Upload folder using huggingface_hub

15_128_e5_3e-5/checkpoint-1101/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

15_128_e5_3e-5/checkpoint-1101/adapter_config.json

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

15_128_e5_3e-5/checkpoint-1101/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:243a0215467d7dd9b503e46b2ee214a89753d6ea28f365e2ccc432b2ea4fdee6
+size 791751704

15_128_e5_3e-5/checkpoint-1101/latest

@@ -0,0 +1 @@
+global_step1101

15_128_e5_3e-5/checkpoint-1101/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:743ef387a7745577c0824d0a05fca5c52802aea3370dd80eb47fed83f640b2a8
+size 15920

15_128_e5_3e-5/checkpoint-1101/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4e91f05e7ce6b1ee576731c10ad6b72509b683dab0b904f81c5fdc3148e68b6
+size 15920

15_128_e5_3e-5/checkpoint-1101/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11add4180835a2caa2139a375eff79ec626177f844a0326c3b027c24a5e2b1ba
+size 15920

15_128_e5_3e-5/checkpoint-1101/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:978e768ff9f36e799e191418a97f043cf7b6e1d14d15783477866c4b80bedc79
+size 15920

15_128_e5_3e-5/checkpoint-1101/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:492f1f86f8f96673fd431c53e917cb41e7e791cb7a94769a5ead610030e22820
+size 15920

15_128_e5_3e-5/checkpoint-1101/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c47af6d28f61754ffd4b571937b506a1ae68b57489b1efee4bc28c29945a7f12
+size 15920

15_128_e5_3e-5/checkpoint-1101/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67c2fb50fdfd32cf1deaf07e1d1cd05517de330f52309319b1577ebbf2cc7a5e
+size 15920

15_128_e5_3e-5/checkpoint-1101/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bce2a18a619e755002a64b2278d1966d58f34bfba97d442b23b3c1994f3ea0fa
+size 15920

15_128_e5_3e-5/checkpoint-1101/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da8f7c6edaeabb14da6f50acb916810b0233729bd56d037ada87efd49aac1da2
+size 1064

15_128_e5_3e-5/checkpoint-1101/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
+  }
+}

15_128_e5_3e-5/checkpoint-1101/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
+}

15_128_e5_3e-5/checkpoint-1101/trainer_state.json

@@ -0,0 +1,1574 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1101,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.013623978201634877,
+      "grad_norm": 1.2644609212875366,
+      "learning_rate": 1.3043478260869566e-06,
+      "loss": 1.3371,
+      "step": 5
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.9776316285133362,
+      "learning_rate": 2.9347826086956523e-06,
+      "loss": 1.3342,
+      "step": 10
+    },
+    {
+      "epoch": 0.04087193460490463,
+      "grad_norm": 0.8481394052505493,
+      "learning_rate": 4.565217391304348e-06,
+      "loss": 1.3057,
+      "step": 15
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.6515122652053833,
+      "learning_rate": 6.195652173913044e-06,
+      "loss": 1.2589,
+      "step": 20
+    },
+    {
+      "epoch": 0.0681198910081744,
+      "grad_norm": 0.5874186158180237,
+      "learning_rate": 7.826086956521738e-06,
+      "loss": 1.3107,
+      "step": 25
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.5392442345619202,
+      "learning_rate": 9.456521739130436e-06,
+      "loss": 1.2746,
+      "step": 30
+    },
+    {
+      "epoch": 0.09536784741144415,
+      "grad_norm": 0.47512176632881165,
+      "learning_rate": 1.108695652173913e-05,
+      "loss": 1.187,
+      "step": 35
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.4385203421115875,
+      "learning_rate": 1.2717391304347827e-05,
+      "loss": 1.2059,
+      "step": 40
+    },
+    {
+      "epoch": 0.1226158038147139,
+      "grad_norm": 0.6362432241439819,
+      "learning_rate": 1.4347826086956522e-05,
+      "loss": 1.1893,
+      "step": 45
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 0.46012377738952637,
+      "learning_rate": 1.597826086956522e-05,
+      "loss": 1.2839,
+      "step": 50
+    },
+    {
+      "epoch": 0.14986376021798364,
+      "grad_norm": 0.5726909637451172,
+      "learning_rate": 1.7608695652173915e-05,
+      "loss": 1.1741,
+      "step": 55
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.4440523684024811,
+      "learning_rate": 1.9239130434782607e-05,
+      "loss": 1.2477,
+      "step": 60
+    },
+    {
+      "epoch": 0.1771117166212534,
+      "grad_norm": 0.4722035229206085,
+      "learning_rate": 2.0869565217391306e-05,
+      "loss": 1.1941,
+      "step": 65
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 0.45267796516418457,
+      "learning_rate": 2.25e-05,
+      "loss": 1.2442,
+      "step": 70
+    },
+    {
+      "epoch": 0.20435967302452315,
+      "grad_norm": 0.5661253929138184,
+      "learning_rate": 2.4130434782608697e-05,
+      "loss": 1.2069,
+      "step": 75
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 0.46782588958740234,
+      "learning_rate": 2.5760869565217392e-05,
+      "loss": 1.1675,
+      "step": 80
+    },
+    {
+      "epoch": 0.23160762942779292,
+      "grad_norm": 0.48947134613990784,
+      "learning_rate": 2.7391304347826085e-05,
+      "loss": 1.1297,
+      "step": 85
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 0.5822666883468628,
+      "learning_rate": 2.9021739130434783e-05,
+      "loss": 1.1361,
+      "step": 90
+    },
+    {
+      "epoch": 0.25885558583106266,
+      "grad_norm": 0.48768264055252075,
+      "learning_rate": 2.9999902540146195e-05,
+      "loss": 1.1555,
+      "step": 95
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 0.5380802750587463,
+      "learning_rate": 2.999880613133526e-05,
+      "loss": 1.1335,
+      "step": 100
+    },
+    {
+      "epoch": 0.28610354223433243,
+      "grad_norm": 0.5643327832221985,
+      "learning_rate": 2.9996491578238983e-05,
+      "loss": 1.0905,
+      "step": 105
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 0.5208688974380493,
+      "learning_rate": 2.9992959068836304e-05,
+      "loss": 1.1015,
+      "step": 110
+    },
+    {
+      "epoch": 0.3133514986376022,
+      "grad_norm": 0.5542858839035034,
+      "learning_rate": 2.99882088900238e-05,
+      "loss": 1.0951,
+      "step": 115
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 0.5248655080795288,
+      "learning_rate": 2.9982241427592387e-05,
+      "loss": 1.0697,
+      "step": 120
+    },
+    {
+      "epoch": 0.3405994550408719,
+      "grad_norm": 0.6014358401298523,
+      "learning_rate": 2.997505716619599e-05,
+      "loss": 1.0873,
+      "step": 125
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 0.6086240410804749,
+      "learning_rate": 2.996665668931218e-05,
+      "loss": 1.0565,
+      "step": 130
+    },
+    {
+      "epoch": 0.3678474114441417,
+      "grad_norm": 0.6661133766174316,
+      "learning_rate": 2.9957040679194782e-05,
+      "loss": 1.0233,
+      "step": 135
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 0.707312285900116,
+      "learning_rate": 2.9946209916818477e-05,
+      "loss": 1.0613,
+      "step": 140
+    },
+    {
+      "epoch": 0.39509536784741145,
+      "grad_norm": 0.5628401041030884,
+      "learning_rate": 2.9934165281815363e-05,
+      "loss": 0.9882,
+      "step": 145
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 0.6965730786323547,
+      "learning_rate": 2.9920907752403513e-05,
+      "loss": 0.9984,
+      "step": 150
+    },
+    {
+      "epoch": 0.4223433242506812,
+      "grad_norm": 0.7086770534515381,
+      "learning_rate": 2.9906438405307548e-05,
+      "loss": 0.9605,
+      "step": 155
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 0.6676215529441833,
+      "learning_rate": 2.989075841567115e-05,
+      "loss": 1.0023,
+      "step": 160
+    },
+    {
+      "epoch": 0.44959128065395093,
+      "grad_norm": 0.8493645191192627,
+      "learning_rate": 2.987386905696167e-05,
+      "loss": 0.9039,
+      "step": 165
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 0.7191267609596252,
+      "learning_rate": 2.9855771700866665e-05,
+      "loss": 0.9652,
+      "step": 170
+    },
+    {
+      "epoch": 0.4768392370572207,
+      "grad_norm": 0.7579267621040344,
+      "learning_rate": 2.983646781718251e-05,
+      "loss": 0.9237,
+      "step": 175
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 0.8027481436729431,
+      "learning_rate": 2.9815958973695034e-05,
+      "loss": 0.9653,
+      "step": 180
+    },
+    {
+      "epoch": 0.5040871934604905,
+      "grad_norm": 0.8199446797370911,
+      "learning_rate": 2.9794246836052167e-05,
+      "loss": 0.9864,
+      "step": 185
+    },
+    {
+      "epoch": 0.5177111716621253,
+      "grad_norm": 0.7790918946266174,
+      "learning_rate": 2.977133316762869e-05,
+      "loss": 0.9403,
+      "step": 190
+    },
+    {
+      "epoch": 0.5313351498637602,
+      "grad_norm": 0.7630456686019897,
+      "learning_rate": 2.9747219829382997e-05,
+      "loss": 0.9161,
+      "step": 195
+    },
+    {
+      "epoch": 0.5449591280653951,
+      "grad_norm": 0.8171895742416382,
+      "learning_rate": 2.9721908779705974e-05,
+      "loss": 0.8556,
+      "step": 200
+    },
+    {
+      "epoch": 0.55858310626703,
+      "grad_norm": 0.8095611929893494,
+      "learning_rate": 2.969540207426193e-05,
+      "loss": 0.9051,
+      "step": 205
+    },
+    {
+      "epoch": 0.5722070844686649,
+      "grad_norm": 0.7118027806282043,
+      "learning_rate": 2.9667701865821666e-05,
+      "loss": 0.9304,
+      "step": 210
+    },
+    {
+      "epoch": 0.5858310626702997,
+      "grad_norm": 0.897177517414093,
+      "learning_rate": 2.9638810404087603e-05,
+      "loss": 0.87,
+      "step": 215
+    },
+    {
+      "epoch": 0.5994550408719346,
+      "grad_norm": 0.7748181819915771,
+      "learning_rate": 2.960873003551111e-05,
+      "loss": 0.8723,
+      "step": 220
+    },
+    {
+      "epoch": 0.6130790190735694,
+      "grad_norm": 0.8167896270751953,
+      "learning_rate": 2.9577463203101897e-05,
+      "loss": 0.8648,
+      "step": 225
+    },
+    {
+      "epoch": 0.6267029972752044,
+      "grad_norm": 0.8939515352249146,
+      "learning_rate": 2.9545012446229613e-05,
+      "loss": 0.7818,
+      "step": 230
+    },
+    {
+      "epoch": 0.6403269754768393,
+      "grad_norm": 0.7820625305175781,
+      "learning_rate": 2.951138040041764e-05,
+      "loss": 0.9032,
+      "step": 235
+    },
+    {
+      "epoch": 0.6539509536784741,
+      "grad_norm": 0.844601571559906,
+      "learning_rate": 2.9476569797129e-05,
+      "loss": 0.8736,
+      "step": 240
+    },
+    {
+      "epoch": 0.667574931880109,
+      "grad_norm": 0.7247692942619324,
+      "learning_rate": 2.944058346354454e-05,
+      "loss": 0.8779,
+      "step": 245
+    },
+    {
+      "epoch": 0.6811989100817438,
+      "grad_norm": 0.8070963621139526,
+      "learning_rate": 2.9403424322333326e-05,
+      "loss": 0.8503,
+      "step": 250
+    },
+    {
+      "epoch": 0.6948228882833788,
+      "grad_norm": 0.8530706167221069,
+      "learning_rate": 2.9365095391415254e-05,
+      "loss": 0.8546,
+      "step": 255
+    },
+    {
+      "epoch": 0.7084468664850136,
+      "grad_norm": 0.8497718572616577,
+      "learning_rate": 2.932559978371596e-05,
+      "loss": 0.78,
+      "step": 260
+    },
+    {
+      "epoch": 0.7220708446866485,
+      "grad_norm": 0.9066139459609985,
+      "learning_rate": 2.928494070691401e-05,
+      "loss": 0.829,
+      "step": 265
+    },
+    {
+      "epoch": 0.7356948228882834,
+      "grad_norm": 0.7985149621963501,
+      "learning_rate": 2.9243121463180362e-05,
+      "loss": 0.8262,
+      "step": 270
+    },
+    {
+      "epoch": 0.7493188010899182,
+      "grad_norm": 1.0057857036590576,
+      "learning_rate": 2.9200145448910184e-05,
+      "loss": 0.7681,
+      "step": 275
+    },
+    {
+      "epoch": 0.7629427792915532,
+      "grad_norm": 0.9126638770103455,
+      "learning_rate": 2.915601615444703e-05,
+      "loss": 0.8171,
+      "step": 280
+    },
+    {
+      "epoch": 0.776566757493188,
+      "grad_norm": 0.8743240237236023,
+      "learning_rate": 2.9110737163799347e-05,
+      "loss": 0.7672,
+      "step": 285
+    },
+    {
+      "epoch": 0.7901907356948229,
+      "grad_norm": 0.8928736448287964,
+      "learning_rate": 2.9064312154349395e-05,
+      "loss": 0.7824,
+      "step": 290
+    },
+    {
+      "epoch": 0.8038147138964578,
+      "grad_norm": 1.0869660377502441,
+      "learning_rate": 2.9016744896554606e-05,
+      "loss": 0.7687,
+      "step": 295
+    },
+    {
+      "epoch": 0.8174386920980926,
+      "grad_norm": 0.9520823955535889,
+      "learning_rate": 2.8968039253641347e-05,
+      "loss": 0.7603,
+      "step": 300
+    },
+    {
+      "epoch": 0.8310626702997275,
+      "grad_norm": 0.9313263297080994,
+      "learning_rate": 2.8918199181291154e-05,
+      "loss": 0.7344,
+      "step": 305
+    },
+    {
+      "epoch": 0.8446866485013624,
+      "grad_norm": 0.8860709071159363,
+      "learning_rate": 2.8867228727319484e-05,
+      "loss": 0.7221,
+      "step": 310
+    },
+    {
+      "epoch": 0.8583106267029973,
+      "grad_norm": 1.0076904296875,
+      "learning_rate": 2.8815132031346967e-05,
+      "loss": 0.7163,
+      "step": 315
+    },
+    {
+      "epoch": 0.8719346049046321,
+      "grad_norm": 1.0047346353530884,
+      "learning_rate": 2.8761913324463193e-05,
+      "loss": 0.731,
+      "step": 320
+    },
+    {
+      "epoch": 0.885558583106267,
+      "grad_norm": 0.9905893802642822,
+      "learning_rate": 2.8707576928883083e-05,
+      "loss": 0.7234,
+      "step": 325
+    },
+    {
+      "epoch": 0.8991825613079019,
+      "grad_norm": 0.8848779797554016,
+      "learning_rate": 2.8652127257595852e-05,
+      "loss": 0.7241,
+      "step": 330
+    },
+    {
+      "epoch": 0.9128065395095368,
+      "grad_norm": 1.086945652961731,
+      "learning_rate": 2.8595568814006618e-05,
+      "loss": 0.7474,
+      "step": 335
+    },
+    {
+      "epoch": 0.9264305177111717,
+      "grad_norm": 1.058239221572876,
+      "learning_rate": 2.853790619157063e-05,
+      "loss": 0.6512,
+      "step": 340
+    },
+    {
+      "epoch": 0.9400544959128065,
+      "grad_norm": 0.9978867173194885,
+      "learning_rate": 2.8479144073420237e-05,
+      "loss": 0.6968,
+      "step": 345
+    },
+    {
+      "epoch": 0.9536784741144414,
+      "grad_norm": 0.9755434989929199,
+      "learning_rate": 2.841928723198449e-05,
+      "loss": 0.6774,
+      "step": 350
+    },
+    {
+      "epoch": 0.9673024523160763,
+      "grad_norm": 0.9939659833908081,
+      "learning_rate": 2.835834052860162e-05,
+      "loss": 0.6851,
+      "step": 355
+    },
+    {
+      "epoch": 0.9809264305177112,
+      "grad_norm": 0.919965922832489,
+      "learning_rate": 2.8296308913124137e-05,
+      "loss": 0.6636,
+      "step": 360
+    },
+    {
+      "epoch": 0.9945504087193461,
+      "grad_norm": 1.1084834337234497,
+      "learning_rate": 2.8233197423516885e-05,
+      "loss": 0.696,
+      "step": 365
+    },
+    {
+      "epoch": 1.008174386920981,
+      "grad_norm": 1.0081297159194946,
+      "learning_rate": 2.816901118544785e-05,
+      "loss": 0.6079,
+      "step": 370
+    },
+    {
+      "epoch": 1.021798365122616,
+      "grad_norm": 0.9290661811828613,
+      "learning_rate": 2.810375541187188e-05,
+      "loss": 0.5794,
+      "step": 375
+    },
+    {
+      "epoch": 1.0354223433242506,
+      "grad_norm": 1.0772560834884644,
+      "learning_rate": 2.80374354026073e-05,
+      "loss": 0.5495,
+      "step": 380
+    },
+    {
+      "epoch": 1.0490463215258856,
+      "grad_norm": 1.0354938507080078,
+      "learning_rate": 2.79700565439055e-05,
+      "loss": 0.6109,
+      "step": 385
+    },
+    {
+      "epoch": 1.0626702997275204,
+      "grad_norm": 1.0381073951721191,
+      "learning_rate": 2.7901624308013465e-05,
+      "loss": 0.5849,
+      "step": 390
+    },
+    {
+      "epoch": 1.0762942779291553,
+      "grad_norm": 1.146996021270752,
+      "learning_rate": 2.7832144252729354e-05,
+      "loss": 0.5798,
+      "step": 395
+    },
+    {
+      "epoch": 1.0899182561307903,
+      "grad_norm": 1.0819785594940186,
+      "learning_rate": 2.776162202095111e-05,
+      "loss": 0.584,
+      "step": 400
+    },
+    {
+      "epoch": 1.103542234332425,
+      "grad_norm": 1.090579628944397,
+      "learning_rate": 2.7690063340218173e-05,
+      "loss": 0.556,
+      "step": 405
+    },
+    {
+      "epoch": 1.11716621253406,
+      "grad_norm": 1.034417748451233,
+      "learning_rate": 2.7617474022246297e-05,
+      "loss": 0.5788,
+      "step": 410
+    },
+    {
+      "epoch": 1.1307901907356948,
+      "grad_norm": 1.1490421295166016,
+      "learning_rate": 2.7543859962455576e-05,
+      "loss": 0.5576,
+      "step": 415
+    },
+    {
+      "epoch": 1.1444141689373297,
+      "grad_norm": 1.0201855897903442,
+      "learning_rate": 2.7469227139491603e-05,
+      "loss": 0.5886,
+      "step": 420
+    },
+    {
+      "epoch": 1.1580381471389645,
+      "grad_norm": 1.01875901222229,
+      "learning_rate": 2.7393581614739924e-05,
+      "loss": 0.6062,
+      "step": 425
+    },
+    {
+      "epoch": 1.1716621253405994,
+      "grad_norm": 1.0762981176376343,
+      "learning_rate": 2.7316929531833775e-05,
+      "loss": 0.5389,
+      "step": 430
+    },
+    {
+      "epoch": 1.1852861035422344,
+      "grad_norm": 1.0140329599380493,
+      "learning_rate": 2.7239277116155077e-05,
+      "loss": 0.5462,
+      "step": 435
+    },
+    {
+      "epoch": 1.1989100817438691,
+      "grad_norm": 1.1165515184402466,
+      "learning_rate": 2.7160630674328893e-05,
+      "loss": 0.5596,
+      "step": 440
+    },
+    {
+      "epoch": 1.2125340599455041,
+      "grad_norm": 1.0689040422439575,
+      "learning_rate": 2.7080996593711172e-05,
+      "loss": 0.5137,
+      "step": 445
+    },
+    {
+      "epoch": 1.226158038147139,
+      "grad_norm": 1.2397655248641968,
+      "learning_rate": 2.700038134187002e-05,
+      "loss": 0.5643,
+      "step": 450
+    },
+    {
+      "epoch": 1.2397820163487738,
+      "grad_norm": 0.9885849952697754,
+      "learning_rate": 2.691879146606043e-05,
+      "loss": 0.5921,
+      "step": 455
+    },
+    {
+      "epoch": 1.2534059945504088,
+      "grad_norm": 1.279818058013916,
+      "learning_rate": 2.6836233592692544e-05,
+      "loss": 0.5126,
+      "step": 460
+    },
+    {
+      "epoch": 1.2670299727520435,
+      "grad_norm": 1.1189254522323608,
+      "learning_rate": 2.675271442679346e-05,
+      "loss": 0.5198,
+      "step": 465
+    },
+    {
+      "epoch": 1.2806539509536785,
+      "grad_norm": 1.1345500946044922,
+      "learning_rate": 2.6668240751462707e-05,
+      "loss": 0.5117,
+      "step": 470
+    },
+    {
+      "epoch": 1.2942779291553133,
+      "grad_norm": 1.1066192388534546,
+      "learning_rate": 2.6582819427321313e-05,
+      "loss": 0.5314,
+      "step": 475
+    },
+    {
+      "epoch": 1.3079019073569482,
+      "grad_norm": 1.0839710235595703,
+      "learning_rate": 2.649645739195464e-05,
+      "loss": 0.5382,
+      "step": 480
+    },
+    {
+      "epoch": 1.3215258855585832,
+      "grad_norm": 1.052043080329895,
+      "learning_rate": 2.640916165934893e-05,
+      "loss": 0.5135,
+      "step": 485
+    },
+    {
+      "epoch": 1.335149863760218,
+      "grad_norm": 1.1333352327346802,
+      "learning_rate": 2.6320939319321657e-05,
+      "loss": 0.5359,
+      "step": 490
+    },
+    {
+      "epoch": 1.348773841961853,
+      "grad_norm": 1.21845543384552,
+      "learning_rate": 2.623179753694573e-05,
+      "loss": 0.4853,
+      "step": 495
+    },
+    {
+      "epoch": 1.3623978201634879,
+      "grad_norm": 1.142262578010559,
+      "learning_rate": 2.614174355196754e-05,
+      "loss": 0.4993,
+      "step": 500
+    },
+    {
+      "epoch": 1.3760217983651226,
+      "grad_norm": 1.037720799446106,
+      "learning_rate": 2.6050784678219024e-05,
+      "loss": 0.512,
+      "step": 505
+    },
+    {
+      "epoch": 1.3896457765667574,
+      "grad_norm": 1.1667251586914062,
+      "learning_rate": 2.5958928303023634e-05,
+      "loss": 0.4788,
+      "step": 510
+    },
+    {
+      "epoch": 1.4032697547683923,
+      "grad_norm": 1.046761393547058,
+      "learning_rate": 2.5866181886596367e-05,
+      "loss": 0.4867,
+      "step": 515
+    },
+    {
+      "epoch": 1.4168937329700273,
+      "grad_norm": 1.1029340028762817,
+      "learning_rate": 2.5772552961437893e-05,
+      "loss": 0.4799,
+      "step": 520
+    },
+    {
+      "epoch": 1.430517711171662,
+      "grad_norm": 1.1235623359680176,
+      "learning_rate": 2.5678049131722772e-05,
+      "loss": 0.4752,
+      "step": 525
+    },
+    {
+      "epoch": 1.444141689373297,
+      "grad_norm": 1.064278483390808,
+      "learning_rate": 2.5582678072681903e-05,
+      "loss": 0.5173,
+      "step": 530
+    },
+    {
+      "epoch": 1.457765667574932,
+      "grad_norm": 1.13005793094635,
+      "learning_rate": 2.5486447529979136e-05,
+      "loss": 0.4451,
+      "step": 535
+    },
+    {
+      "epoch": 1.4713896457765667,
+      "grad_norm": 1.0552195310592651,
+      "learning_rate": 2.5389365319082226e-05,
+      "loss": 0.4595,
+      "step": 540
+    },
+    {
+      "epoch": 1.4850136239782017,
+      "grad_norm": 1.1078029870986938,
+      "learning_rate": 2.5291439324628084e-05,
+      "loss": 0.4693,
+      "step": 545
+    },
+    {
+      "epoch": 1.4986376021798364,
+      "grad_norm": 1.090710997581482,
+      "learning_rate": 2.5192677499782413e-05,
+      "loss": 0.4537,
+      "step": 550
+    },
+    {
+      "epoch": 1.5122615803814714,
+      "grad_norm": 1.037945032119751,
+      "learning_rate": 2.5093087865593784e-05,
+      "loss": 0.4556,
+      "step": 555
+    },
+    {
+      "epoch": 1.5258855585831061,
+      "grad_norm": 1.1792460680007935,
+      "learning_rate": 2.499267851034221e-05,
+      "loss": 0.4734,
+      "step": 560
+    },
+    {
+      "epoch": 1.5395095367847411,
+      "grad_norm": 1.0996177196502686,
+      "learning_rate": 2.4891457588882238e-05,
+      "loss": 0.4444,
+      "step": 565
+    },
+    {
+      "epoch": 1.553133514986376,
+      "grad_norm": 1.13473641872406,
+      "learning_rate": 2.478943332198062e-05,
+      "loss": 0.4513,
+      "step": 570
+    },
+    {
+      "epoch": 1.5667574931880108,
+      "grad_norm": 1.212607979774475,
+      "learning_rate": 2.468661399564871e-05,
+      "loss": 0.4506,
+      "step": 575
+    },
+    {
+      "epoch": 1.5803814713896458,
+      "grad_norm": 1.0172079801559448,
+      "learning_rate": 2.458300796046946e-05,
+      "loss": 0.4238,
+      "step": 580
+    },
+    {
+      "epoch": 1.5940054495912808,
+      "grad_norm": 1.3280726671218872,
+      "learning_rate": 2.4478623630919236e-05,
+      "loss": 0.4419,
+      "step": 585
+    },
+    {
+      "epoch": 1.6076294277929155,
+      "grad_norm": 1.1108782291412354,
+      "learning_rate": 2.437346948468441e-05,
+      "loss": 0.3942,
+      "step": 590
+    },
+    {
+      "epoch": 1.6212534059945503,
+      "grad_norm": 1.157791256904602,
+      "learning_rate": 2.4267554061972873e-05,
+      "loss": 0.397,
+      "step": 595
+    },
+    {
+      "epoch": 1.6348773841961854,
+      "grad_norm": 1.1420488357543945,
+      "learning_rate": 2.416088596482039e-05,
+      "loss": 0.4849,
+      "step": 600
+    },
+    {
+      "epoch": 1.6485013623978202,
+      "grad_norm": 1.264007329940796,
+      "learning_rate": 2.405347385639202e-05,
+      "loss": 0.4365,
+      "step": 605
+    },
+    {
+      "epoch": 1.662125340599455,
+      "grad_norm": 1.0671672821044922,
+      "learning_rate": 2.394532646027848e-05,
+      "loss": 0.4259,
+      "step": 610
+    },
+    {
+      "epoch": 1.67574931880109,
+      "grad_norm": 1.245835781097412,
+      "learning_rate": 2.3836452559787673e-05,
+      "loss": 0.4078,
+      "step": 615
+    },
+    {
+      "epoch": 1.6893732970027249,
+      "grad_norm": 1.1798008680343628,
+      "learning_rate": 2.3726860997231356e-05,
+      "loss": 0.4319,
+      "step": 620
+    },
+    {
+      "epoch": 1.7029972752043596,
+      "grad_norm": 1.3666014671325684,
+      "learning_rate": 2.3616560673206984e-05,
+      "loss": 0.4645,
+      "step": 625
+    },
+    {
+      "epoch": 1.7166212534059946,
+      "grad_norm": 1.0500986576080322,
+      "learning_rate": 2.3505560545874843e-05,
+      "loss": 0.4543,
+      "step": 630
+    },
+    {
+      "epoch": 1.7302452316076296,
+      "grad_norm": 1.4002113342285156,
+      "learning_rate": 2.3393869630230495e-05,
+      "loss": 0.3978,
+      "step": 635
+    },
+    {
+      "epoch": 1.7438692098092643,
+      "grad_norm": 1.170268177986145,
+      "learning_rate": 2.3281496997372625e-05,
+      "loss": 0.4355,
+      "step": 640
+    },
+    {
+      "epoch": 1.757493188010899,
+      "grad_norm": 1.0691126585006714,
+      "learning_rate": 2.316845177376633e-05,
+      "loss": 0.4097,
+      "step": 645
+    },
+    {
+      "epoch": 1.771117166212534,
+      "grad_norm": 1.1894110441207886,
+      "learning_rate": 2.3054743140501877e-05,
+      "loss": 0.4025,
+      "step": 650
+    },
+    {
+      "epoch": 1.784741144414169,
+      "grad_norm": 1.0559704303741455,
+      "learning_rate": 2.2940380332549086e-05,
+      "loss": 0.4237,
+      "step": 655
+    },
+    {
+      "epoch": 1.7983651226158037,
+      "grad_norm": 1.1340513229370117,
+      "learning_rate": 2.282537263800727e-05,
+      "loss": 0.3523,
+      "step": 660
+    },
+    {
+      "epoch": 1.8119891008174387,
+      "grad_norm": 1.1120586395263672,
+      "learning_rate": 2.2709729397350904e-05,
+      "loss": 0.4037,
+      "step": 665
+    },
+    {
+      "epoch": 1.8256130790190737,
+      "grad_norm": 1.1562973260879517,
+      "learning_rate": 2.2593460002671024e-05,
+      "loss": 0.3839,
+      "step": 670
+    },
+    {
+      "epoch": 1.8392370572207084,
+      "grad_norm": 1.2246955633163452,
+      "learning_rate": 2.247657389691247e-05,
+      "loss": 0.4188,
+      "step": 675
+    },
+    {
+      "epoch": 1.8528610354223434,
+      "grad_norm": 0.9505524635314941,
+      "learning_rate": 2.2359080573106913e-05,
+      "loss": 0.3348,
+      "step": 680
+    },
+    {
+      "epoch": 1.8664850136239783,
+      "grad_norm": 1.161018967628479,
+      "learning_rate": 2.2240989573601902e-05,
+      "loss": 0.3978,
+      "step": 685
+    },
+    {
+      "epoch": 1.880108991825613,
+      "grad_norm": 1.076042890548706,
+      "learning_rate": 2.212231048928587e-05,
+      "loss": 0.3608,
+      "step": 690
+    },
+    {
+      "epoch": 1.8937329700272478,
+      "grad_norm": 1.248965859413147,
+      "learning_rate": 2.2003052958809185e-05,
+      "loss": 0.3621,
+      "step": 695
+    },
+    {
+      "epoch": 1.9073569482288828,
+      "grad_norm": 1.0826005935668945,
+      "learning_rate": 2.1883226667801374e-05,
+      "loss": 0.3733,
+      "step": 700
+    },
+    {
+      "epoch": 1.9209809264305178,
+      "grad_norm": 1.1231858730316162,
+      "learning_rate": 2.1762841348084425e-05,
+      "loss": 0.3916,
+      "step": 705
+    },
+    {
+      "epoch": 1.9346049046321525,
+      "grad_norm": 1.1095854043960571,
+      "learning_rate": 2.164190677688248e-05,
+      "loss": 0.3406,
+      "step": 710
+    },
+    {
+      "epoch": 1.9482288828337875,
+      "grad_norm": 1.0947611331939697,
+      "learning_rate": 2.1520432776027723e-05,
+      "loss": 0.3511,
+      "step": 715
+    },
+    {
+      "epoch": 1.9618528610354224,
+      "grad_norm": 1.035023808479309,
+      "learning_rate": 2.1398429211162706e-05,
+      "loss": 0.3894,
+      "step": 720
+    },
+    {
+      "epoch": 1.9754768392370572,
+      "grad_norm": 1.166717529296875,
+      "learning_rate": 2.127590599093909e-05,
+      "loss": 0.3692,
+      "step": 725
+    },
+    {
+      "epoch": 1.989100817438692,
+      "grad_norm": 1.1927279233932495,
+      "learning_rate": 2.1152873066212913e-05,
+      "loss": 0.3278,
+      "step": 730
+    },
+    {
+      "epoch": 2.002724795640327,
+      "grad_norm": 1.0591354370117188,
+      "learning_rate": 2.102934042923641e-05,
+      "loss": 0.3421,
+      "step": 735
+    },
+    {
+      "epoch": 2.016348773841962,
+      "grad_norm": 1.107465147972107,
+      "learning_rate": 2.0905318112846497e-05,
+      "loss": 0.274,
+      "step": 740
+    },
+    {
+      "epoch": 2.0299727520435966,
+      "grad_norm": 1.011695146560669,
+      "learning_rate": 2.0780816189649928e-05,
+      "loss": 0.2894,
+      "step": 745
+    },
+    {
+      "epoch": 2.043596730245232,
+      "grad_norm": 1.3591814041137695,
+      "learning_rate": 2.0655844771205253e-05,
+      "loss": 0.3034,
+      "step": 750
+    },
+    {
+      "epoch": 2.0572207084468666,
+      "grad_norm": 1.2014539241790771,
+      "learning_rate": 2.0530414007201566e-05,
+      "loss": 0.2508,
+      "step": 755
+    },
+    {
+      "epoch": 2.0708446866485013,
+      "grad_norm": 1.1204092502593994,
+      "learning_rate": 2.0404534084634228e-05,
+      "loss": 0.2766,
+      "step": 760
+    },
+    {
+      "epoch": 2.084468664850136,
+      "grad_norm": 1.0801162719726562,
+      "learning_rate": 2.0278215226977497e-05,
+      "loss": 0.285,
+      "step": 765
+    },
+    {
+      "epoch": 2.0980926430517712,
+      "grad_norm": 1.1138821840286255,
+      "learning_rate": 2.015146769335421e-05,
+      "loss": 0.2964,
+      "step": 770
+    },
+    {
+      "epoch": 2.111716621253406,
+      "grad_norm": 1.2337431907653809,
+      "learning_rate": 2.0024301777702586e-05,
+      "loss": 0.2889,
+      "step": 775
+    },
+    {
+      "epoch": 2.1253405994550407,
+      "grad_norm": 1.228090763092041,
+      "learning_rate": 1.9896727807940208e-05,
+      "loss": 0.2882,
+      "step": 780
+    },
+    {
+      "epoch": 2.138964577656676,
+      "grad_norm": 1.3023854494094849,
+      "learning_rate": 1.976875614512521e-05,
+      "loss": 0.2811,
+      "step": 785
+    },
+    {
+      "epoch": 2.1525885558583107,
+      "grad_norm": 1.1693427562713623,
+      "learning_rate": 1.964039718261478e-05,
+      "loss": 0.2795,
+      "step": 790
+    },
+    {
+      "epoch": 2.1662125340599454,
+      "grad_norm": 1.235745906829834,
+      "learning_rate": 1.95116613452211e-05,
+      "loss": 0.2745,
+      "step": 795
+    },
+    {
+      "epoch": 2.1798365122615806,
+      "grad_norm": 1.1393667459487915,
+      "learning_rate": 1.9382559088364636e-05,
+      "loss": 0.2928,
+      "step": 800
+    },
+    {
+      "epoch": 2.1934604904632153,
+      "grad_norm": 1.275499939918518,
+      "learning_rate": 1.9253100897225006e-05,
+      "loss": 0.2544,
+      "step": 805
+    },
+    {
+      "epoch": 2.20708446866485,
+      "grad_norm": 1.1270493268966675,
+      "learning_rate": 1.9123297285889424e-05,
+      "loss": 0.2696,
+      "step": 810
+    },
+    {
+      "epoch": 2.220708446866485,
+      "grad_norm": 1.0665929317474365,
+      "learning_rate": 1.899315879649878e-05,
+      "loss": 0.2448,
+      "step": 815
+    },
+    {
+      "epoch": 2.23433242506812,
+      "grad_norm": 1.0491502285003662,
+      "learning_rate": 1.8862695998391438e-05,
+      "loss": 0.259,
+      "step": 820
+    },
+    {
+      "epoch": 2.2479564032697548,
+      "grad_norm": 1.0268921852111816,
+      "learning_rate": 1.8731919487244855e-05,
+      "loss": 0.237,
+      "step": 825
+    },
+    {
+      "epoch": 2.2615803814713895,
+      "grad_norm": 1.1503524780273438,
+      "learning_rate": 1.8600839884215034e-05,
+      "loss": 0.2626,
+      "step": 830
+    },
+    {
+      "epoch": 2.2752043596730247,
+      "grad_norm": 1.155947208404541,
+      "learning_rate": 1.84694678350739e-05,
+      "loss": 0.2527,
+      "step": 835
+    },
+    {
+      "epoch": 2.2888283378746594,
+      "grad_norm": 1.2745076417922974,
+      "learning_rate": 1.8337814009344716e-05,
+      "loss": 0.2974,
+      "step": 840
+    },
+    {
+      "epoch": 2.302452316076294,
+      "grad_norm": 1.1524717807769775,
+      "learning_rate": 1.8205889099435543e-05,
+      "loss": 0.2676,
+      "step": 845
+    },
+    {
+      "epoch": 2.316076294277929,
+      "grad_norm": 1.2803691625595093,
+      "learning_rate": 1.807370381977082e-05,
+      "loss": 0.261,
+      "step": 850
+    },
+    {
+      "epoch": 2.329700272479564,
+      "grad_norm": 1.0910717248916626,
+      "learning_rate": 1.7941268905921208e-05,
+      "loss": 0.2543,
+      "step": 855
+    },
+    {
+      "epoch": 2.343324250681199,
+      "grad_norm": 1.127986192703247,
+      "learning_rate": 1.780859511373168e-05,
+      "loss": 0.2653,
+      "step": 860
+    },
+    {
+      "epoch": 2.3569482288828336,
+      "grad_norm": 1.1669334173202515,
+      "learning_rate": 1.7675693218447973e-05,
+      "loss": 0.2461,
+      "step": 865
+    },
+    {
+      "epoch": 2.370572207084469,
+      "grad_norm": 1.3159593343734741,
+      "learning_rate": 1.7542574013841453e-05,
+      "loss": 0.2459,
+      "step": 870
+    },
+    {
+      "epoch": 2.3841961852861036,
+      "grad_norm": 1.1560946702957153,
+      "learning_rate": 1.7409248311332503e-05,
+      "loss": 0.2403,
+      "step": 875
+    },
+    {
+      "epoch": 2.3978201634877383,
+      "grad_norm": 1.3311054706573486,
+      "learning_rate": 1.7275726939112465e-05,
+      "loss": 0.2627,
+      "step": 880
+    },
+    {
+      "epoch": 2.4114441416893735,
+      "grad_norm": 1.2287989854812622,
+      "learning_rate": 1.71420207412642e-05,
+      "loss": 0.2428,
+      "step": 885
+    },
+    {
+      "epoch": 2.4250681198910082,
+      "grad_norm": 0.9484009742736816,
+      "learning_rate": 1.7008140576881384e-05,
+      "loss": 0.2546,
+      "step": 890
+    },
+    {
+      "epoch": 2.438692098092643,
+      "grad_norm": 1.140745759010315,
+      "learning_rate": 1.6874097319186575e-05,
+      "loss": 0.2152,
+      "step": 895
+    },
+    {
+      "epoch": 2.452316076294278,
+      "grad_norm": 1.1644483804702759,
+      "learning_rate": 1.6739901854648138e-05,
+      "loss": 0.2806,
+      "step": 900
+    },
+    {
+      "epoch": 2.465940054495913,
+      "grad_norm": 1.1392990350723267,
+      "learning_rate": 1.660556508209606e-05,
+      "loss": 0.2534,
+      "step": 905
+    },
+    {
+      "epoch": 2.4795640326975477,
+      "grad_norm": 1.051167607307434,
+      "learning_rate": 1.647109791183684e-05,
+      "loss": 0.2345,
+      "step": 910
+    },
+    {
+      "epoch": 2.4931880108991824,
+      "grad_norm": 1.1279879808425903,
+      "learning_rate": 1.6336511264767343e-05,
+      "loss": 0.2557,
+      "step": 915
+    },
+    {
+      "epoch": 2.5068119891008176,
+      "grad_norm": 1.1849719285964966,
+      "learning_rate": 1.620181607148789e-05,
+      "loss": 0.2264,
+      "step": 920
+    },
+    {
+      "epoch": 2.5204359673024523,
+      "grad_norm": 1.1594517230987549,
+      "learning_rate": 1.6067023271414483e-05,
+      "loss": 0.2559,
+      "step": 925
+    },
+    {
+      "epoch": 2.534059945504087,
+      "grad_norm": 1.2838243246078491,
+      "learning_rate": 1.5932143811890384e-05,
+      "loss": 0.2365,
+      "step": 930
+    },
+    {
+      "epoch": 2.547683923705722,
+      "grad_norm": 1.1414204835891724,
+      "learning_rate": 1.579718864729699e-05,
+      "loss": 0.2515,
+      "step": 935
+    },
+    {
+      "epoch": 2.561307901907357,
+      "grad_norm": 1.1180287599563599,
+      "learning_rate": 1.5662168738164158e-05,
+      "loss": 0.2196,
+      "step": 940
+    },
+    {
+      "epoch": 2.5749318801089918,
+      "grad_norm": 1.2409911155700684,
+      "learning_rate": 1.5527095050280054e-05,
+      "loss": 0.2345,
+      "step": 945
+    },
+    {
+      "epoch": 2.5885558583106265,
+      "grad_norm": 1.1334092617034912,
+      "learning_rate": 1.539197855380054e-05,
+      "loss": 0.2052,
+      "step": 950
+    },
+    {
+      "epoch": 2.6021798365122617,
+      "grad_norm": 1.0595248937606812,
+      "learning_rate": 1.5256830222358224e-05,
+      "loss": 0.22,
+      "step": 955
+    },
+    {
+      "epoch": 2.6158038147138964,
+      "grad_norm": 1.116734266281128,
+      "learning_rate": 1.5121661032171224e-05,
+      "loss": 0.22,
+      "step": 960
+    },
+    {
+      "epoch": 2.629427792915531,
+      "grad_norm": 1.1290299892425537,
+      "learning_rate": 1.4986481961151717e-05,
+      "loss": 0.2258,
+      "step": 965
+    },
+    {
+      "epoch": 2.6430517711171664,
+      "grad_norm": 1.305019497871399,
+      "learning_rate": 1.4851303988014384e-05,
+      "loss": 0.204,
+      "step": 970
+    },
+    {
+      "epoch": 2.656675749318801,
+      "grad_norm": 0.9618856906890869,
+      "learning_rate": 1.4716138091384718e-05,
+      "loss": 0.2276,
+      "step": 975
+    },
+    {
+      "epoch": 2.670299727520436,
+      "grad_norm": 1.3820059299468994,
+      "learning_rate": 1.458099524890741e-05,
+      "loss": 0.2135,
+      "step": 980
+    },
+    {
+      "epoch": 2.683923705722071,
+      "grad_norm": 1.2182809114456177,
+      "learning_rate": 1.4445886436354793e-05,
+      "loss": 0.202,
+      "step": 985
+    },
+    {
+      "epoch": 2.697547683923706,
+      "grad_norm": 1.1048258543014526,
+      "learning_rate": 1.4310822626735414e-05,
+      "loss": 0.2486,
+      "step": 990
+    },
+    {
+      "epoch": 2.7111716621253406,
+      "grad_norm": 1.3930188417434692,
+      "learning_rate": 1.417581478940286e-05,
+      "loss": 0.2226,
+      "step": 995
+    },
+    {
+      "epoch": 2.7247956403269757,
+      "grad_norm": 1.1179946660995483,
+      "learning_rate": 1.4040873889164875e-05,
+      "loss": 0.1894,
+      "step": 1000
+    },
+    {
+      "epoch": 2.7384196185286105,
+      "grad_norm": 1.052364706993103,
+      "learning_rate": 1.3906010885392814e-05,
+      "loss": 0.2116,
+      "step": 1005
+    },
+    {
+      "epoch": 2.7520435967302452,
+      "grad_norm": 1.10884690284729,
+      "learning_rate": 1.3771236731131614e-05,
+      "loss": 0.1976,
+      "step": 1010
+    },
+    {
+      "epoch": 2.76566757493188,
+      "grad_norm": 1.083383560180664,
+      "learning_rate": 1.3636562372210186e-05,
+      "loss": 0.1818,
+      "step": 1015
+    },
+    {
+      "epoch": 2.7792915531335147,
+      "grad_norm": 1.0200647115707397,
+      "learning_rate": 1.3501998746352446e-05,
+      "loss": 0.219,
+      "step": 1020
+    },
+    {
+      "epoch": 2.79291553133515,
+      "grad_norm": 1.2136708498001099,
+      "learning_rate": 1.336755678228903e-05,
+      "loss": 0.2042,
+      "step": 1025
+    },
+    {
+      "epoch": 2.8065395095367847,
+      "grad_norm": 0.9672884345054626,
+      "learning_rate": 1.3233247398869654e-05,
+      "loss": 0.1924,
+      "step": 1030
+    },
+    {
+      "epoch": 2.8201634877384194,
+      "grad_norm": 1.2369074821472168,
+      "learning_rate": 1.3099081504176371e-05,
+      "loss": 0.2137,
+      "step": 1035
+    },
+    {
+      "epoch": 2.8337874659400546,
+      "grad_norm": 1.1135735511779785,
+      "learning_rate": 1.2965069994637638e-05,
+      "loss": 0.1952,
+      "step": 1040
+    },
+    {
+      "epoch": 2.8474114441416893,
+      "grad_norm": 1.032332181930542,
+      "learning_rate": 1.283122375414334e-05,
+      "loss": 0.2085,
+      "step": 1045
+    },
+    {
+      "epoch": 2.861035422343324,
+      "grad_norm": 1.1367651224136353,
+      "learning_rate": 1.2697553653160882e-05,
+      "loss": 0.1865,
+      "step": 1050
+    },
+    {
+      "epoch": 2.8746594005449593,
+      "grad_norm": 1.225988507270813,
+      "learning_rate": 1.2564070547852284e-05,
+      "loss": 0.1882,
+      "step": 1055
+    },
+    {
+      "epoch": 2.888283378746594,
+      "grad_norm": 0.9921906590461731,
+      "learning_rate": 1.2430785279192532e-05,
+      "loss": 0.174,
+      "step": 1060
+    },
+    {
+      "epoch": 2.9019073569482288,
+      "grad_norm": 1.1808017492294312,
+      "learning_rate": 1.2297708672089076e-05,
+      "loss": 0.1945,
+      "step": 1065
+    },
+    {
+      "epoch": 2.915531335149864,
+      "grad_norm": 1.0186188220977783,
+      "learning_rate": 1.2164851534502696e-05,
+      "loss": 0.1918,
+      "step": 1070
+    },
+    {
+      "epoch": 2.9291553133514987,
+      "grad_norm": 1.0717378854751587,
+      "learning_rate": 1.2032224656569727e-05,
+      "loss": 0.2062,
+      "step": 1075
+    },
+    {
+      "epoch": 2.9427792915531334,
+      "grad_norm": 1.1200538873672485,
+      "learning_rate": 1.1899838809725706e-05,
+      "loss": 0.2039,
+      "step": 1080
+    },
+    {
+      "epoch": 2.9564032697547686,
+      "grad_norm": 1.117672085762024,
+      "learning_rate": 1.1767704745830572e-05,
+      "loss": 0.1703,
+      "step": 1085
+    },
+    {
+      "epoch": 2.9700272479564034,
+      "grad_norm": 1.2501522302627563,
+      "learning_rate": 1.1635833196295449e-05,
+      "loss": 0.1942,
+      "step": 1090
+    },
+    {
+      "epoch": 2.983651226158038,
+      "grad_norm": 1.1326488256454468,
+      "learning_rate": 1.150423487121105e-05,
+      "loss": 0.2035,
+      "step": 1095
+    },
+    {
+      "epoch": 2.997275204359673,
+      "grad_norm": 1.0389609336853027,
+      "learning_rate": 1.1372920458477898e-05,
+      "loss": 0.1957,
+      "step": 1100
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1835,
+  "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": 1.6145624950311485e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

15_128_e5_3e-5/checkpoint-1101/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31203962b61a3a708831e2459d0fa5f30ca4a29718cce8ad449f65945bdf035e
+size 7736

15_128_e5_3e-5/checkpoint-1101/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)

15_128_e5_3e-5/checkpoint-1468/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

15_128_e5_3e-5/checkpoint-1468/adapter_config.json

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

15_128_e5_3e-5/checkpoint-1468/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57c4b16f1d505d6d6f45a65b5d75fcd3bbfbcd1bed8c78a71ed015a574c734f2
+size 791751704

15_128_e5_3e-5/checkpoint-1468/latest

@@ -0,0 +1 @@
+global_step1468

15_128_e5_3e-5/checkpoint-1468/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:674f45b5a0a440606fdd2999fcb522bc8ab1970ad4656c791739fdb1e7dbde44
+size 15920

15_128_e5_3e-5/checkpoint-1468/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5529b58046b905f1507f57e6c20005db6eb838b6550379b5a7f110825a1d137b
+size 15920

15_128_e5_3e-5/checkpoint-1468/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ff5677a2a47dedcddc5468df749f0655f90ab884c56bca29792f9beb5295277b
+size 15920

15_128_e5_3e-5/checkpoint-1468/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2fe4112d73b60b0d8c6247bf78f913796ebd2d4f9697b842d78ef51e48e215b4
+size 15920

15_128_e5_3e-5/checkpoint-1468/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ed2117280794d8387226b06f8681b9931ec9c008363ba42a0e3edf96f1988be
+size 15920

15_128_e5_3e-5/checkpoint-1468/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b824ee23388433d62ccbdb5239ed8fff7ab900064d35ecb9ca198a3813637bfe
+size 15920

15_128_e5_3e-5/checkpoint-1468/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3cd7aa640a002a2deb1dc80b99bc7ac08a33c44f54bc572f7d2d8a69bf0f580
+size 15920

15_128_e5_3e-5/checkpoint-1468/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:662b803f4757116cca05f3650985a6abd888020d045a9c99ab3f909a3cd878f4
+size 15920

15_128_e5_3e-5/checkpoint-1468/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1871ef69b88d29dca4b0fbe67f7159b27ab5ca373cd463b025af6a6cde8bb937
+size 1064

15_128_e5_3e-5/checkpoint-1468/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
+  }
+}

15_128_e5_3e-5/checkpoint-1468/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
+}

15_128_e5_3e-5/checkpoint-1468/trainer_state.json

@@ -0,0 +1,2085 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1468,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.013623978201634877,
+      "grad_norm": 1.2644609212875366,
+      "learning_rate": 1.3043478260869566e-06,
+      "loss": 1.3371,
+      "step": 5
+    },
+    {
+      "epoch": 0.027247956403269755,
+      "grad_norm": 0.9776316285133362,
+      "learning_rate": 2.9347826086956523e-06,
+      "loss": 1.3342,
+      "step": 10
+    },
+    {
+      "epoch": 0.04087193460490463,
+      "grad_norm": 0.8481394052505493,
+      "learning_rate": 4.565217391304348e-06,
+      "loss": 1.3057,
+      "step": 15
+    },
+    {
+      "epoch": 0.05449591280653951,
+      "grad_norm": 0.6515122652053833,
+      "learning_rate": 6.195652173913044e-06,
+      "loss": 1.2589,
+      "step": 20
+    },
+    {
+      "epoch": 0.0681198910081744,
+      "grad_norm": 0.5874186158180237,
+      "learning_rate": 7.826086956521738e-06,
+      "loss": 1.3107,
+      "step": 25
+    },
+    {
+      "epoch": 0.08174386920980926,
+      "grad_norm": 0.5392442345619202,
+      "learning_rate": 9.456521739130436e-06,
+      "loss": 1.2746,
+      "step": 30
+    },
+    {
+      "epoch": 0.09536784741144415,
+      "grad_norm": 0.47512176632881165,
+      "learning_rate": 1.108695652173913e-05,
+      "loss": 1.187,
+      "step": 35
+    },
+    {
+      "epoch": 0.10899182561307902,
+      "grad_norm": 0.4385203421115875,
+      "learning_rate": 1.2717391304347827e-05,
+      "loss": 1.2059,
+      "step": 40
+    },
+    {
+      "epoch": 0.1226158038147139,
+      "grad_norm": 0.6362432241439819,
+      "learning_rate": 1.4347826086956522e-05,
+      "loss": 1.1893,
+      "step": 45
+    },
+    {
+      "epoch": 0.1362397820163488,
+      "grad_norm": 0.46012377738952637,
+      "learning_rate": 1.597826086956522e-05,
+      "loss": 1.2839,
+      "step": 50
+    },
+    {
+      "epoch": 0.14986376021798364,
+      "grad_norm": 0.5726909637451172,
+      "learning_rate": 1.7608695652173915e-05,
+      "loss": 1.1741,
+      "step": 55
+    },
+    {
+      "epoch": 0.16348773841961853,
+      "grad_norm": 0.4440523684024811,
+      "learning_rate": 1.9239130434782607e-05,
+      "loss": 1.2477,
+      "step": 60
+    },
+    {
+      "epoch": 0.1771117166212534,
+      "grad_norm": 0.4722035229206085,
+      "learning_rate": 2.0869565217391306e-05,
+      "loss": 1.1941,
+      "step": 65
+    },
+    {
+      "epoch": 0.1907356948228883,
+      "grad_norm": 0.45267796516418457,
+      "learning_rate": 2.25e-05,
+      "loss": 1.2442,
+      "step": 70
+    },
+    {
+      "epoch": 0.20435967302452315,
+      "grad_norm": 0.5661253929138184,
+      "learning_rate": 2.4130434782608697e-05,
+      "loss": 1.2069,
+      "step": 75
+    },
+    {
+      "epoch": 0.21798365122615804,
+      "grad_norm": 0.46782588958740234,
+      "learning_rate": 2.5760869565217392e-05,
+      "loss": 1.1675,
+      "step": 80
+    },
+    {
+      "epoch": 0.23160762942779292,
+      "grad_norm": 0.48947134613990784,
+      "learning_rate": 2.7391304347826085e-05,
+      "loss": 1.1297,
+      "step": 85
+    },
+    {
+      "epoch": 0.2452316076294278,
+      "grad_norm": 0.5822666883468628,
+      "learning_rate": 2.9021739130434783e-05,
+      "loss": 1.1361,
+      "step": 90
+    },
+    {
+      "epoch": 0.25885558583106266,
+      "grad_norm": 0.48768264055252075,
+      "learning_rate": 2.9999902540146195e-05,
+      "loss": 1.1555,
+      "step": 95
+    },
+    {
+      "epoch": 0.2724795640326976,
+      "grad_norm": 0.5380802750587463,
+      "learning_rate": 2.999880613133526e-05,
+      "loss": 1.1335,
+      "step": 100
+    },
+    {
+      "epoch": 0.28610354223433243,
+      "grad_norm": 0.5643327832221985,
+      "learning_rate": 2.9996491578238983e-05,
+      "loss": 1.0905,
+      "step": 105
+    },
+    {
+      "epoch": 0.2997275204359673,
+      "grad_norm": 0.5208688974380493,
+      "learning_rate": 2.9992959068836304e-05,
+      "loss": 1.1015,
+      "step": 110
+    },
+    {
+      "epoch": 0.3133514986376022,
+      "grad_norm": 0.5542858839035034,
+      "learning_rate": 2.99882088900238e-05,
+      "loss": 1.0951,
+      "step": 115
+    },
+    {
+      "epoch": 0.32697547683923706,
+      "grad_norm": 0.5248655080795288,
+      "learning_rate": 2.9982241427592387e-05,
+      "loss": 1.0697,
+      "step": 120
+    },
+    {
+      "epoch": 0.3405994550408719,
+      "grad_norm": 0.6014358401298523,
+      "learning_rate": 2.997505716619599e-05,
+      "loss": 1.0873,
+      "step": 125
+    },
+    {
+      "epoch": 0.3542234332425068,
+      "grad_norm": 0.6086240410804749,
+      "learning_rate": 2.996665668931218e-05,
+      "loss": 1.0565,
+      "step": 130
+    },
+    {
+      "epoch": 0.3678474114441417,
+      "grad_norm": 0.6661133766174316,
+      "learning_rate": 2.9957040679194782e-05,
+      "loss": 1.0233,
+      "step": 135
+    },
+    {
+      "epoch": 0.3814713896457766,
+      "grad_norm": 0.707312285900116,
+      "learning_rate": 2.9946209916818477e-05,
+      "loss": 1.0613,
+      "step": 140
+    },
+    {
+      "epoch": 0.39509536784741145,
+      "grad_norm": 0.5628401041030884,
+      "learning_rate": 2.9934165281815363e-05,
+      "loss": 0.9882,
+      "step": 145
+    },
+    {
+      "epoch": 0.4087193460490463,
+      "grad_norm": 0.6965730786323547,
+      "learning_rate": 2.9920907752403513e-05,
+      "loss": 0.9984,
+      "step": 150
+    },
+    {
+      "epoch": 0.4223433242506812,
+      "grad_norm": 0.7086770534515381,
+      "learning_rate": 2.9906438405307548e-05,
+      "loss": 0.9605,
+      "step": 155
+    },
+    {
+      "epoch": 0.4359673024523161,
+      "grad_norm": 0.6676215529441833,
+      "learning_rate": 2.989075841567115e-05,
+      "loss": 1.0023,
+      "step": 160
+    },
+    {
+      "epoch": 0.44959128065395093,
+      "grad_norm": 0.8493645191192627,
+      "learning_rate": 2.987386905696167e-05,
+      "loss": 0.9039,
+      "step": 165
+    },
+    {
+      "epoch": 0.46321525885558584,
+      "grad_norm": 0.7191267609596252,
+      "learning_rate": 2.9855771700866665e-05,
+      "loss": 0.9652,
+      "step": 170
+    },
+    {
+      "epoch": 0.4768392370572207,
+      "grad_norm": 0.7579267621040344,
+      "learning_rate": 2.983646781718251e-05,
+      "loss": 0.9237,
+      "step": 175
+    },
+    {
+      "epoch": 0.4904632152588556,
+      "grad_norm": 0.8027481436729431,
+      "learning_rate": 2.9815958973695034e-05,
+      "loss": 0.9653,
+      "step": 180
+    },
+    {
+      "epoch": 0.5040871934604905,
+      "grad_norm": 0.8199446797370911,
+      "learning_rate": 2.9794246836052167e-05,
+      "loss": 0.9864,
+      "step": 185
+    },
+    {
+      "epoch": 0.5177111716621253,
+      "grad_norm": 0.7790918946266174,
+      "learning_rate": 2.977133316762869e-05,
+      "loss": 0.9403,
+      "step": 190
+    },
+    {
+      "epoch": 0.5313351498637602,
+      "grad_norm": 0.7630456686019897,
+      "learning_rate": 2.9747219829382997e-05,
+      "loss": 0.9161,
+      "step": 195
+    },
+    {
+      "epoch": 0.5449591280653951,
+      "grad_norm": 0.8171895742416382,
+      "learning_rate": 2.9721908779705974e-05,
+      "loss": 0.8556,
+      "step": 200
+    },
+    {
+      "epoch": 0.55858310626703,
+      "grad_norm": 0.8095611929893494,
+      "learning_rate": 2.969540207426193e-05,
+      "loss": 0.9051,
+      "step": 205
+    },
+    {
+      "epoch": 0.5722070844686649,
+      "grad_norm": 0.7118027806282043,
+      "learning_rate": 2.9667701865821666e-05,
+      "loss": 0.9304,
+      "step": 210
+    },
+    {
+      "epoch": 0.5858310626702997,
+      "grad_norm": 0.897177517414093,
+      "learning_rate": 2.9638810404087603e-05,
+      "loss": 0.87,
+      "step": 215
+    },
+    {
+      "epoch": 0.5994550408719346,
+      "grad_norm": 0.7748181819915771,
+      "learning_rate": 2.960873003551111e-05,
+      "loss": 0.8723,
+      "step": 220
+    },
+    {
+      "epoch": 0.6130790190735694,
+      "grad_norm": 0.8167896270751953,
+      "learning_rate": 2.9577463203101897e-05,
+      "loss": 0.8648,
+      "step": 225
+    },
+    {
+      "epoch": 0.6267029972752044,
+      "grad_norm": 0.8939515352249146,
+      "learning_rate": 2.9545012446229613e-05,
+      "loss": 0.7818,
+      "step": 230
+    },
+    {
+      "epoch": 0.6403269754768393,
+      "grad_norm": 0.7820625305175781,
+      "learning_rate": 2.951138040041764e-05,
+      "loss": 0.9032,
+      "step": 235
+    },
+    {
+      "epoch": 0.6539509536784741,
+      "grad_norm": 0.844601571559906,
+      "learning_rate": 2.9476569797129e-05,
+      "loss": 0.8736,
+      "step": 240
+    },
+    {
+      "epoch": 0.667574931880109,
+      "grad_norm": 0.7247692942619324,
+      "learning_rate": 2.944058346354454e-05,
+      "loss": 0.8779,
+      "step": 245
+    },
+    {
+      "epoch": 0.6811989100817438,
+      "grad_norm": 0.8070963621139526,
+      "learning_rate": 2.9403424322333326e-05,
+      "loss": 0.8503,
+      "step": 250
+    },
+    {
+      "epoch": 0.6948228882833788,
+      "grad_norm": 0.8530706167221069,
+      "learning_rate": 2.9365095391415254e-05,
+      "loss": 0.8546,
+      "step": 255
+    },
+    {
+      "epoch": 0.7084468664850136,
+      "grad_norm": 0.8497718572616577,
+      "learning_rate": 2.932559978371596e-05,
+      "loss": 0.78,
+      "step": 260
+    },
+    {
+      "epoch": 0.7220708446866485,
+      "grad_norm": 0.9066139459609985,
+      "learning_rate": 2.928494070691401e-05,
+      "loss": 0.829,
+      "step": 265
+    },
+    {
+      "epoch": 0.7356948228882834,
+      "grad_norm": 0.7985149621963501,
+      "learning_rate": 2.9243121463180362e-05,
+      "loss": 0.8262,
+      "step": 270
+    },
+    {
+      "epoch": 0.7493188010899182,
+      "grad_norm": 1.0057857036590576,
+      "learning_rate": 2.9200145448910184e-05,
+      "loss": 0.7681,
+      "step": 275
+    },
+    {
+      "epoch": 0.7629427792915532,
+      "grad_norm": 0.9126638770103455,
+      "learning_rate": 2.915601615444703e-05,
+      "loss": 0.8171,
+      "step": 280
+    },
+    {
+      "epoch": 0.776566757493188,
+      "grad_norm": 0.8743240237236023,
+      "learning_rate": 2.9110737163799347e-05,
+      "loss": 0.7672,
+      "step": 285
+    },
+    {
+      "epoch": 0.7901907356948229,
+      "grad_norm": 0.8928736448287964,
+      "learning_rate": 2.9064312154349395e-05,
+      "loss": 0.7824,
+      "step": 290
+    },
+    {
+      "epoch": 0.8038147138964578,
+      "grad_norm": 1.0869660377502441,
+      "learning_rate": 2.9016744896554606e-05,
+      "loss": 0.7687,
+      "step": 295
+    },
+    {
+      "epoch": 0.8174386920980926,
+      "grad_norm": 0.9520823955535889,
+      "learning_rate": 2.8968039253641347e-05,
+      "loss": 0.7603,
+      "step": 300
+    },
+    {
+      "epoch": 0.8310626702997275,
+      "grad_norm": 0.9313263297080994,
+      "learning_rate": 2.8918199181291154e-05,
+      "loss": 0.7344,
+      "step": 305
+    },
+    {
+      "epoch": 0.8446866485013624,
+      "grad_norm": 0.8860709071159363,
+      "learning_rate": 2.8867228727319484e-05,
+      "loss": 0.7221,
+      "step": 310
+    },
+    {
+      "epoch": 0.8583106267029973,
+      "grad_norm": 1.0076904296875,
+      "learning_rate": 2.8815132031346967e-05,
+      "loss": 0.7163,
+      "step": 315
+    },
+    {
+      "epoch": 0.8719346049046321,
+      "grad_norm": 1.0047346353530884,
+      "learning_rate": 2.8761913324463193e-05,
+      "loss": 0.731,
+      "step": 320
+    },
+    {
+      "epoch": 0.885558583106267,
+      "grad_norm": 0.9905893802642822,
+      "learning_rate": 2.8707576928883083e-05,
+      "loss": 0.7234,
+      "step": 325
+    },
+    {
+      "epoch": 0.8991825613079019,
+      "grad_norm": 0.8848779797554016,
+      "learning_rate": 2.8652127257595852e-05,
+      "loss": 0.7241,
+      "step": 330
+    },
+    {
+      "epoch": 0.9128065395095368,
+      "grad_norm": 1.086945652961731,
+      "learning_rate": 2.8595568814006618e-05,
+      "loss": 0.7474,
+      "step": 335
+    },
+    {
+      "epoch": 0.9264305177111717,
+      "grad_norm": 1.058239221572876,
+      "learning_rate": 2.853790619157063e-05,
+      "loss": 0.6512,
+      "step": 340
+    },
+    {
+      "epoch": 0.9400544959128065,
+      "grad_norm": 0.9978867173194885,
+      "learning_rate": 2.8479144073420237e-05,
+      "loss": 0.6968,
+      "step": 345
+    },
+    {
+      "epoch": 0.9536784741144414,
+      "grad_norm": 0.9755434989929199,
+      "learning_rate": 2.841928723198449e-05,
+      "loss": 0.6774,
+      "step": 350
+    },
+    {
+      "epoch": 0.9673024523160763,
+      "grad_norm": 0.9939659833908081,
+      "learning_rate": 2.835834052860162e-05,
+      "loss": 0.6851,
+      "step": 355
+    },
+    {
+      "epoch": 0.9809264305177112,
+      "grad_norm": 0.919965922832489,
+      "learning_rate": 2.8296308913124137e-05,
+      "loss": 0.6636,
+      "step": 360
+    },
+    {
+      "epoch": 0.9945504087193461,
+      "grad_norm": 1.1084834337234497,
+      "learning_rate": 2.8233197423516885e-05,
+      "loss": 0.696,
+      "step": 365
+    },
+    {
+      "epoch": 1.008174386920981,
+      "grad_norm": 1.0081297159194946,
+      "learning_rate": 2.816901118544785e-05,
+      "loss": 0.6079,
+      "step": 370
+    },
+    {
+      "epoch": 1.021798365122616,
+      "grad_norm": 0.9290661811828613,
+      "learning_rate": 2.810375541187188e-05,
+      "loss": 0.5794,
+      "step": 375
+    },
+    {
+      "epoch": 1.0354223433242506,
+      "grad_norm": 1.0772560834884644,
+      "learning_rate": 2.80374354026073e-05,
+      "loss": 0.5495,
+      "step": 380
+    },
+    {
+      "epoch": 1.0490463215258856,
+      "grad_norm": 1.0354938507080078,
+      "learning_rate": 2.79700565439055e-05,
+      "loss": 0.6109,
+      "step": 385
+    },
+    {
+      "epoch": 1.0626702997275204,
+      "grad_norm": 1.0381073951721191,
+      "learning_rate": 2.7901624308013465e-05,
+      "loss": 0.5849,
+      "step": 390
+    },
+    {
+      "epoch": 1.0762942779291553,
+      "grad_norm": 1.146996021270752,
+      "learning_rate": 2.7832144252729354e-05,
+      "loss": 0.5798,
+      "step": 395
+    },
+    {
+      "epoch": 1.0899182561307903,
+      "grad_norm": 1.0819785594940186,
+      "learning_rate": 2.776162202095111e-05,
+      "loss": 0.584,
+      "step": 400
+    },
+    {
+      "epoch": 1.103542234332425,
+      "grad_norm": 1.090579628944397,
+      "learning_rate": 2.7690063340218173e-05,
+      "loss": 0.556,
+      "step": 405
+    },
+    {
+      "epoch": 1.11716621253406,
+      "grad_norm": 1.034417748451233,
+      "learning_rate": 2.7617474022246297e-05,
+      "loss": 0.5788,
+      "step": 410
+    },
+    {
+      "epoch": 1.1307901907356948,
+      "grad_norm": 1.1490421295166016,
+      "learning_rate": 2.7543859962455576e-05,
+      "loss": 0.5576,
+      "step": 415
+    },
+    {
+      "epoch": 1.1444141689373297,
+      "grad_norm": 1.0201855897903442,
+      "learning_rate": 2.7469227139491603e-05,
+      "loss": 0.5886,
+      "step": 420
+    },
+    {
+      "epoch": 1.1580381471389645,
+      "grad_norm": 1.01875901222229,
+      "learning_rate": 2.7393581614739924e-05,
+      "loss": 0.6062,
+      "step": 425
+    },
+    {
+      "epoch": 1.1716621253405994,
+      "grad_norm": 1.0762981176376343,
+      "learning_rate": 2.7316929531833775e-05,
+      "loss": 0.5389,
+      "step": 430
+    },
+    {
+      "epoch": 1.1852861035422344,
+      "grad_norm": 1.0140329599380493,
+      "learning_rate": 2.7239277116155077e-05,
+      "loss": 0.5462,
+      "step": 435
+    },
+    {
+      "epoch": 1.1989100817438691,
+      "grad_norm": 1.1165515184402466,
+      "learning_rate": 2.7160630674328893e-05,
+      "loss": 0.5596,
+      "step": 440
+    },
+    {
+      "epoch": 1.2125340599455041,
+      "grad_norm": 1.0689040422439575,
+      "learning_rate": 2.7080996593711172e-05,
+      "loss": 0.5137,
+      "step": 445
+    },
+    {
+      "epoch": 1.226158038147139,
+      "grad_norm": 1.2397655248641968,
+      "learning_rate": 2.700038134187002e-05,
+      "loss": 0.5643,
+      "step": 450
+    },
+    {
+      "epoch": 1.2397820163487738,
+      "grad_norm": 0.9885849952697754,
+      "learning_rate": 2.691879146606043e-05,
+      "loss": 0.5921,
+      "step": 455
+    },
+    {
+      "epoch": 1.2534059945504088,
+      "grad_norm": 1.279818058013916,
+      "learning_rate": 2.6836233592692544e-05,
+      "loss": 0.5126,
+      "step": 460
+    },
+    {
+      "epoch": 1.2670299727520435,
+      "grad_norm": 1.1189254522323608,
+      "learning_rate": 2.675271442679346e-05,
+      "loss": 0.5198,
+      "step": 465
+    },
+    {
+      "epoch": 1.2806539509536785,
+      "grad_norm": 1.1345500946044922,
+      "learning_rate": 2.6668240751462707e-05,
+      "loss": 0.5117,
+      "step": 470
+    },
+    {
+      "epoch": 1.2942779291553133,
+      "grad_norm": 1.1066192388534546,
+      "learning_rate": 2.6582819427321313e-05,
+      "loss": 0.5314,
+      "step": 475
+    },
+    {
+      "epoch": 1.3079019073569482,
+      "grad_norm": 1.0839710235595703,
+      "learning_rate": 2.649645739195464e-05,
+      "loss": 0.5382,
+      "step": 480
+    },
+    {
+      "epoch": 1.3215258855585832,
+      "grad_norm": 1.052043080329895,
+      "learning_rate": 2.640916165934893e-05,
+      "loss": 0.5135,
+      "step": 485
+    },
+    {
+      "epoch": 1.335149863760218,
+      "grad_norm": 1.1333352327346802,
+      "learning_rate": 2.6320939319321657e-05,
+      "loss": 0.5359,
+      "step": 490
+    },
+    {
+      "epoch": 1.348773841961853,
+      "grad_norm": 1.21845543384552,
+      "learning_rate": 2.623179753694573e-05,
+      "loss": 0.4853,
+      "step": 495
+    },
+    {
+      "epoch": 1.3623978201634879,
+      "grad_norm": 1.142262578010559,
+      "learning_rate": 2.614174355196754e-05,
+      "loss": 0.4993,
+      "step": 500
+    },
+    {
+      "epoch": 1.3760217983651226,
+      "grad_norm": 1.037720799446106,
+      "learning_rate": 2.6050784678219024e-05,
+      "loss": 0.512,
+      "step": 505
+    },
+    {
+      "epoch": 1.3896457765667574,
+      "grad_norm": 1.1667251586914062,
+      "learning_rate": 2.5958928303023634e-05,
+      "loss": 0.4788,
+      "step": 510
+    },
+    {
+      "epoch": 1.4032697547683923,
+      "grad_norm": 1.046761393547058,
+      "learning_rate": 2.5866181886596367e-05,
+      "loss": 0.4867,
+      "step": 515
+    },
+    {
+      "epoch": 1.4168937329700273,
+      "grad_norm": 1.1029340028762817,
+      "learning_rate": 2.5772552961437893e-05,
+      "loss": 0.4799,
+      "step": 520
+    },
+    {
+      "epoch": 1.430517711171662,
+      "grad_norm": 1.1235623359680176,
+      "learning_rate": 2.5678049131722772e-05,
+      "loss": 0.4752,
+      "step": 525
+    },
+    {
+      "epoch": 1.444141689373297,
+      "grad_norm": 1.064278483390808,
+      "learning_rate": 2.5582678072681903e-05,
+      "loss": 0.5173,
+      "step": 530
+    },
+    {
+      "epoch": 1.457765667574932,
+      "grad_norm": 1.13005793094635,
+      "learning_rate": 2.5486447529979136e-05,
+      "loss": 0.4451,
+      "step": 535
+    },
+    {
+      "epoch": 1.4713896457765667,
+      "grad_norm": 1.0552195310592651,
+      "learning_rate": 2.5389365319082226e-05,
+      "loss": 0.4595,
+      "step": 540
+    },
+    {
+      "epoch": 1.4850136239782017,
+      "grad_norm": 1.1078029870986938,
+      "learning_rate": 2.5291439324628084e-05,
+      "loss": 0.4693,
+      "step": 545
+    },
+    {
+      "epoch": 1.4986376021798364,
+      "grad_norm": 1.090710997581482,
+      "learning_rate": 2.5192677499782413e-05,
+      "loss": 0.4537,
+      "step": 550
+    },
+    {
+      "epoch": 1.5122615803814714,
+      "grad_norm": 1.037945032119751,
+      "learning_rate": 2.5093087865593784e-05,
+      "loss": 0.4556,
+      "step": 555
+    },
+    {
+      "epoch": 1.5258855585831061,
+      "grad_norm": 1.1792460680007935,
+      "learning_rate": 2.499267851034221e-05,
+      "loss": 0.4734,
+      "step": 560
+    },
+    {
+      "epoch": 1.5395095367847411,
+      "grad_norm": 1.0996177196502686,
+      "learning_rate": 2.4891457588882238e-05,
+      "loss": 0.4444,
+      "step": 565
+    },
+    {
+      "epoch": 1.553133514986376,
+      "grad_norm": 1.13473641872406,
+      "learning_rate": 2.478943332198062e-05,
+      "loss": 0.4513,
+      "step": 570
+    },
+    {
+      "epoch": 1.5667574931880108,
+      "grad_norm": 1.212607979774475,
+      "learning_rate": 2.468661399564871e-05,
+      "loss": 0.4506,
+      "step": 575
+    },
+    {
+      "epoch": 1.5803814713896458,
+      "grad_norm": 1.0172079801559448,
+      "learning_rate": 2.458300796046946e-05,
+      "loss": 0.4238,
+      "step": 580
+    },
+    {
+      "epoch": 1.5940054495912808,
+      "grad_norm": 1.3280726671218872,
+      "learning_rate": 2.4478623630919236e-05,
+      "loss": 0.4419,
+      "step": 585
+    },
+    {
+      "epoch": 1.6076294277929155,
+      "grad_norm": 1.1108782291412354,
+      "learning_rate": 2.437346948468441e-05,
+      "loss": 0.3942,
+      "step": 590
+    },
+    {
+      "epoch": 1.6212534059945503,
+      "grad_norm": 1.157791256904602,
+      "learning_rate": 2.4267554061972873e-05,
+      "loss": 0.397,
+      "step": 595
+    },
+    {
+      "epoch": 1.6348773841961854,
+      "grad_norm": 1.1420488357543945,
+      "learning_rate": 2.416088596482039e-05,
+      "loss": 0.4849,
+      "step": 600
+    },
+    {
+      "epoch": 1.6485013623978202,
+      "grad_norm": 1.264007329940796,
+      "learning_rate": 2.405347385639202e-05,
+      "loss": 0.4365,
+      "step": 605
+    },
+    {
+      "epoch": 1.662125340599455,
+      "grad_norm": 1.0671672821044922,
+      "learning_rate": 2.394532646027848e-05,
+      "loss": 0.4259,
+      "step": 610
+    },
+    {
+      "epoch": 1.67574931880109,
+      "grad_norm": 1.245835781097412,
+      "learning_rate": 2.3836452559787673e-05,
+      "loss": 0.4078,
+      "step": 615
+    },
+    {
+      "epoch": 1.6893732970027249,
+      "grad_norm": 1.1798008680343628,
+      "learning_rate": 2.3726860997231356e-05,
+      "loss": 0.4319,
+      "step": 620
+    },
+    {
+      "epoch": 1.7029972752043596,
+      "grad_norm": 1.3666014671325684,
+      "learning_rate": 2.3616560673206984e-05,
+      "loss": 0.4645,
+      "step": 625
+    },
+    {
+      "epoch": 1.7166212534059946,
+      "grad_norm": 1.0500986576080322,
+      "learning_rate": 2.3505560545874843e-05,
+      "loss": 0.4543,
+      "step": 630
+    },
+    {
+      "epoch": 1.7302452316076296,
+      "grad_norm": 1.4002113342285156,
+      "learning_rate": 2.3393869630230495e-05,
+      "loss": 0.3978,
+      "step": 635
+    },
+    {
+      "epoch": 1.7438692098092643,
+      "grad_norm": 1.170268177986145,
+      "learning_rate": 2.3281496997372625e-05,
+      "loss": 0.4355,
+      "step": 640
+    },
+    {
+      "epoch": 1.757493188010899,
+      "grad_norm": 1.0691126585006714,
+      "learning_rate": 2.316845177376633e-05,
+      "loss": 0.4097,
+      "step": 645
+    },
+    {
+      "epoch": 1.771117166212534,
+      "grad_norm": 1.1894110441207886,
+      "learning_rate": 2.3054743140501877e-05,
+      "loss": 0.4025,
+      "step": 650
+    },
+    {
+      "epoch": 1.784741144414169,
+      "grad_norm": 1.0559704303741455,
+      "learning_rate": 2.2940380332549086e-05,
+      "loss": 0.4237,
+      "step": 655
+    },
+    {
+      "epoch": 1.7983651226158037,
+      "grad_norm": 1.1340513229370117,
+      "learning_rate": 2.282537263800727e-05,
+      "loss": 0.3523,
+      "step": 660
+    },
+    {
+      "epoch": 1.8119891008174387,
+      "grad_norm": 1.1120586395263672,
+      "learning_rate": 2.2709729397350904e-05,
+      "loss": 0.4037,
+      "step": 665
+    },
+    {
+      "epoch": 1.8256130790190737,
+      "grad_norm": 1.1562973260879517,
+      "learning_rate": 2.2593460002671024e-05,
+      "loss": 0.3839,
+      "step": 670
+    },
+    {
+      "epoch": 1.8392370572207084,
+      "grad_norm": 1.2246955633163452,
+      "learning_rate": 2.247657389691247e-05,
+      "loss": 0.4188,
+      "step": 675
+    },
+    {
+      "epoch": 1.8528610354223434,
+      "grad_norm": 0.9505524635314941,
+      "learning_rate": 2.2359080573106913e-05,
+      "loss": 0.3348,
+      "step": 680
+    },
+    {
+      "epoch": 1.8664850136239783,
+      "grad_norm": 1.161018967628479,
+      "learning_rate": 2.2240989573601902e-05,
+      "loss": 0.3978,
+      "step": 685
+    },
+    {
+      "epoch": 1.880108991825613,
+      "grad_norm": 1.076042890548706,
+      "learning_rate": 2.212231048928587e-05,
+      "loss": 0.3608,
+      "step": 690
+    },
+    {
+      "epoch": 1.8937329700272478,
+      "grad_norm": 1.248965859413147,
+      "learning_rate": 2.2003052958809185e-05,
+      "loss": 0.3621,
+      "step": 695
+    },
+    {
+      "epoch": 1.9073569482288828,
+      "grad_norm": 1.0826005935668945,
+      "learning_rate": 2.1883226667801374e-05,
+      "loss": 0.3733,
+      "step": 700
+    },
+    {
+      "epoch": 1.9209809264305178,
+      "grad_norm": 1.1231858730316162,
+      "learning_rate": 2.1762841348084425e-05,
+      "loss": 0.3916,
+      "step": 705
+    },
+    {
+      "epoch": 1.9346049046321525,
+      "grad_norm": 1.1095854043960571,
+      "learning_rate": 2.164190677688248e-05,
+      "loss": 0.3406,
+      "step": 710
+    },
+    {
+      "epoch": 1.9482288828337875,
+      "grad_norm": 1.0947611331939697,
+      "learning_rate": 2.1520432776027723e-05,
+      "loss": 0.3511,
+      "step": 715
+    },
+    {
+      "epoch": 1.9618528610354224,
+      "grad_norm": 1.035023808479309,
+      "learning_rate": 2.1398429211162706e-05,
+      "loss": 0.3894,
+      "step": 720
+    },
+    {
+      "epoch": 1.9754768392370572,
+      "grad_norm": 1.166717529296875,
+      "learning_rate": 2.127590599093909e-05,
+      "loss": 0.3692,
+      "step": 725
+    },
+    {
+      "epoch": 1.989100817438692,
+      "grad_norm": 1.1927279233932495,
+      "learning_rate": 2.1152873066212913e-05,
+      "loss": 0.3278,
+      "step": 730
+    },
+    {
+      "epoch": 2.002724795640327,
+      "grad_norm": 1.0591354370117188,
+      "learning_rate": 2.102934042923641e-05,
+      "loss": 0.3421,
+      "step": 735
+    },
+    {
+      "epoch": 2.016348773841962,
+      "grad_norm": 1.107465147972107,
+      "learning_rate": 2.0905318112846497e-05,
+      "loss": 0.274,
+      "step": 740
+    },
+    {
+      "epoch": 2.0299727520435966,
+      "grad_norm": 1.011695146560669,
+      "learning_rate": 2.0780816189649928e-05,
+      "loss": 0.2894,
+      "step": 745
+    },
+    {
+      "epoch": 2.043596730245232,
+      "grad_norm": 1.3591814041137695,
+      "learning_rate": 2.0655844771205253e-05,
+      "loss": 0.3034,
+      "step": 750
+    },
+    {
+      "epoch": 2.0572207084468666,
+      "grad_norm": 1.2014539241790771,
+      "learning_rate": 2.0530414007201566e-05,
+      "loss": 0.2508,
+      "step": 755
+    },
+    {
+      "epoch": 2.0708446866485013,
+      "grad_norm": 1.1204092502593994,
+      "learning_rate": 2.0404534084634228e-05,
+      "loss": 0.2766,
+      "step": 760
+    },
+    {
+      "epoch": 2.084468664850136,
+      "grad_norm": 1.0801162719726562,
+      "learning_rate": 2.0278215226977497e-05,
+      "loss": 0.285,
+      "step": 765
+    },
+    {
+      "epoch": 2.0980926430517712,
+      "grad_norm": 1.1138821840286255,
+      "learning_rate": 2.015146769335421e-05,
+      "loss": 0.2964,
+      "step": 770
+    },
+    {
+      "epoch": 2.111716621253406,
+      "grad_norm": 1.2337431907653809,
+      "learning_rate": 2.0024301777702586e-05,
+      "loss": 0.2889,
+      "step": 775
+    },
+    {
+      "epoch": 2.1253405994550407,
+      "grad_norm": 1.228090763092041,
+      "learning_rate": 1.9896727807940208e-05,
+      "loss": 0.2882,
+      "step": 780
+    },
+    {
+      "epoch": 2.138964577656676,
+      "grad_norm": 1.3023854494094849,
+      "learning_rate": 1.976875614512521e-05,
+      "loss": 0.2811,
+      "step": 785
+    },
+    {
+      "epoch": 2.1525885558583107,
+      "grad_norm": 1.1693427562713623,
+      "learning_rate": 1.964039718261478e-05,
+      "loss": 0.2795,
+      "step": 790
+    },
+    {
+      "epoch": 2.1662125340599454,
+      "grad_norm": 1.235745906829834,
+      "learning_rate": 1.95116613452211e-05,
+      "loss": 0.2745,
+      "step": 795
+    },
+    {
+      "epoch": 2.1798365122615806,
+      "grad_norm": 1.1393667459487915,
+      "learning_rate": 1.9382559088364636e-05,
+      "loss": 0.2928,
+      "step": 800
+    },
+    {
+      "epoch": 2.1934604904632153,
+      "grad_norm": 1.275499939918518,
+      "learning_rate": 1.9253100897225006e-05,
+      "loss": 0.2544,
+      "step": 805
+    },
+    {
+      "epoch": 2.20708446866485,
+      "grad_norm": 1.1270493268966675,
+      "learning_rate": 1.9123297285889424e-05,
+      "loss": 0.2696,
+      "step": 810
+    },
+    {
+      "epoch": 2.220708446866485,
+      "grad_norm": 1.0665929317474365,
+      "learning_rate": 1.899315879649878e-05,
+      "loss": 0.2448,
+      "step": 815
+    },
+    {
+      "epoch": 2.23433242506812,
+      "grad_norm": 1.0491502285003662,
+      "learning_rate": 1.8862695998391438e-05,
+      "loss": 0.259,
+      "step": 820
+    },
+    {
+      "epoch": 2.2479564032697548,
+      "grad_norm": 1.0268921852111816,
+      "learning_rate": 1.8731919487244855e-05,
+      "loss": 0.237,
+      "step": 825
+    },
+    {
+      "epoch": 2.2615803814713895,
+      "grad_norm": 1.1503524780273438,
+      "learning_rate": 1.8600839884215034e-05,
+      "loss": 0.2626,
+      "step": 830
+    },
+    {
+      "epoch": 2.2752043596730247,
+      "grad_norm": 1.155947208404541,
+      "learning_rate": 1.84694678350739e-05,
+      "loss": 0.2527,
+      "step": 835
+    },
+    {
+      "epoch": 2.2888283378746594,
+      "grad_norm": 1.2745076417922974,
+      "learning_rate": 1.8337814009344716e-05,
+      "loss": 0.2974,
+      "step": 840
+    },
+    {
+      "epoch": 2.302452316076294,
+      "grad_norm": 1.1524717807769775,
+      "learning_rate": 1.8205889099435543e-05,
+      "loss": 0.2676,
+      "step": 845
+    },
+    {
+      "epoch": 2.316076294277929,
+      "grad_norm": 1.2803691625595093,
+      "learning_rate": 1.807370381977082e-05,
+      "loss": 0.261,
+      "step": 850
+    },
+    {
+      "epoch": 2.329700272479564,
+      "grad_norm": 1.0910717248916626,
+      "learning_rate": 1.7941268905921208e-05,
+      "loss": 0.2543,
+      "step": 855
+    },
+    {
+      "epoch": 2.343324250681199,
+      "grad_norm": 1.127986192703247,
+      "learning_rate": 1.780859511373168e-05,
+      "loss": 0.2653,
+      "step": 860
+    },
+    {
+      "epoch": 2.3569482288828336,
+      "grad_norm": 1.1669334173202515,
+      "learning_rate": 1.7675693218447973e-05,
+      "loss": 0.2461,
+      "step": 865
+    },
+    {
+      "epoch": 2.370572207084469,
+      "grad_norm": 1.3159593343734741,
+      "learning_rate": 1.7542574013841453e-05,
+      "loss": 0.2459,
+      "step": 870
+    },
+    {
+      "epoch": 2.3841961852861036,
+      "grad_norm": 1.1560946702957153,
+      "learning_rate": 1.7409248311332503e-05,
+      "loss": 0.2403,
+      "step": 875
+    },
+    {
+      "epoch": 2.3978201634877383,
+      "grad_norm": 1.3311054706573486,
+      "learning_rate": 1.7275726939112465e-05,
+      "loss": 0.2627,
+      "step": 880
+    },
+    {
+      "epoch": 2.4114441416893735,
+      "grad_norm": 1.2287989854812622,
+      "learning_rate": 1.71420207412642e-05,
+      "loss": 0.2428,
+      "step": 885
+    },
+    {
+      "epoch": 2.4250681198910082,
+      "grad_norm": 0.9484009742736816,
+      "learning_rate": 1.7008140576881384e-05,
+      "loss": 0.2546,
+      "step": 890
+    },
+    {
+      "epoch": 2.438692098092643,
+      "grad_norm": 1.140745759010315,
+      "learning_rate": 1.6874097319186575e-05,
+      "loss": 0.2152,
+      "step": 895
+    },
+    {
+      "epoch": 2.452316076294278,
+      "grad_norm": 1.1644483804702759,
+      "learning_rate": 1.6739901854648138e-05,
+      "loss": 0.2806,
+      "step": 900
+    },
+    {
+      "epoch": 2.465940054495913,
+      "grad_norm": 1.1392990350723267,
+      "learning_rate": 1.660556508209606e-05,
+      "loss": 0.2534,
+      "step": 905
+    },
+    {
+      "epoch": 2.4795640326975477,
+      "grad_norm": 1.051167607307434,
+      "learning_rate": 1.647109791183684e-05,
+      "loss": 0.2345,
+      "step": 910
+    },
+    {
+      "epoch": 2.4931880108991824,
+      "grad_norm": 1.1279879808425903,
+      "learning_rate": 1.6336511264767343e-05,
+      "loss": 0.2557,
+      "step": 915
+    },
+    {
+      "epoch": 2.5068119891008176,
+      "grad_norm": 1.1849719285964966,
+      "learning_rate": 1.620181607148789e-05,
+      "loss": 0.2264,
+      "step": 920
+    },
+    {
+      "epoch": 2.5204359673024523,
+      "grad_norm": 1.1594517230987549,
+      "learning_rate": 1.6067023271414483e-05,
+      "loss": 0.2559,
+      "step": 925
+    },
+    {
+      "epoch": 2.534059945504087,
+      "grad_norm": 1.2838243246078491,
+      "learning_rate": 1.5932143811890384e-05,
+      "loss": 0.2365,
+      "step": 930
+    },
+    {
+      "epoch": 2.547683923705722,
+      "grad_norm": 1.1414204835891724,
+      "learning_rate": 1.579718864729699e-05,
+      "loss": 0.2515,
+      "step": 935
+    },
+    {
+      "epoch": 2.561307901907357,
+      "grad_norm": 1.1180287599563599,
+      "learning_rate": 1.5662168738164158e-05,
+      "loss": 0.2196,
+      "step": 940
+    },
+    {
+      "epoch": 2.5749318801089918,
+      "grad_norm": 1.2409911155700684,
+      "learning_rate": 1.5527095050280054e-05,
+      "loss": 0.2345,
+      "step": 945
+    },
+    {
+      "epoch": 2.5885558583106265,
+      "grad_norm": 1.1334092617034912,
+      "learning_rate": 1.539197855380054e-05,
+      "loss": 0.2052,
+      "step": 950
+    },
+    {
+      "epoch": 2.6021798365122617,
+      "grad_norm": 1.0595248937606812,
+      "learning_rate": 1.5256830222358224e-05,
+      "loss": 0.22,
+      "step": 955
+    },
+    {
+      "epoch": 2.6158038147138964,
+      "grad_norm": 1.116734266281128,
+      "learning_rate": 1.5121661032171224e-05,
+      "loss": 0.22,
+      "step": 960
+    },
+    {
+      "epoch": 2.629427792915531,
+      "grad_norm": 1.1290299892425537,
+      "learning_rate": 1.4986481961151717e-05,
+      "loss": 0.2258,
+      "step": 965
+    },
+    {
+      "epoch": 2.6430517711171664,
+      "grad_norm": 1.305019497871399,
+      "learning_rate": 1.4851303988014384e-05,
+      "loss": 0.204,
+      "step": 970
+    },
+    {
+      "epoch": 2.656675749318801,
+      "grad_norm": 0.9618856906890869,
+      "learning_rate": 1.4716138091384718e-05,
+      "loss": 0.2276,
+      "step": 975
+    },
+    {
+      "epoch": 2.670299727520436,
+      "grad_norm": 1.3820059299468994,
+      "learning_rate": 1.458099524890741e-05,
+      "loss": 0.2135,
+      "step": 980
+    },
+    {
+      "epoch": 2.683923705722071,
+      "grad_norm": 1.2182809114456177,
+      "learning_rate": 1.4445886436354793e-05,
+      "loss": 0.202,
+      "step": 985
+    },
+    {
+      "epoch": 2.697547683923706,
+      "grad_norm": 1.1048258543014526,
+      "learning_rate": 1.4310822626735414e-05,
+      "loss": 0.2486,
+      "step": 990
+    },
+    {
+      "epoch": 2.7111716621253406,
+      "grad_norm": 1.3930188417434692,
+      "learning_rate": 1.417581478940286e-05,
+      "loss": 0.2226,
+      "step": 995
+    },
+    {
+      "epoch": 2.7247956403269757,
+      "grad_norm": 1.1179946660995483,
+      "learning_rate": 1.4040873889164875e-05,
+      "loss": 0.1894,
+      "step": 1000
+    },
+    {
+      "epoch": 2.7384196185286105,
+      "grad_norm": 1.052364706993103,
+      "learning_rate": 1.3906010885392814e-05,
+      "loss": 0.2116,
+      "step": 1005
+    },
+    {
+      "epoch": 2.7520435967302452,
+      "grad_norm": 1.10884690284729,
+      "learning_rate": 1.3771236731131614e-05,
+      "loss": 0.1976,
+      "step": 1010
+    },
+    {
+      "epoch": 2.76566757493188,
+      "grad_norm": 1.083383560180664,
+      "learning_rate": 1.3636562372210186e-05,
+      "loss": 0.1818,
+      "step": 1015
+    },
+    {
+      "epoch": 2.7792915531335147,
+      "grad_norm": 1.0200647115707397,
+      "learning_rate": 1.3501998746352446e-05,
+      "loss": 0.219,
+      "step": 1020
+    },
+    {
+      "epoch": 2.79291553133515,
+      "grad_norm": 1.2136708498001099,
+      "learning_rate": 1.336755678228903e-05,
+      "loss": 0.2042,
+      "step": 1025
+    },
+    {
+      "epoch": 2.8065395095367847,
+      "grad_norm": 0.9672884345054626,
+      "learning_rate": 1.3233247398869654e-05,
+      "loss": 0.1924,
+      "step": 1030
+    },
+    {
+      "epoch": 2.8201634877384194,
+      "grad_norm": 1.2369074821472168,
+      "learning_rate": 1.3099081504176371e-05,
+      "loss": 0.2137,
+      "step": 1035
+    },
+    {
+      "epoch": 2.8337874659400546,
+      "grad_norm": 1.1135735511779785,
+      "learning_rate": 1.2965069994637638e-05,
+      "loss": 0.1952,
+      "step": 1040
+    },
+    {
+      "epoch": 2.8474114441416893,
+      "grad_norm": 1.032332181930542,
+      "learning_rate": 1.283122375414334e-05,
+      "loss": 0.2085,
+      "step": 1045
+    },
+    {
+      "epoch": 2.861035422343324,
+      "grad_norm": 1.1367651224136353,
+      "learning_rate": 1.2697553653160882e-05,
+      "loss": 0.1865,
+      "step": 1050
+    },
+    {
+      "epoch": 2.8746594005449593,
+      "grad_norm": 1.225988507270813,
+      "learning_rate": 1.2564070547852284e-05,
+      "loss": 0.1882,
+      "step": 1055
+    },
+    {
+      "epoch": 2.888283378746594,
+      "grad_norm": 0.9921906590461731,
+      "learning_rate": 1.2430785279192532e-05,
+      "loss": 0.174,
+      "step": 1060
+    },
+    {
+      "epoch": 2.9019073569482288,
+      "grad_norm": 1.1808017492294312,
+      "learning_rate": 1.2297708672089076e-05,
+      "loss": 0.1945,
+      "step": 1065
+    },
+    {
+      "epoch": 2.915531335149864,
+      "grad_norm": 1.0186188220977783,
+      "learning_rate": 1.2164851534502696e-05,
+      "loss": 0.1918,
+      "step": 1070
+    },
+    {
+      "epoch": 2.9291553133514987,
+      "grad_norm": 1.0717378854751587,
+      "learning_rate": 1.2032224656569727e-05,
+      "loss": 0.2062,
+      "step": 1075
+    },
+    {
+      "epoch": 2.9427792915531334,
+      "grad_norm": 1.1200538873672485,
+      "learning_rate": 1.1899838809725706e-05,
+      "loss": 0.2039,
+      "step": 1080
+    },
+    {
+      "epoch": 2.9564032697547686,
+      "grad_norm": 1.117672085762024,
+      "learning_rate": 1.1767704745830572e-05,
+      "loss": 0.1703,
+      "step": 1085
+    },
+    {
+      "epoch": 2.9700272479564034,
+      "grad_norm": 1.2501522302627563,
+      "learning_rate": 1.1635833196295449e-05,
+      "loss": 0.1942,
+      "step": 1090
+    },
+    {
+      "epoch": 2.983651226158038,
+      "grad_norm": 1.1326488256454468,
+      "learning_rate": 1.150423487121105e-05,
+      "loss": 0.2035,
+      "step": 1095
+    },
+    {
+      "epoch": 2.997275204359673,
+      "grad_norm": 1.0389609336853027,
+      "learning_rate": 1.1372920458477898e-05,
+      "loss": 0.1957,
+      "step": 1100
+    },
+    {
+      "epoch": 3.010899182561308,
+      "grad_norm": 1.0185716152191162,
+      "learning_rate": 1.1241900622938234e-05,
+      "loss": 0.1515,
+      "step": 1105
+    },
+    {
+      "epoch": 3.024523160762943,
+      "grad_norm": 1.0238181352615356,
+      "learning_rate": 1.111118600550993e-05,
+      "loss": 0.1421,
+      "step": 1110
+    },
+    {
+      "epoch": 3.0381471389645776,
+      "grad_norm": 0.9082461595535278,
+      "learning_rate": 1.0980787222322206e-05,
+      "loss": 0.1483,
+      "step": 1115
+    },
+    {
+      "epoch": 3.0517711171662127,
+      "grad_norm": 1.05185866355896,
+      "learning_rate": 1.085071486385347e-05,
+      "loss": 0.1451,
+      "step": 1120
+    },
+    {
+      "epoch": 3.0653950953678475,
+      "grad_norm": 0.9242850542068481,
+      "learning_rate": 1.0720979494071217e-05,
+      "loss": 0.155,
+      "step": 1125
+    },
+    {
+      "epoch": 3.0790190735694822,
+      "grad_norm": 1.018025517463684,
+      "learning_rate": 1.0591591649574e-05,
+      "loss": 0.1547,
+      "step": 1130
+    },
+    {
+      "epoch": 3.092643051771117,
+      "grad_norm": 1.2223737239837646,
+      "learning_rate": 1.0462561838735765e-05,
+      "loss": 0.155,
+      "step": 1135
+    },
+    {
+      "epoch": 3.106267029972752,
+      "grad_norm": 1.0280207395553589,
+      "learning_rate": 1.0333900540852358e-05,
+      "loss": 0.1341,
+      "step": 1140
+    },
+    {
+      "epoch": 3.119891008174387,
+      "grad_norm": 0.9579217433929443,
+      "learning_rate": 1.020561820529042e-05,
+      "loss": 0.1302,
+      "step": 1145
+    },
+    {
+      "epoch": 3.1335149863760217,
+      "grad_norm": 1.0365608930587769,
+      "learning_rate": 1.0077725250638801e-05,
+      "loss": 0.1549,
+      "step": 1150
+    },
+    {
+      "epoch": 3.147138964577657,
+      "grad_norm": 1.0444051027297974,
+      "learning_rate": 9.950232063862323e-06,
+      "loss": 0.1362,
+      "step": 1155
+    },
+    {
+      "epoch": 3.1607629427792916,
+      "grad_norm": 1.178430438041687,
+      "learning_rate": 9.82314899945824e-06,
+      "loss": 0.143,
+      "step": 1160
+    },
+    {
+      "epoch": 3.1743869209809263,
+      "grad_norm": 1.183701992034912,
+      "learning_rate": 9.696486378615275e-06,
+      "loss": 0.1629,
+      "step": 1165
+    },
+    {
+      "epoch": 3.1880108991825615,
+      "grad_norm": 1.00314462184906,
+      "learning_rate": 9.570254488375365e-06,
+      "loss": 0.1366,
+      "step": 1170
+    },
+    {
+      "epoch": 3.2016348773841963,
+      "grad_norm": 0.9827495217323303,
+      "learning_rate": 9.444463580798208e-06,
+      "loss": 0.1357,
+      "step": 1175
+    },
+    {
+      "epoch": 3.215258855585831,
+      "grad_norm": 1.160131812095642,
+      "learning_rate": 9.319123872128602e-06,
+      "loss": 0.1512,
+      "step": 1180
+    },
+    {
+      "epoch": 3.2288828337874658,
+      "grad_norm": 1.0760021209716797,
+      "learning_rate": 9.194245541966742e-06,
+      "loss": 0.1476,
+      "step": 1185
+    },
+    {
+      "epoch": 3.242506811989101,
+      "grad_norm": 1.061460256576538,
+      "learning_rate": 9.069838732441476e-06,
+      "loss": 0.1294,
+      "step": 1190
+    },
+    {
+      "epoch": 3.2561307901907357,
+      "grad_norm": 1.032152533531189,
+      "learning_rate": 8.94591354738658e-06,
+      "loss": 0.1323,
+      "step": 1195
+    },
+    {
+      "epoch": 3.2697547683923704,
+      "grad_norm": 1.059644341468811,
+      "learning_rate": 8.82248005152019e-06,
+      "loss": 0.1524,
+      "step": 1200
+    },
+    {
+      "epoch": 3.2833787465940056,
+      "grad_norm": 1.1085480451583862,
+      "learning_rate": 8.699548269627357e-06,
+      "loss": 0.1288,
+      "step": 1205
+    },
+    {
+      "epoch": 3.2970027247956404,
+      "grad_norm": 1.2114875316619873,
+      "learning_rate": 8.5771281857459e-06,
+      "loss": 0.1222,
+      "step": 1210
+    },
+    {
+      "epoch": 3.310626702997275,
+      "grad_norm": 1.0272252559661865,
+      "learning_rate": 8.455229742355524e-06,
+      "loss": 0.1535,
+      "step": 1215
+    },
+    {
+      "epoch": 3.32425068119891,
+      "grad_norm": 1.0079641342163086,
+      "learning_rate": 8.33386283957034e-06,
+      "loss": 0.1362,
+      "step": 1220
+    },
+    {
+      "epoch": 3.337874659400545,
+      "grad_norm": 1.125577449798584,
+      "learning_rate": 8.213037334334816e-06,
+      "loss": 0.1321,
+      "step": 1225
+    },
+    {
+      "epoch": 3.35149863760218,
+      "grad_norm": 1.0433320999145508,
+      "learning_rate": 8.092763039623226e-06,
+      "loss": 0.1242,
+      "step": 1230
+    },
+    {
+      "epoch": 3.3651226158038146,
+      "grad_norm": 0.8790069222450256,
+      "learning_rate": 7.973049723642672e-06,
+      "loss": 0.1281,
+      "step": 1235
+    },
+    {
+      "epoch": 3.3787465940054497,
+      "grad_norm": 0.9994122385978699,
+      "learning_rate": 7.853907109039797e-06,
+      "loss": 0.1361,
+      "step": 1240
+    },
+    {
+      "epoch": 3.3923705722070845,
+      "grad_norm": 1.1024796962738037,
+      "learning_rate": 7.735344872111066e-06,
+      "loss": 0.1246,
+      "step": 1245
+    },
+    {
+      "epoch": 3.4059945504087192,
+      "grad_norm": 0.9317172169685364,
+      "learning_rate": 7.617372642016995e-06,
+      "loss": 0.1235,
+      "step": 1250
+    },
+    {
+      "epoch": 3.4196185286103544,
+      "grad_norm": 1.159458875656128,
+      "learning_rate": 7.500000000000004e-06,
+      "loss": 0.1442,
+      "step": 1255
+    },
+    {
+      "epoch": 3.433242506811989,
+      "grad_norm": 0.9606974720954895,
+      "learning_rate": 7.383236478606351e-06,
+      "loss": 0.1334,
+      "step": 1260
+    },
+    {
+      "epoch": 3.446866485013624,
+      "grad_norm": 0.8745614886283875,
+      "learning_rate": 7.267091560911896e-06,
+      "loss": 0.1081,
+      "step": 1265
+    },
+    {
+      "epoch": 3.460490463215259,
+      "grad_norm": 1.1493163108825684,
+      "learning_rate": 7.151574679751919e-06,
+      "loss": 0.1239,
+      "step": 1270
+    },
+    {
+      "epoch": 3.474114441416894,
+      "grad_norm": 1.1217668056488037,
+      "learning_rate": 7.0366952169550605e-06,
+      "loss": 0.1266,
+      "step": 1275
+    },
+    {
+      "epoch": 3.4877384196185286,
+      "grad_norm": 0.9933749437332153,
+      "learning_rate": 6.9224625025813e-06,
+      "loss": 0.1415,
+      "step": 1280
+    },
+    {
+      "epoch": 3.5013623978201633,
+      "grad_norm": 1.0076018571853638,
+      "learning_rate": 6.80888581416427e-06,
+      "loss": 0.1217,
+      "step": 1285
+    },
+    {
+      "epoch": 3.5149863760217985,
+      "grad_norm": 1.0097330808639526,
+      "learning_rate": 6.695974375957732e-06,
+      "loss": 0.1201,
+      "step": 1290
+    },
+    {
+      "epoch": 3.5286103542234333,
+      "grad_norm": 1.1095619201660156,
+      "learning_rate": 6.583737358186442e-06,
+      "loss": 0.1282,
+      "step": 1295
+    },
+    {
+      "epoch": 3.542234332425068,
+      "grad_norm": 1.1345323324203491,
+      "learning_rate": 6.472183876301358e-06,
+      "loss": 0.1158,
+      "step": 1300
+    },
+    {
+      "epoch": 3.5558583106267028,
+      "grad_norm": 0.9935111999511719,
+      "learning_rate": 6.361322990239339e-06,
+      "loss": 0.1156,
+      "step": 1305
+    },
+    {
+      "epoch": 3.569482288828338,
+      "grad_norm": 0.9708585739135742,
+      "learning_rate": 6.25116370368732e-06,
+      "loss": 0.1375,
+      "step": 1310
+    },
+    {
+      "epoch": 3.5831062670299727,
+      "grad_norm": 1.10858952999115,
+      "learning_rate": 6.141714963351074e-06,
+      "loss": 0.1357,
+      "step": 1315
+    },
+    {
+      "epoch": 3.5967302452316074,
+      "grad_norm": 0.9845702052116394,
+      "learning_rate": 6.032985658228598e-06,
+      "loss": 0.1049,
+      "step": 1320
+    },
+    {
+      "epoch": 3.6103542234332426,
+      "grad_norm": 1.046037197113037,
+      "learning_rate": 5.924984618888156e-06,
+      "loss": 0.1188,
+      "step": 1325
+    },
+    {
+      "epoch": 3.6239782016348774,
+      "grad_norm": 0.91131192445755,
+      "learning_rate": 5.8177206167511594e-06,
+      "loss": 0.1187,
+      "step": 1330
+    },
+    {
+      "epoch": 3.637602179836512,
+      "grad_norm": 0.858351469039917,
+      "learning_rate": 5.711202363379698e-06,
+      "loss": 0.1131,
+      "step": 1335
+    },
+    {
+      "epoch": 3.6512261580381473,
+      "grad_norm": 0.8908146023750305,
+      "learning_rate": 5.6054385097691106e-06,
+      "loss": 0.1026,
+      "step": 1340
+    },
+    {
+      "epoch": 3.664850136239782,
+      "grad_norm": 1.0716880559921265,
+      "learning_rate": 5.500437645645306e-06,
+      "loss": 0.1201,
+      "step": 1345
+    },
+    {
+      "epoch": 3.678474114441417,
+      "grad_norm": 0.9851012825965881,
+      "learning_rate": 5.3962082987672004e-06,
+      "loss": 0.1225,
+      "step": 1350
+    },
+    {
+      "epoch": 3.692098092643052,
+      "grad_norm": 0.9541130661964417,
+      "learning_rate": 5.292758934234089e-06,
+      "loss": 0.104,
+      "step": 1355
+    },
+    {
+      "epoch": 3.7057220708446867,
+      "grad_norm": 0.9533129334449768,
+      "learning_rate": 5.190097953798163e-06,
+      "loss": 0.1219,
+      "step": 1360
+    },
+    {
+      "epoch": 3.7193460490463215,
+      "grad_norm": 0.9795069694519043,
+      "learning_rate": 5.088233695182136e-06,
+      "loss": 0.1235,
+      "step": 1365
+    },
+    {
+      "epoch": 3.7329700272479567,
+      "grad_norm": 0.8738570809364319,
+      "learning_rate": 4.98717443140211e-06,
+      "loss": 0.1122,
+      "step": 1370
+    },
+    {
+      "epoch": 3.7465940054495914,
+      "grad_norm": 0.9202728271484375,
+      "learning_rate": 4.886928370095645e-06,
+      "loss": 0.1159,
+      "step": 1375
+    },
+    {
+      "epoch": 3.760217983651226,
+      "grad_norm": 0.8850563764572144,
+      "learning_rate": 4.7875036528551895e-06,
+      "loss": 0.1183,
+      "step": 1380
+    },
+    {
+      "epoch": 3.773841961852861,
+      "grad_norm": 1.0772314071655273,
+      "learning_rate": 4.688908354566843e-06,
+      "loss": 0.1187,
+      "step": 1385
+    },
+    {
+      "epoch": 3.7874659400544957,
+      "grad_norm": 0.8542178869247437,
+      "learning_rate": 4.5911504827545445e-06,
+      "loss": 0.1073,
+      "step": 1390
+    },
+    {
+      "epoch": 3.801089918256131,
+      "grad_norm": 0.9948552846908569,
+      "learning_rate": 4.494237976929742e-06,
+      "loss": 0.1208,
+      "step": 1395
+    },
+    {
+      "epoch": 3.8147138964577656,
+      "grad_norm": 0.9467646479606628,
+      "learning_rate": 4.39817870794655e-06,
+      "loss": 0.1027,
+      "step": 1400
+    },
+    {
+      "epoch": 3.8283378746594003,
+      "grad_norm": 0.9915233254432678,
+      "learning_rate": 4.3029804773625525e-06,
+      "loss": 0.121,
+      "step": 1405
+    },
+    {
+      "epoch": 3.8419618528610355,
+      "grad_norm": 0.929770827293396,
+      "learning_rate": 4.208651016805154e-06,
+      "loss": 0.1215,
+      "step": 1410
+    },
+    {
+      "epoch": 3.8555858310626703,
+      "grad_norm": 0.9286534190177917,
+      "learning_rate": 4.115197987343663e-06,
+      "loss": 0.1035,
+      "step": 1415
+    },
+    {
+      "epoch": 3.869209809264305,
+      "grad_norm": 0.8719895482063293,
+      "learning_rate": 4.022628978867093e-06,
+      "loss": 0.1176,
+      "step": 1420
+    },
+    {
+      "epoch": 3.88283378746594,
+      "grad_norm": 0.9461065530776978,
+      "learning_rate": 3.930951509467714e-06,
+      "loss": 0.1012,
+      "step": 1425
+    },
+    {
+      "epoch": 3.896457765667575,
+      "grad_norm": 0.9570680856704712,
+      "learning_rate": 3.84017302483052e-06,
+      "loss": 0.1176,
+      "step": 1430
+    },
+    {
+      "epoch": 3.9100817438692097,
+      "grad_norm": 0.9735510945320129,
+      "learning_rate": 3.7503008976284486e-06,
+      "loss": 0.1095,
+      "step": 1435
+    },
+    {
+      "epoch": 3.923705722070845,
+      "grad_norm": 1.0221697092056274,
+      "learning_rate": 3.6613424269236593e-06,
+      "loss": 0.1092,
+      "step": 1440
+    },
+    {
+      "epoch": 3.9373297002724796,
+      "grad_norm": 0.9469287395477295,
+      "learning_rate": 3.5733048375747007e-06,
+      "loss": 0.0976,
+      "step": 1445
+    },
+    {
+      "epoch": 3.9509536784741144,
+      "grad_norm": 0.9500718116760254,
+      "learning_rate": 3.486195279649748e-06,
+      "loss": 0.1,
+      "step": 1450
+    },
+    {
+      "epoch": 3.9645776566757496,
+      "grad_norm": 0.7802490592002869,
+      "learning_rate": 3.400020827845895e-06,
+      "loss": 0.1066,
+      "step": 1455
+    },
+    {
+      "epoch": 3.9782016348773843,
+      "grad_norm": 0.9441335201263428,
+      "learning_rate": 3.3147884809145827e-06,
+      "loss": 0.109,
+      "step": 1460
+    },
+    {
+      "epoch": 3.991825613079019,
+      "grad_norm": 0.8844651579856873,
+      "learning_rate": 3.2305051610931836e-06,
+      "loss": 0.1144,
+      "step": 1465
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1835,
+  "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": 2.1526013219186934e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

15_128_e5_3e-5/checkpoint-1468/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31203962b61a3a708831e2459d0fa5f30ca4a29718cce8ad449f65945bdf035e
+size 7736

15_128_e5_3e-5/checkpoint-1468/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)

15_128_e5_3e-5/checkpoint-1835/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

15_128_e5_3e-5/checkpoint-1835/adapter_config.json

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

15_128_e5_3e-5/checkpoint-1835/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0f055f804c7abe041f3a4f97c633afb0db1ecd64aad456bb79fabf352ef1f60
+size 791751704

15_128_e5_3e-5/checkpoint-1835/latest

@@ -0,0 +1 @@
+global_step1835

15_128_e5_3e-5/checkpoint-1835/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5f57bf436a1b649395eb8914df6a6689563d4893ccdd1593245a614564c75d4
+size 15920

15_128_e5_3e-5/checkpoint-1835/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8187eb0b82dcff4ffc9894b9924019ee319a5b666e4b599537111099b8f864f3
+size 15920

15_128_e5_3e-5/checkpoint-1835/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59f92926ef6cc31ff58d7f597ffced63250337fea107aa6bfdaedacb24f998d2
+size 15920