      "loss": 0.5537,
+      "step": 485
+    },
+    {
+      "epoch": 1.3578363384188628,
+      "grad_norm": 1.1354525089263916,
+      "learning_rate": 2.6182669377770807e-05,
+      "loss": 0.4727,
+      "step": 490
+    },
+    {
+      "epoch": 1.3717059639389737,
+      "grad_norm": 1.2694000005722046,
+      "learning_rate": 2.6090579871780583e-05,
+      "loss": 0.5171,
+      "step": 495
+    },
+    {
+      "epoch": 1.3855755894590847,
+      "grad_norm": 1.2506495714187622,
+      "learning_rate": 2.5997558894801514e-05,
+      "loss": 0.472,
+      "step": 500
+    },
+    {
+      "epoch": 1.3994452149791956,
+      "grad_norm": 1.1564306020736694,
+      "learning_rate": 2.5903614259440553e-05,
+      "loss": 0.5141,
+      "step": 505
+    },
+    {
+      "epoch": 1.4133148404993066,
+      "grad_norm": 1.1961454153060913,
+      "learning_rate": 2.580875385588048e-05,
+      "loss": 0.4358,
+      "step": 510
+    },
+    {
+      "epoch": 1.4271844660194175,
+      "grad_norm": 1.0629825592041016,
+      "learning_rate": 2.571298565121725e-05,
+      "loss": 0.4934,
+      "step": 515
+    },
+    {
+      "epoch": 1.4410540915395285,
+      "grad_norm": 1.1157474517822266,
+      "learning_rate": 2.561631768879082e-05,
+      "loss": 0.4666,
+      "step": 520
+    },
+    {
+      "epoch": 1.4549237170596394,
+      "grad_norm": 1.1023491621017456,
+      "learning_rate": 2.551875808750963e-05,
+      "loss": 0.4786,
+      "step": 525
+    },
+    {
+      "epoch": 1.4687933425797504,
+      "grad_norm": 1.3004133701324463,
+      "learning_rate": 2.5420315041168717e-05,
+      "loss": 0.5161,
+      "step": 530
+    },
+    {
+      "epoch": 1.4826629680998613,
+      "grad_norm": 1.2039952278137207,
+      "learning_rate": 2.5320996817761534e-05,
+      "loss": 0.4471,
+      "step": 535
+    },
+    {
+      "epoch": 1.4965325936199723,
+      "grad_norm": 1.3886206150054932,
+      "learning_rate": 2.5220811758785525e-05,
+      "loss": 0.4796,
+      "step": 540
+    },
+    {
+      "epoch": 1.5104022191400832,
+      "grad_norm": 1.13352370262146,
+      "learning_rate": 2.5119768278541576e-05,
+      "loss": 0.4782,
+      "step": 545
+    },
+    {
+      "epoch": 1.5242718446601942,
+      "grad_norm": 1.2641507387161255,
+      "learning_rate": 2.5017874863427292e-05,
+      "loss": 0.4516,
+      "step": 550
+    },
+    {
+      "epoch": 1.5381414701803051,
+      "grad_norm": 1.3237358331680298,
+      "learning_rate": 2.4915140071224247e-05,
+      "loss": 0.451,
+      "step": 555
+    },
+    {
+      "epoch": 1.552011095700416,
+      "grad_norm": 1.1909607648849487,
+      "learning_rate": 2.4811572530379253e-05,
+      "loss": 0.4389,
+      "step": 560
+    },
+    {
+      "epoch": 1.565880721220527,
+      "grad_norm": 1.2734519243240356,
+      "learning_rate": 2.4707180939279658e-05,
+      "loss": 0.4822,
+      "step": 565
+    },
+    {
+      "epoch": 1.579750346740638,
+      "grad_norm": 1.2859396934509277,
+      "learning_rate": 2.46019740655228e-05,
+      "loss": 0.4503,
+      "step": 570
+    },
+    {
+      "epoch": 1.593619972260749,
+      "grad_norm": 1.0753681659698486,
+      "learning_rate": 2.4495960745179646e-05,
+      "loss": 0.4571,
+      "step": 575
+    },
+    {
+      "epoch": 1.6074895977808599,
+      "grad_norm": 1.0502837896347046,
+      "learning_rate": 2.4389149882052654e-05,
+      "loss": 0.4532,
+      "step": 580
+    },
+    {
+      "epoch": 1.6213592233009708,
+      "grad_norm": 1.1627888679504395,
+      "learning_rate": 2.4281550446927968e-05,
+      "loss": 0.4613,
+      "step": 585
+    },
+    {
+      "epoch": 1.635228848821082,
+      "grad_norm": 1.1929353475570679,
+      "learning_rate": 2.4173171476821997e-05,
+      "loss": 0.4411,
+      "step": 590
+    },
+    {
+      "epoch": 1.649098474341193,
+      "grad_norm": 1.1515079736709595,
+      "learning_rate": 2.40640220742224e-05,
+      "loss": 0.427,
+      "step": 595
+    },
+    {
+      "epoch": 1.662968099861304,
+      "grad_norm": 1.0611435174942017,
+      "learning_rate": 2.3954111406323607e-05,
+      "loss": 0.4296,
+      "step": 600
+    },
+    {
+      "epoch": 1.6768377253814148,
+      "grad_norm": 1.2731926441192627,
+      "learning_rate": 2.3843448704256868e-05,
+      "loss": 0.3959,
+      "step": 605
+    },
+    {
+      "epoch": 1.6907073509015258,
+      "grad_norm": 1.104193091392517,
+      "learning_rate": 2.3732043262314974e-05,
+      "loss": 0.4378,
+      "step": 610
+    },
+    {
+      "epoch": 1.7045769764216367,
+      "grad_norm": 1.2957518100738525,
+      "learning_rate": 2.3619904437171642e-05,
+      "loss": 0.4088,
+      "step": 615
+    },
+    {
+      "epoch": 1.7184466019417477,
+      "grad_norm": 1.1752687692642212,
+      "learning_rate": 2.3507041647095668e-05,
+      "loss": 0.4095,
+      "step": 620
+    },
+    {
+      "epoch": 1.7323162274618586,
+      "grad_norm": 1.1501843929290771,
+      "learning_rate": 2.3393464371159914e-05,
+      "loss": 0.39,
+      "step": 625
+    },
+    {
+      "epoch": 1.7461858529819696,
+      "grad_norm": 1.2524619102478027,
+      "learning_rate": 2.327918214844518e-05,
+      "loss": 0.4312,
+      "step": 630
+    },
+    {
+      "epoch": 1.7600554785020806,
+      "grad_norm": 1.1194138526916504,
+      "learning_rate": 2.3164204577239037e-05,
+      "loss": 0.3986,
+      "step": 635
+    },
+    {
+      "epoch": 1.7739251040221915,
+      "grad_norm": 1.197754979133606,
+      "learning_rate": 2.304854131422971e-05,
+      "loss": 0.4333,
+      "step": 640
+    },
+    {
+      "epoch": 1.7877947295423025,
+      "grad_norm": 1.1752138137817383,
+      "learning_rate": 2.2932202073695003e-05,
+      "loss": 0.389,
+      "step": 645
+    },
+    {
+      "epoch": 1.8016643550624134,
+      "grad_norm": 1.1883151531219482,
+      "learning_rate": 2.281519662668645e-05,
+      "loss": 0.3669,
+      "step": 650
+    },
+    {
+      "epoch": 1.8155339805825244,
+      "grad_norm": 1.2165765762329102,
+      "learning_rate": 2.2697534800208655e-05,
+      "loss": 0.4625,
+      "step": 655
+    },
+    {
+      "epoch": 1.8294036061026353,
+      "grad_norm": 1.1365174055099487,
+      "learning_rate": 2.2579226476393932e-05,
+      "loss": 0.4195,
+      "step": 660
+    },
+    {
+      "epoch": 1.8432732316227463,
+      "grad_norm": 1.1491037607192993,
+      "learning_rate": 2.2460281591672365e-05,
+      "loss": 0.4132,
+      "step": 665
+    },
+    {
+      "epoch": 1.8571428571428572,
+      "grad_norm": 1.3076406717300415,
+      "learning_rate": 2.2340710135937233e-05,
+      "loss": 0.3854,
+      "step": 670
+    },
+    {
+      "epoch": 1.8710124826629682,
+      "grad_norm": 1.1297541856765747,
+      "learning_rate": 2.2220522151706012e-05,
+      "loss": 0.4352,
+      "step": 675
+    },
+    {
+      "epoch": 1.884882108183079,
+      "grad_norm": 1.168922781944275,
+      "learning_rate": 2.2099727733276905e-05,
+      "loss": 0.3606,
+      "step": 680
+    },
+    {
+      "epoch": 1.89875173370319,
+      "grad_norm": 1.1461069583892822,
+      "learning_rate": 2.1978337025881062e-05,
+      "loss": 0.3786,
+      "step": 685
+    },
+    {
+      "epoch": 1.912621359223301,
+      "grad_norm": 1.3558424711227417,
+      "learning_rate": 2.18563602248305e-05,
+      "loss": 0.3852,
+      "step": 690
+    },
+    {
+      "epoch": 1.926490984743412,
+      "grad_norm": 1.1597490310668945,
+      "learning_rate": 2.1733807574661823e-05,
+      "loss": 0.3721,
+      "step": 695
+    },
+    {
+      "epoch": 1.940360610263523,
+      "grad_norm": 1.5078226327896118,
+      "learning_rate": 2.1610689368275827e-05,
+      "loss": 0.4072,
+      "step": 700
+    },
+    {
+      "epoch": 1.9542302357836339,
+      "grad_norm": 1.1332188844680786,
+      "learning_rate": 2.148701594607298e-05,
+      "loss": 0.3278,
+      "step": 705
+    },
+    {
+      "epoch": 1.9680998613037448,
+      "grad_norm": 1.2428522109985352,
+      "learning_rate": 2.1362797695085e-05,
+      "loss": 0.3739,
+      "step": 710
+    },
+    {
+      "epoch": 1.9819694868238558,
+      "grad_norm": 1.3413153886795044,
+      "learning_rate": 2.123804504810246e-05,
+      "loss": 0.3625,
+      "step": 715
+    },
+    {
+      "epoch": 1.9958391123439667,
+      "grad_norm": 1.2263511419296265,
+      "learning_rate": 2.1112768482798553e-05,
+      "loss": 0.3571,
+      "step": 720
+    },
+    {
+      "epoch": 2.0083217753120666,
+      "grad_norm": 1.1522630453109741,
+      "learning_rate": 2.0986978520849108e-05,
+      "loss": 0.3112,
+      "step": 725
+    },
+    {
+      "epoch": 2.0221914008321775,
+      "grad_norm": 1.0950467586517334,
+      "learning_rate": 2.0860685727048895e-05,
+      "loss": 0.2632,
+      "step": 730
+    },
+    {
+      "epoch": 2.0360610263522885,
+      "grad_norm": 1.1049528121948242,
+      "learning_rate": 2.0733900708424318e-05,
+      "loss": 0.3054,
+      "step": 735
+    },
+    {
+      "epoch": 2.0499306518723994,
+      "grad_norm": 1.146390676498413,
+      "learning_rate": 2.060663411334255e-05,
+      "loss": 0.3114,
+      "step": 740
+    },
+    {
+      "epoch": 2.0638002773925104,
+      "grad_norm": 1.360620141029358,
+      "learning_rate": 2.0478896630617216e-05,
+      "loss": 0.2868,
+      "step": 745
+    },
+    {
+      "epoch": 2.0776699029126213,
+      "grad_norm": 1.2842508554458618,
+      "learning_rate": 2.0350698988610645e-05,
+      "loss": 0.311,
+      "step": 750
+    },
+    {
+      "epoch": 2.0915395284327323,
+      "grad_norm": 1.1546225547790527,
+      "learning_rate": 2.0222051954332835e-05,
+      "loss": 0.321,
+      "step": 755
+    },
+    {
+      "epoch": 2.1054091539528432,
+      "grad_norm": 1.112905502319336,
+      "learning_rate": 2.0092966332537152e-05,
+      "loss": 0.2572,
+      "step": 760
+    },
+    {
+      "epoch": 2.119278779472954,
+      "grad_norm": 1.122055172920227,
+      "learning_rate": 1.9963452964812873e-05,
+      "loss": 0.268,
+      "step": 765
+    },
+    {
+      "epoch": 2.133148404993065,
+      "grad_norm": 1.007519006729126,
+      "learning_rate": 1.9833522728674624e-05,
+      "loss": 0.2998,
+      "step": 770
+    },
+    {
+      "epoch": 2.147018030513176,
+      "grad_norm": 1.0647330284118652,
+      "learning_rate": 1.9703186536648786e-05,
+      "loss": 0.2406,
+      "step": 775
+    },
+    {
+      "epoch": 2.160887656033287,
+      "grad_norm": 1.1296517848968506,
+      "learning_rate": 1.957245533535701e-05,
+      "loss": 0.2856,
+      "step": 780
+    },
+    {
+      "epoch": 2.174757281553398,
+      "grad_norm": 1.148427963256836,
+      "learning_rate": 1.9441340104596808e-05,
+      "loss": 0.2616,
+      "step": 785
+    },
+    {
+      "epoch": 2.188626907073509,
+      "grad_norm": 1.1418274641036987,
+      "learning_rate": 1.9309851856419403e-05,
+      "loss": 0.2816,
+      "step": 790
+    },
+    {
+      "epoch": 2.20249653259362,
+      "grad_norm": 1.2052356004714966,
+      "learning_rate": 1.9178001634204858e-05,
+      "loss": 0.2679,
+      "step": 795
+    },
+    {
+      "epoch": 2.216366158113731,
+      "grad_norm": 1.0612205266952515,
+      "learning_rate": 1.904580051173454e-05,
+      "loss": 0.2781,
+      "step": 800
+    },
+    {
+      "epoch": 2.230235783633842,
+      "grad_norm": 1.198889136314392,
+      "learning_rate": 1.8913259592261105e-05,
+      "loss": 0.2583,
+      "step": 805
+    },
+    {
+      "epoch": 2.2441054091539527,
+      "grad_norm": 1.1233667135238647,
+      "learning_rate": 1.878039000757591e-05,
+      "loss": 0.2989,
+      "step": 810
+    },
+    {
+      "epoch": 2.2579750346740637,
+      "grad_norm": 1.131326675415039,
+      "learning_rate": 1.8647202917074138e-05,
+      "loss": 0.2489,
+      "step": 815
+    },
+    {
+      "epoch": 2.2718446601941746,
+      "grad_norm": 1.171873927116394,
+      "learning_rate": 1.85137095068175e-05,
+      "loss": 0.3085,
+      "step": 820
+    },
+    {
+      "epoch": 2.2857142857142856,
+      "grad_norm": 1.1070297956466675,
+      "learning_rate": 1.8379920988594756e-05,
+      "loss": 0.2358,
+      "step": 825
+    },
+    {
+      "epoch": 2.2995839112343965,
+      "grad_norm": 1.2584491968154907,
+      "learning_rate": 1.824584859898009e-05,
+      "loss": 0.2459,
+      "step": 830
+    },
+    {
+      "epoch": 2.3134535367545075,
+      "grad_norm": 1.1663209199905396,
+      "learning_rate": 1.811150359838934e-05,
+      "loss": 0.2666,
+      "step": 835
+    },
+    {
+      "epoch": 2.3273231622746184,
+      "grad_norm": 1.0747770071029663,
+      "learning_rate": 1.79768972701343e-05,
+      "loss": 0.2603,
+      "step": 840
+    },
+    {
+      "epoch": 2.3411927877947294,
+      "grad_norm": 1.1685748100280762,
+      "learning_rate": 1.784204091947502e-05,
+      "loss": 0.2675,
+      "step": 845
+    },
+    {
+      "epoch": 2.3550624133148403,
+      "grad_norm": 1.055817723274231,
+      "learning_rate": 1.770694587267034e-05,
+      "loss": 0.2473,
+      "step": 850
+    },
+    {
+      "epoch": 2.3689320388349513,
+      "grad_norm": 1.434633493423462,
+      "learning_rate": 1.7571623476026624e-05,
+      "loss": 0.2786,
+      "step": 855
+    },
+    {
+      "epoch": 2.3828016643550622,
+      "grad_norm": 1.1234263181686401,
+      "learning_rate": 1.743608509494477e-05,
+      "loss": 0.2524,
+      "step": 860
+    },
+    {
+      "epoch": 2.396671289875173,
+      "grad_norm": 1.1498136520385742,
+      "learning_rate": 1.7300342112965717e-05,
+      "loss": 0.2738,
+      "step": 865
+    },
+    {
+      "epoch": 2.410540915395284,
+      "grad_norm": 1.1265621185302734,
+      "learning_rate": 1.7164405930814313e-05,
+      "loss": 0.2519,
+      "step": 870
+    },
+    {
+      "epoch": 2.424410540915395,
+      "grad_norm": 1.1003201007843018,
+      "learning_rate": 1.7028287965441843e-05,
+      "loss": 0.2923,
+      "step": 875
+    },
+    {
+      "epoch": 2.438280166435506,
+      "grad_norm": 1.170314908027649,
+      "learning_rate": 1.6891999649067105e-05,
+      "loss": 0.2434,
+      "step": 880
+    },
+    {
+      "epoch": 2.452149791955617,
+      "grad_norm": 1.3272323608398438,
+      "learning_rate": 1.6755552428216278e-05,
+      "loss": 0.2573,
+      "step": 885
+    },
+    {
+      "epoch": 2.466019417475728,
+      "grad_norm": 1.2042258977890015,
+      "learning_rate": 1.6618957762761555e-05,
+      "loss": 0.2406,
+      "step": 890
+    },
+    {
+      "epoch": 2.479889042995839,
+      "grad_norm": 1.2370964288711548,
+      "learning_rate": 1.6482227124958617e-05,
+      "loss": 0.2438,
+      "step": 895
+    },
+    {
+      "epoch": 2.49375866851595,
+      "grad_norm": 1.3068962097167969,
+      "learning_rate": 1.6345371998483155e-05,
+      "loss": 0.2423,
+      "step": 900
+    },
+    {
+      "epoch": 2.507628294036061,
+      "grad_norm": 1.1322987079620361,
+      "learning_rate": 1.6208403877466354e-05,
+      "loss": 0.2355,
+      "step": 905
+    },
+    {
+      "epoch": 2.5214979195561718,
+      "grad_norm": 1.19632887840271,
+      "learning_rate": 1.6071334265529538e-05,
+      "loss": 0.2376,
+      "step": 910
+    },
+    {
+      "epoch": 2.5353675450762827,
+      "grad_norm": 1.1878324747085571,
+      "learning_rate": 1.5934174674818013e-05,
+      "loss": 0.2495,
+      "step": 915
+    },
+    {
+      "epoch": 2.5492371705963937,
+      "grad_norm": 1.1836864948272705,
+      "learning_rate": 1.5796936625034177e-05,
+      "loss": 0.2526,
+      "step": 920
+    },
+    {
+      "epoch": 2.5631067961165046,
+      "grad_norm": 1.1582130193710327,
+      "learning_rate": 1.5659631642470025e-05,
+      "loss": 0.2388,
+      "step": 925
+    },
+    {
+      "epoch": 2.5769764216366156,
+      "grad_norm": 1.140184760093689,
+      "learning_rate": 1.5522271259039063e-05,
+      "loss": 0.2468,
+      "step": 930
+    },
+    {
+      "epoch": 2.5908460471567265,
+      "grad_norm": 1.086246132850647,
+      "learning_rate": 1.5384867011307797e-05,
+      "loss": 0.2685,
+      "step": 935
+    },
+    {
+      "epoch": 2.6047156726768375,
+      "grad_norm": 1.1802036762237549,
+      "learning_rate": 1.5247430439526771e-05,
+      "loss": 0.2597,
+      "step": 940
+    },
+    {
+      "epoch": 2.6185852981969484,
+      "grad_norm": 1.1825424432754517,
+      "learning_rate": 1.5109973086661361e-05,
+      "loss": 0.2185,
+      "step": 945
+    },
+    {
+      "epoch": 2.63245492371706,
+      "grad_norm": 1.2922337055206299,
+      "learning_rate": 1.4972506497422298e-05,
+      "loss": 0.2154,
+      "step": 950
+    },
+    {
+      "epoch": 2.6463245492371708,
+      "grad_norm": 1.451962947845459,
+      "learning_rate": 1.4835042217296032e-05,
+      "loss": 0.2287,
+      "step": 955
+    },
+    {
+      "epoch": 2.6601941747572817,
+      "grad_norm": 1.0612127780914307,
+      "learning_rate": 1.4697591791575115e-05,
+      "loss": 0.2298,
+      "step": 960
+    },
+    {
+      "epoch": 2.6740638002773927,
+      "grad_norm": 1.1951512098312378,
+      "learning_rate": 1.4560166764388465e-05,
+      "loss": 0.2392,
+      "step": 965
+    },
+    {
+      "epoch": 2.6879334257975036,
+      "grad_norm": 1.1254812479019165,
+      "learning_rate": 1.442277867773185e-05,
+      "loss": 0.211,
+      "step": 970
+    },
+    {
+      "epoch": 2.7018030513176146,
+      "grad_norm": 1.0225093364715576,
+      "learning_rate": 1.4285439070498504e-05,
+      "loss": 0.2218,
+      "step": 975
+    },
+    {
+      "epoch": 2.7156726768377255,
+      "grad_norm": 1.0218101739883423,
+      "learning_rate": 1.4148159477509968e-05,
+      "loss": 0.2297,
+      "step": 980
+    },
+    {
+      "epoch": 2.7295423023578365,
+      "grad_norm": 1.3001481294631958,
+      "learning_rate": 1.4010951428547356e-05,
+      "loss": 0.1802,
+      "step": 985
+    },
+    {
+      "epoch": 2.7434119278779474,
+      "grad_norm": 1.085077166557312,
+      "learning_rate": 1.3873826447382957e-05,
+      "loss": 0.2022,
+      "step": 990
+    },
+    {
+      "epoch": 2.7572815533980584,
+      "grad_norm": 1.1113858222961426,
+      "learning_rate": 1.3736796050812406e-05,
+      "loss": 0.2163,
+      "step": 995
+    },
+    {
+      "epoch": 2.7711511789181693,
+      "grad_norm": 1.412807822227478,
+      "learning_rate": 1.3599871747687394e-05,
+      "loss": 0.2158,
+      "step": 1000
+    },
+    {
+      "epoch": 2.7850208044382803,
+      "grad_norm": 5.181802749633789,
+      "learning_rate": 1.346306503794909e-05,
+      "loss": 0.2101,
+      "step": 1005
+    },
+    {
+      "epoch": 2.798890429958391,
+      "grad_norm": 1.1314557790756226,
+      "learning_rate": 1.3326387411662279e-05,
+      "loss": 0.1977,
+      "step": 1010
+    },
+    {
+      "epoch": 2.812760055478502,
+      "grad_norm": 1.1890625953674316,
+      "learning_rate": 1.3189850348050336e-05,
+      "loss": 0.2135,
+      "step": 1015
+    },
+    {
+      "epoch": 2.826629680998613,
+      "grad_norm": 1.2581861019134521,
+      "learning_rate": 1.3053465314531118e-05,
+      "loss": 0.2091,
+      "step": 1020
+    },
+    {
+      "epoch": 2.840499306518724,
+      "grad_norm": 1.0491549968719482,
+      "learning_rate": 1.2917243765753836e-05,
+      "loss": 0.2041,
+      "step": 1025
+    },
+    {
+      "epoch": 2.854368932038835,
+      "grad_norm": 1.208863615989685,
+      "learning_rate": 1.2781197142637018e-05,
+      "loss": 0.1964,
+      "step": 1030
+    },
+    {
+      "epoch": 2.868238557558946,
+      "grad_norm": 1.2405157089233398,
+      "learning_rate": 1.2645336871407615e-05,
+      "loss": 0.2303,
+      "step": 1035
+    },
+    {
+      "epoch": 2.882108183079057,
+      "grad_norm": 1.1387736797332764,
+      "learning_rate": 1.2509674362641327e-05,
+      "loss": 0.1905,
+      "step": 1040
+    },
+    {
+      "epoch": 2.895977808599168,
+      "grad_norm": 1.1768152713775635,
+      "learning_rate": 1.237422101030426e-05,
+      "loss": 0.1963,
+      "step": 1045
+    },
+    {
+      "epoch": 2.909847434119279,
+      "grad_norm": 1.097781777381897,
+      "learning_rate": 1.223898819079598e-05,
+      "loss": 0.1991,
+      "step": 1050
+    },
+    {
+      "epoch": 2.9237170596393898,
+      "grad_norm": 1.2869900465011597,
+      "learning_rate": 1.2103987261994029e-05,
+      "loss": 0.1915,
+      "step": 1055
+    },
+    {
+      "epoch": 2.9375866851595007,
+      "grad_norm": 1.0657851696014404,
+      "learning_rate": 1.1969229562300025e-05,
+      "loss": 0.1841,
+      "step": 1060
+    },
+    {
+      "epoch": 2.9514563106796117,
+      "grad_norm": 1.0998252630233765,
+      "learning_rate": 1.1834726409687351e-05,
+      "loss": 0.1834,
+      "step": 1065
+    },
+    {
+      "epoch": 2.9653259361997226,
+      "grad_norm": 1.1832915544509888,
+      "learning_rate": 1.1700489100750585e-05,
+      "loss": 0.1848,
+      "step": 1070
+    },
+    {
+      "epoch": 2.9791955617198336,
+      "grad_norm": 1.1553837060928345,
+      "learning_rate": 1.1566528909756757e-05,
+      "loss": 0.1986,
+      "step": 1075
+    },
+    {
+      "epoch": 2.9930651872399445,
+      "grad_norm": 1.2271969318389893,
+      "learning_rate": 1.1432857087698414e-05,
+      "loss": 0.1971,
+      "step": 1080
+    },
+    {
+      "epoch": 3.0055478502080444,
+      "grad_norm": 0.9039092063903809,
+      "learning_rate": 1.1299484861348712e-05,
+      "loss": 0.1445,
+      "step": 1085
+    },
+    {
+      "epoch": 3.0194174757281553,
+      "grad_norm": 1.064099669456482,
+      "learning_rate": 1.1166423432318474e-05,
+      "loss": 0.1397,
+      "step": 1090
+    },
+    {
+      "epoch": 3.0332871012482663,
+      "grad_norm": 1.0022575855255127,
+      "learning_rate": 1.1033683976115408e-05,
+      "loss": 0.1541,
+      "step": 1095
+    },
+    {
+      "epoch": 3.0471567267683772,
+      "grad_norm": 1.0292963981628418,
+      "learning_rate": 1.0901277641205507e-05,
+      "loss": 0.1566,
+      "step": 1100
+    },
+    {
+      "epoch": 3.061026352288488,
+      "grad_norm": 1.3775979280471802,
+      "learning_rate": 1.0769215548076707e-05,
+      "loss": 0.1489,
+      "step": 1105
+    },
+    {
+      "epoch": 3.074895977808599,
+      "grad_norm": 1.095622181892395,
+      "learning_rate": 1.063750878830492e-05,
+      "loss": 0.131,
+      "step": 1110
+    },
+    {
+      "epoch": 3.08876560332871,
+      "grad_norm": 1.1273585557937622,
+      "learning_rate": 1.0506168423622461e-05,
+      "loss": 0.1386,
+      "step": 1115
+    },
+    {
+      "epoch": 3.102635228848821,
+      "grad_norm": 1.0857326984405518,
+      "learning_rate": 1.0375205484989001e-05,
+      "loss": 0.1292,
+      "step": 1120
+    },
+    {
+      "epoch": 3.116504854368932,
+      "grad_norm": 1.1363747119903564,
+      "learning_rate": 1.0244630971665118e-05,
+      "loss": 0.1332,
+      "step": 1125
+    },
+    {
+      "epoch": 3.130374479889043,
+      "grad_norm": 1.252336859703064,
+      "learning_rate": 1.0114455850288486e-05,
+      "loss": 0.138,
+      "step": 1130
+    },
+    {
+      "epoch": 3.144244105409154,
+      "grad_norm": 1.0865238904953003,
+      "learning_rate": 9.98469105395282e-06,
+      "loss": 0.1491,
+      "step": 1135
+    },
+    {
+      "epoch": 3.158113730929265,
+      "grad_norm": 1.1959137916564941,
+      "learning_rate": 9.855347481289617e-06,
+      "loss": 0.1361,
+      "step": 1140
+    },
+    {
+      "epoch": 3.171983356449376,
+      "grad_norm": 1.049391746520996,
+      "learning_rate": 9.726435995552815e-06,
+      "loss": 0.1419,
+      "step": 1145
+    },
+    {
+      "epoch": 3.1858529819694867,
+      "grad_norm": 1.1711589097976685,
+      "learning_rate": 9.597967423706436e-06,
+      "loss": 0.1514,
+      "step": 1150
+    },
+    {
+      "epoch": 3.1997226074895977,
+      "grad_norm": 1.1780716180801392,
+      "learning_rate": 9.4699525555152e-06,
+      "loss": 0.1606,
+      "step": 1155
+    },
+    {
+      "epoch": 3.2135922330097086,
+      "grad_norm": 1.179707646369934,
+      "learning_rate": 9.342402142638394e-06,
+      "loss": 0.1853,
+      "step": 1160
+    },
+    {
+      "epoch": 3.2274618585298196,
+      "grad_norm": 1.2147411108016968,
+      "learning_rate": 9.2153268977268e-06,
+      "loss": 0.1458,
+      "step": 1165
+    },
+    {
+      "epoch": 3.2413314840499305,
+      "grad_norm": 1.0931400060653687,
+      "learning_rate": 9.088737493522992e-06,
+      "loss": 0.1441,
+      "step": 1170
+    },
+    {
+      "epoch": 3.2552011095700415,
+      "grad_norm": 1.0420454740524292,
+      "learning_rate": 8.962644561964974e-06,
+      "loss": 0.1447,
+      "step": 1175
+    },
+    {
+      "epoch": 3.2690707350901524,
+      "grad_norm": 1.082470417022705,
+      "learning_rate": 8.837058693293188e-06,
+      "loss": 0.135,
+      "step": 1180
+    },
+    {
+      "epoch": 3.2829403606102634,
+      "grad_norm": 1.0198251008987427,
+      "learning_rate": 8.711990435161118e-06,
+      "loss": 0.14,
+      "step": 1185
+    },
+    {
+      "epoch": 3.2968099861303743,
+      "grad_norm": 1.134844183921814,
+      "learning_rate": 8.587450291749382e-06,
+      "loss": 0.1384,
+      "step": 1190
+    },
+    {
+      "epoch": 3.3106796116504853,
+      "grad_norm": 1.1528284549713135,
+      "learning_rate": 8.463448722883498e-06,
+      "loss": 0.1491,
+      "step": 1195
+    },
+    {
+      "epoch": 3.3245492371705962,
+      "grad_norm": 1.3226497173309326,
+      "learning_rate": 8.339996143155435e-06,
+      "loss": 0.1432,
+      "step": 1200
+    },
+    {
+      "epoch": 3.338418862690707,
+      "grad_norm": 1.098723292350769,
+      "learning_rate": 8.217102921048875e-06,
+      "loss": 0.1346,
+      "step": 1205
+    },
+    {
+      "epoch": 3.352288488210818,
+      "grad_norm": 1.1040832996368408,
+      "learning_rate": 8.094779378068425e-06,
+      "loss": 0.1436,
+      "step": 1210
+    },
+    {
+      "epoch": 3.366158113730929,
+      "grad_norm": 1.008098840713501,
+      "learning_rate": 7.973035787872712e-06,
+      "loss": 0.1314,
+      "step": 1215
+    },
+    {
+      "epoch": 3.38002773925104,
+      "grad_norm": 1.2032930850982666,
+      "learning_rate": 7.851882375411512e-06,
+      "loss": 0.129,
+      "step": 1220
+    },
+    {
+      "epoch": 3.393897364771151,
+      "grad_norm": 1.0998107194900513,
+      "learning_rate": 7.731329316067025e-06,
+      "loss": 0.141,
+      "step": 1225
+    },
+    {
+      "epoch": 3.407766990291262,
+      "grad_norm": 1.2666847705841064,
+      "learning_rate": 7.6113867347992264e-06,
+      "loss": 0.1477,
+      "step": 1230
+    },
+    {
+      "epoch": 3.421636615811373,
+      "grad_norm": 1.1232925653457642,
+      "learning_rate": 7.492064705295537e-06,
+      "loss": 0.1469,
+      "step": 1235
+    },
+    {
+      "epoch": 3.435506241331484,
+      "grad_norm": 1.0400546789169312,
+      "learning_rate": 7.373373249124709e-06,
+      "loss": 0.1458,
+      "step": 1240
+    },
+    {
+      "epoch": 3.449375866851595,
+      "grad_norm": 0.9294617772102356,
+      "learning_rate": 7.2553223348951755e-06,
+      "loss": 0.1396,
+      "step": 1245
+    },
+    {
+      "epoch": 3.4632454923717058,
+      "grad_norm": 1.007970929145813,
+      "learning_rate": 7.137921877417807e-06,
+      "loss": 0.1283,
+      "step": 1250
+    },
+    {
+      "epoch": 3.4771151178918167,
+      "grad_norm": 0.9476901888847351,
+      "learning_rate": 7.02118173687317e-06,
+      "loss": 0.1357,
+      "step": 1255
+    },
+    {
+      "epoch": 3.4909847434119277,
+      "grad_norm": 1.1563243865966797,
+      "learning_rate": 6.9051117179834334e-06,
+      "loss": 0.1369,
+      "step": 1260
+    },
+    {
+      "epoch": 3.5048543689320386,
+      "grad_norm": 1.0749354362487793,
+      "learning_rate": 6.789721569188837e-06,
+      "loss": 0.1185,
+      "step": 1265
+    },
+    {
+      "epoch": 3.5187239944521496,
+      "grad_norm": 1.0987670421600342,
+      "learning_rate": 6.675020981828994e-06,
+      "loss": 0.1262,
+      "step": 1270
+    },
+    {
+      "epoch": 3.5325936199722605,
+      "grad_norm": 0.9692766070365906,
+      "learning_rate": 6.561019589328916e-06,
+      "loss": 0.1224,
+      "step": 1275
+    },
+    {
+      "epoch": 3.5464632454923715,
+      "grad_norm": 1.037160873413086,
+      "learning_rate": 6.447726966389924e-06,
+      "loss": 0.137,
+      "step": 1280
+    },
+    {
+      "epoch": 3.5603328710124824,
+      "grad_norm": 1.1211687326431274,
+      "learning_rate": 6.335152628185517e-06,
+      "loss": 0.1448,
+      "step": 1285
+    },
+    {
+      "epoch": 3.5742024965325934,
+      "grad_norm": 1.0993516445159912,
+      "learning_rate": 6.223306029562167e-06,
+      "loss": 0.1298,
+      "step": 1290
+    },
+    {
+      "epoch": 3.5880721220527043,
+      "grad_norm": 1.0851831436157227,
+      "learning_rate": 6.112196564245276e-06,
+      "loss": 0.1263,
+      "step": 1295
+    },
+    {
+      "epoch": 3.6019417475728153,
+      "grad_norm": 1.3717095851898193,
+      "learning_rate": 6.001833564050196e-06,
+      "loss": 0.1105,
+      "step": 1300
+    },
+    {
+      "epoch": 3.615811373092926,
+      "grad_norm": 1.1500000953674316,
+      "learning_rate": 5.892226298098482e-06,
+      "loss": 0.1332,
+      "step": 1305
+    },
+    {
+      "epoch": 3.6296809986130376,
+      "grad_norm": 0.9285922050476074,
+      "learning_rate": 5.783383972039397e-06,
+      "loss": 0.1164,
+      "step": 1310
+    },
+    {
+      "epoch": 3.6435506241331486,
+      "grad_norm": 1.0234140157699585,
+      "learning_rate": 5.675315727276744e-06,
+      "loss": 0.1256,
+      "step": 1315
+    },
+    {
+      "epoch": 3.6574202496532595,
+      "grad_norm": 0.895950198173523,
+      "learning_rate": 5.5680306402011196e-06,
+      "loss": 0.12,
+      "step": 1320
+    },
+    {
+      "epoch": 3.6712898751733705,
+      "grad_norm": 1.1106315851211548,
+      "learning_rate": 5.461537721427595e-06,
+      "loss": 0.1196,
+      "step": 1325
+    },
+    {
+      "epoch": 3.6851595006934814,
+      "grad_norm": 1.03919517993927,
+      "learning_rate": 5.355845915038939e-06,
+      "loss": 0.1212,
+      "step": 1330
+    },
+    {
+      "epoch": 3.6990291262135924,
+      "grad_norm": 0.8517008423805237,
+      "learning_rate": 5.250964097834434e-06,
+      "loss": 0.1142,
+      "step": 1335
+    },
+    {
+      "epoch": 3.7128987517337033,
+      "grad_norm": 1.1055963039398193,
+      "learning_rate": 5.146901078584324e-06,
+      "loss": 0.1125,
+      "step": 1340
+    },
+    {
+      "epoch": 3.7267683772538143,
+      "grad_norm": 0.9949010610580444,
+      "learning_rate": 5.0436655972899946e-06,
+      "loss": 0.1259,
+      "step": 1345
+    },
+    {
+      "epoch": 3.740638002773925,
+      "grad_norm": 0.9716811776161194,
+      "learning_rate": 4.941266324449921e-06,
+      "loss": 0.1037,
+      "step": 1350
+    },
+    {
+      "epoch": 3.754507628294036,
+      "grad_norm": 0.9626354575157166,
+      "learning_rate": 4.839711860331446e-06,
+      "loss": 0.114,
+      "step": 1355
+    },
+    {
+      "epoch": 3.768377253814147,
+      "grad_norm": 1.0716753005981445,
+      "learning_rate": 4.739010734248478e-06,
+      "loss": 0.1186,
+      "step": 1360
+    },
+    {
+      "epoch": 3.782246879334258,
+      "grad_norm": 0.9848427772521973,
+      "learning_rate": 4.639171403845121e-06,
+      "loss": 0.1127,
+      "step": 1365
+    },
+    {
+      "epoch": 3.796116504854369,
+      "grad_norm": 1.086754560470581,
+      "learning_rate": 4.5402022543853476e-06,
+      "loss": 0.1253,
+      "step": 1370
+    },
+    {
+      "epoch": 3.80998613037448,
+      "grad_norm": 0.9147264361381531,
+      "learning_rate": 4.442111598048734e-06,
+      "loss": 0.1108,
+      "step": 1375
+    },
+    {
+      "epoch": 3.823855755894591,
+      "grad_norm": 1.0737582445144653,
+      "learning_rate": 4.344907673232347e-06,
+      "loss": 0.1086,
+      "step": 1380
+    },
+    {
+      "epoch": 3.837725381414702,
+      "grad_norm": 1.0010972023010254,
+      "learning_rate": 4.2485986438588115e-06,
+      "loss": 0.1114,
+      "step": 1385
+    },
+    {
+      "epoch": 3.851595006934813,
+      "grad_norm": 0.9007065892219543,
+      "learning_rate": 4.153192598690652e-06,
+      "loss": 0.1094,
+      "step": 1390
+    },
+    {
+      "epoch": 3.8654646324549238,
+      "grad_norm": 0.8125517964363098,
+      "learning_rate": 4.058697550650929e-06,
+      "loss": 0.1031,
+      "step": 1395
+    },
+    {
+      "epoch": 3.8793342579750347,
+      "grad_norm": 1.0790445804595947,
+      "learning_rate": 3.965121436150262e-06,
+      "loss": 0.1221,
+      "step": 1400
+    },
+    {
+      "epoch": 3.8932038834951457,
+      "grad_norm": 0.975977897644043,
+      "learning_rate": 3.87247211442026e-06,
+      "loss": 0.1269,
+      "step": 1405
+    },
+    {
+      "epoch": 3.9070735090152566,
+      "grad_norm": 0.937172532081604,
+      "learning_rate": 3.780757366853449e-06,
+      "loss": 0.1121,
+      "step": 1410
+    },
+    {
+      "epoch": 3.9209431345353676,
+      "grad_norm": 0.9537680745124817,
+      "learning_rate": 3.689984896349731e-06,
+      "loss": 0.1151,
+      "step": 1415
+    },
+    {
+      "epoch": 3.9348127600554785,
+      "grad_norm": 0.8375487923622131,
+      "learning_rate": 3.600162326669436e-06,
+      "loss": 0.0991,
+      "step": 1420
+    },
+    {
+      "epoch": 3.9486823855755895,
+      "grad_norm": 0.875694751739502,
+      "learning_rate": 3.5112972017930224e-06,
+      "loss": 0.1208,
+      "step": 1425
+    },
+    {
+      "epoch": 3.9625520110957004,
+      "grad_norm": 0.9617794156074524,
+      "learning_rate": 3.423396985287469e-06,
+      "loss": 0.1314,
+      "step": 1430
+    },
+    {
+      "epoch": 3.9764216366158114,
+      "grad_norm": 1.0028330087661743,
+      "learning_rate": 3.3364690596794384e-06,
+      "loss": 0.12,
+      "step": 1435
+    },
+    {
+      "epoch": 3.9902912621359223,
+      "grad_norm": 1.1332823038101196,
+      "learning_rate": 3.2505207258352297e-06,
+      "loss": 0.1295,
+      "step": 1440
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1805,
+  "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": 2.0937263254947758e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

23_128_e5_3e-5/checkpoint-1444/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ebded260fea4e3a399f7970554139fd5ed218599097aac84526fbe4423490a7
+size 7736

23_128_e5_3e-5/checkpoint-1444/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)

23_128_e5_3e-5/checkpoint-1805/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

23_128_e5_3e-5/checkpoint-1805/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": [
+    "up_proj",
+    "q_proj",
+    "down_proj",
+    "v_proj",
+    "gate_proj",
+    "o_proj",
+    "k_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

23_128_e5_3e-5/checkpoint-1805/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3800ea2ce445eb426fcc3e9b9ca7e625ccb767505d571eea8f3e500c3f1fd46
+size 791751704

23_128_e5_3e-5/checkpoint-1805/latest

@@ -0,0 +1 @@
+global_step1805

23_128_e5_3e-5/checkpoint-1805/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11bfae26b7005ab6e06087ca91518be4161591c5d5fa38643a2dd775c8765154
+size 15920

23_128_e5_3e-5/checkpoint-1805/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d21a3c0e7a512b0b2730b1cd371796c17e4c5e74cd371ba3e9d3bf7b5f2acebf
+size 15920

23_128_e5_3e-5/checkpoint-1805/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:befa44ba8f631cc474117086ec3c8e95165663fc618c784bdf57914d553fa9c5
+size 15920