Upload folder using huggingface_hub

25_128_e5_3e-5/checkpoint-1128/README.md

@@ -0,0 +1,202 @@
+---
+base_model: ibm-granite/granite-3.3-8b-base
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.15.2

25_128_e5_3e-5/checkpoint-1128/adapter_config.json

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

25_128_e5_3e-5/checkpoint-1128/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ed0a9c43481fde971b4794d254a9e4f9bcac4758c3cbb656e6dbb1362e8b92fe
+size 791751704

25_128_e5_3e-5/checkpoint-1128/latest

@@ -0,0 +1 @@
+global_step1128

25_128_e5_3e-5/checkpoint-1128/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7fddbcf463eac3002f63eae9f5cf9db656597c5eff78e4ee272c244f76744cae
+size 15920

25_128_e5_3e-5/checkpoint-1128/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3cf2e22598db382d30cc12c9431a04b60d95bfb90281831ca5e504ab015ad51
+size 15920

25_128_e5_3e-5/checkpoint-1128/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:890714f9d2328bd04f7de58522697e75cd693c470c8a2a45b563283d0c50e236
+size 15920

25_128_e5_3e-5/checkpoint-1128/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37669a4424b99f3237c946f949ac54442be3a760540b9a5623196b960eea5f1c
+size 15920

25_128_e5_3e-5/checkpoint-1128/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e811fed9f25f764bc232625916cacdb4e39adacb9bb7ba8e192fc67c50aba3f
+size 15920

25_128_e5_3e-5/checkpoint-1128/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbf432cd7c25f03a54a092a99046fcfdd4c43987848faf1767cdd53f32813b0d
+size 15920

25_128_e5_3e-5/checkpoint-1128/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05603b6ed7c9fc1e84d2f0b9a52f4d8f221aa8b73c3997dd2618cbf6dce57983
+size 15920

25_128_e5_3e-5/checkpoint-1128/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39c10470fb96dead83285904d20410d4ab712399ca37c0d429b5a13c3736560a
+size 15920

25_128_e5_3e-5/checkpoint-1128/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2bf2ca0f03eb3422276cf7428c198246659488c03cd55751242e7898c37ef3eb
+size 1064

25_128_e5_3e-5/checkpoint-1128/special_tokens_map.json

@@ -0,0 +1,45 @@
+{
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<fim_prefix>",
+    "<fim_middle>",
+    "<fim_suffix>",
+    "<fim_pad>",
+    "<filename>",
+    "<gh_stars>",
+    "<issue_start>",
+    "<issue_comment>",
+    "<issue_closed>",
+    "<jupyter_start>",
+    "<jupyter_text>",
+    "<jupyter_code>",
+    "<jupyter_output>",
+    "<empty_output>",
+    "<commit_before>",
+    "<commit_msg>",
+    "<commit_after>",
+    "<reponame>"
+  ],
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<reponame>",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

25_128_e5_3e-5/checkpoint-1128/tokenizer_config.json

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

25_128_e5_3e-5/checkpoint-1128/trainer_state.json

@@ -0,0 +1,1609 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1128,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.017761989342806393,
+      "grad_norm": 1.3582063913345337,
+      "learning_rate": 1.6901408450704227e-06,
+      "loss": 1.3155,
+      "step": 5
+    },
+    {
+      "epoch": 0.035523978685612786,
+      "grad_norm": 0.8357166647911072,
+      "learning_rate": 3.8028169014084508e-06,
+      "loss": 1.3269,
+      "step": 10
+    },
+    {
+      "epoch": 0.05328596802841918,
+      "grad_norm": 0.627669095993042,
+      "learning_rate": 5.915492957746478e-06,
+      "loss": 1.2938,
+      "step": 15
+    },
+    {
+      "epoch": 0.07104795737122557,
+      "grad_norm": 0.6404774785041809,
+      "learning_rate": 8.028169014084507e-06,
+      "loss": 1.3265,
+      "step": 20
+    },
+    {
+      "epoch": 0.08880994671403197,
+      "grad_norm": 0.5879626274108887,
+      "learning_rate": 1.0140845070422535e-05,
+      "loss": 1.2527,
+      "step": 25
+    },
+    {
+      "epoch": 0.10657193605683836,
+      "grad_norm": 0.5322818756103516,
+      "learning_rate": 1.2253521126760564e-05,
+      "loss": 1.2349,
+      "step": 30
+    },
+    {
+      "epoch": 0.12433392539964476,
+      "grad_norm": 0.4639137387275696,
+      "learning_rate": 1.436619718309859e-05,
+      "loss": 1.2609,
+      "step": 35
+    },
+    {
+      "epoch": 0.14209591474245115,
+      "grad_norm": 0.44531938433647156,
+      "learning_rate": 1.6478873239436623e-05,
+      "loss": 1.1992,
+      "step": 40
+    },
+    {
+      "epoch": 0.15985790408525755,
+      "grad_norm": 0.5321743488311768,
+      "learning_rate": 1.8591549295774646e-05,
+      "loss": 1.1719,
+      "step": 45
+    },
+    {
+      "epoch": 0.17761989342806395,
+      "grad_norm": 0.4335007071495056,
+      "learning_rate": 2.0704225352112676e-05,
+      "loss": 1.1815,
+      "step": 50
+    },
+    {
+      "epoch": 0.19538188277087035,
+      "grad_norm": 0.6063855290412903,
+      "learning_rate": 2.2816901408450703e-05,
+      "loss": 1.2039,
+      "step": 55
+    },
+    {
+      "epoch": 0.21314387211367672,
+      "grad_norm": 0.4855096936225891,
+      "learning_rate": 2.4929577464788733e-05,
+      "loss": 1.1404,
+      "step": 60
+    },
+    {
+      "epoch": 0.23090586145648312,
+      "grad_norm": 0.46861350536346436,
+      "learning_rate": 2.7042253521126763e-05,
+      "loss": 1.1783,
+      "step": 65
+    },
+    {
+      "epoch": 0.24866785079928952,
+      "grad_norm": 0.5658775568008423,
+      "learning_rate": 2.915492957746479e-05,
+      "loss": 1.1626,
+      "step": 70
+    },
+    {
+      "epoch": 0.2664298401420959,
+      "grad_norm": 0.49945878982543945,
+      "learning_rate": 2.9999628430060982e-05,
+      "loss": 1.1247,
+      "step": 75
+    },
+    {
+      "epoch": 0.2841918294849023,
+      "grad_norm": 0.5159095525741577,
+      "learning_rate": 2.9997357791542522e-05,
+      "loss": 1.1237,
+      "step": 80
+    },
+    {
+      "epoch": 0.3019538188277087,
+      "grad_norm": 0.5034985542297363,
+      "learning_rate": 2.9993023254350274e-05,
+      "loss": 1.135,
+      "step": 85
+    },
+    {
+      "epoch": 0.3197158081705151,
+      "grad_norm": 0.5856865644454956,
+      "learning_rate": 2.9986625414992358e-05,
+      "loss": 1.0438,
+      "step": 90
+    },
+    {
+      "epoch": 0.33747779751332146,
+      "grad_norm": 0.5520426630973816,
+      "learning_rate": 2.9978165153923333e-05,
+      "loss": 1.0886,
+      "step": 95
+    },
+    {
+      "epoch": 0.3552397868561279,
+      "grad_norm": 0.49657881259918213,
+      "learning_rate": 2.996764363542307e-05,
+      "loss": 1.0517,
+      "step": 100
+    },
+    {
+      "epoch": 0.37300177619893427,
+      "grad_norm": 0.5683771967887878,
+      "learning_rate": 2.995506230743648e-05,
+      "loss": 1.0431,
+      "step": 105
+    },
+    {
+      "epoch": 0.3907637655417407,
+      "grad_norm": 0.5750526785850525,
+      "learning_rate": 2.9940422901374294e-05,
+      "loss": 1.0203,
+      "step": 110
+    },
+    {
+      "epoch": 0.40852575488454707,
+      "grad_norm": 0.5807237029075623,
+      "learning_rate": 2.992372743187477e-05,
+      "loss": 0.9762,
+      "step": 115
+    },
+    {
+      "epoch": 0.42628774422735344,
+      "grad_norm": 0.5997951030731201,
+      "learning_rate": 2.9904978196526445e-05,
+      "loss": 1.0097,
+      "step": 120
+    },
+    {
+      "epoch": 0.44404973357015987,
+      "grad_norm": 0.5774429440498352,
+      "learning_rate": 2.988417777555195e-05,
+      "loss": 0.9927,
+      "step": 125
+    },
+    {
+      "epoch": 0.46181172291296624,
+      "grad_norm": 0.6737991571426392,
+      "learning_rate": 2.986132903145292e-05,
+      "loss": 0.9825,
+      "step": 130
+    },
+    {
+      "epoch": 0.47957371225577267,
+      "grad_norm": 0.7121424078941345,
+      "learning_rate": 2.9836435108616075e-05,
+      "loss": 0.9792,
+      "step": 135
+    },
+    {
+      "epoch": 0.49733570159857904,
+      "grad_norm": 0.8093393445014954,
+      "learning_rate": 2.9809499432880494e-05,
+      "loss": 0.8921,
+      "step": 140
+    },
+    {
+      "epoch": 0.5150976909413855,
+      "grad_norm": 0.8331164121627808,
+      "learning_rate": 2.9780525711066147e-05,
+      "loss": 0.8977,
+      "step": 145
+    },
+    {
+      "epoch": 0.5328596802841918,
+      "grad_norm": 0.710951566696167,
+      "learning_rate": 2.9749517930463785e-05,
+      "loss": 0.942,
+      "step": 150
+    },
+    {
+      "epoch": 0.5506216696269982,
+      "grad_norm": 0.7496849298477173,
+      "learning_rate": 2.971648035828622e-05,
+      "loss": 0.8894,
+      "step": 155
+    },
+    {
+      "epoch": 0.5683836589698046,
+      "grad_norm": 0.7852967381477356,
+      "learning_rate": 2.9681417541081072e-05,
+      "loss": 0.9124,
+      "step": 160
+    },
+    {
+      "epoch": 0.5861456483126111,
+      "grad_norm": 0.8863868713378906,
+      "learning_rate": 2.9644334304105092e-05,
+      "loss": 0.9476,
+      "step": 165
+    },
+    {
+      "epoch": 0.6039076376554174,
+      "grad_norm": 1.0187339782714844,
+      "learning_rate": 2.960523575066012e-05,
+      "loss": 0.8799,
+      "step": 170
+    },
+    {
+      "epoch": 0.6216696269982238,
+      "grad_norm": 0.941071093082428,
+      "learning_rate": 2.956412726139078e-05,
+      "loss": 0.8753,
+      "step": 175
+    },
+    {
+      "epoch": 0.6394316163410302,
+      "grad_norm": 0.7954825758934021,
+      "learning_rate": 2.9521014493544017e-05,
+      "loss": 0.8491,
+      "step": 180
+    },
+    {
+      "epoch": 0.6571936056838366,
+      "grad_norm": 0.837969183921814,
+      "learning_rate": 2.9475903380190547e-05,
+      "loss": 0.8343,
+      "step": 185
+    },
+    {
+      "epoch": 0.6749555950266429,
+      "grad_norm": 0.971666157245636,
+      "learning_rate": 2.9428800129408374e-05,
+      "loss": 0.8167,
+      "step": 190
+    },
+    {
+      "epoch": 0.6927175843694494,
+      "grad_norm": 0.8235210180282593,
+      "learning_rate": 2.9379711223428446e-05,
+      "loss": 0.8028,
+      "step": 195
+    },
+    {
+      "epoch": 0.7104795737122558,
+      "grad_norm": 0.8552044034004211,
+      "learning_rate": 2.9328643417742592e-05,
+      "loss": 0.8244,
+      "step": 200
+    },
+    {
+      "epoch": 0.7282415630550622,
+      "grad_norm": 0.92862468957901,
+      "learning_rate": 2.9275603740173835e-05,
+      "loss": 0.8089,
+      "step": 205
+    },
+    {
+      "epoch": 0.7460035523978685,
+      "grad_norm": 0.9003951549530029,
+      "learning_rate": 2.9220599489909254e-05,
+      "loss": 0.7295,
+      "step": 210
+    },
+    {
+      "epoch": 0.7637655417406749,
+      "grad_norm": 1.0426762104034424,
+      "learning_rate": 2.916363823649548e-05,
+      "loss": 0.7462,
+      "step": 215
+    },
+    {
+      "epoch": 0.7815275310834814,
+      "grad_norm": 0.8161042332649231,
+      "learning_rate": 2.9104727818797e-05,
+      "loss": 0.7455,
+      "step": 220
+    },
+    {
+      "epoch": 0.7992895204262878,
+      "grad_norm": 0.9472765326499939,
+      "learning_rate": 2.904387634391739e-05,
+      "loss": 0.752,
+      "step": 225
+    },
+    {
+      "epoch": 0.8170515097690941,
+      "grad_norm": 0.9386636018753052,
+      "learning_rate": 2.8981092186083632e-05,
+      "loss": 0.7525,
+      "step": 230
+    },
+    {
+      "epoch": 0.8348134991119005,
+      "grad_norm": 1.1699973344802856,
+      "learning_rate": 2.8916383985493676e-05,
+      "loss": 0.7123,
+      "step": 235
+    },
+    {
+      "epoch": 0.8525754884547069,
+      "grad_norm": 0.8852489590644836,
+      "learning_rate": 2.884976064712741e-05,
+      "loss": 0.698,
+      "step": 240
+    },
+    {
+      "epoch": 0.8703374777975134,
+      "grad_norm": 1.0662732124328613,
+      "learning_rate": 2.878123133952115e-05,
+      "loss": 0.7178,
+      "step": 245
+    },
+    {
+      "epoch": 0.8880994671403197,
+      "grad_norm": 1.2624017000198364,
+      "learning_rate": 2.8710805493505922e-05,
+      "loss": 0.6831,
+      "step": 250
+    },
+    {
+      "epoch": 0.9058614564831261,
+      "grad_norm": 1.024356484413147,
+      "learning_rate": 2.863849280090959e-05,
+      "loss": 0.6938,
+      "step": 255
+    },
+    {
+      "epoch": 0.9236234458259325,
+      "grad_norm": 1.0030966997146606,
+      "learning_rate": 2.856430321322311e-05,
+      "loss": 0.6796,
+      "step": 260
+    },
+    {
+      "epoch": 0.9413854351687388,
+      "grad_norm": 1.0371718406677246,
+      "learning_rate": 2.848824694023102e-05,
+      "loss": 0.6981,
+      "step": 265
+    },
+    {
+      "epoch": 0.9591474245115453,
+      "grad_norm": 0.970425546169281,
+      "learning_rate": 2.8410334448606377e-05,
+      "loss": 0.6895,
+      "step": 270
+    },
+    {
+      "epoch": 0.9769094138543517,
+      "grad_norm": 0.9296903014183044,
+      "learning_rate": 2.8330576460470407e-05,
+      "loss": 0.6514,
+      "step": 275
+    },
+    {
+      "epoch": 0.9946714031971581,
+      "grad_norm": 0.9770685434341431,
+      "learning_rate": 2.824898395191691e-05,
+      "loss": 0.6336,
+      "step": 280
+    },
+    {
+      "epoch": 1.0106571936056838,
+      "grad_norm": 1.1167672872543335,
+      "learning_rate": 2.816556815150179e-05,
+      "loss": 0.5954,
+      "step": 285
+    },
+    {
+      "epoch": 1.0284191829484903,
+      "grad_norm": 1.0073131322860718,
+      "learning_rate": 2.808034053869776e-05,
+      "loss": 0.5953,
+      "step": 290
+    },
+    {
+      "epoch": 1.0461811722912966,
+      "grad_norm": 1.1674365997314453,
+      "learning_rate": 2.7993312842314636e-05,
+      "loss": 0.5708,
+      "step": 295
+    },
+    {
+      "epoch": 1.063943161634103,
+      "grad_norm": 1.1259663105010986,
+      "learning_rate": 2.7904497038885184e-05,
+      "loss": 0.5943,
+      "step": 300
+    },
+    {
+      "epoch": 1.0817051509769093,
+      "grad_norm": 1.0642991065979004,
+      "learning_rate": 2.7813905351016984e-05,
+      "loss": 0.5721,
+      "step": 305
+    },
+    {
+      "epoch": 1.0994671403197158,
+      "grad_norm": 1.2667790651321411,
+      "learning_rate": 2.7721550245710362e-05,
+      "loss": 0.577,
+      "step": 310
+    },
+    {
+      "epoch": 1.1172291296625223,
+      "grad_norm": 1.0688438415527344,
+      "learning_rate": 2.7627444432642708e-05,
+      "loss": 0.5548,
+      "step": 315
+    },
+    {
+      "epoch": 1.1349911190053286,
+      "grad_norm": 1.0694912672042847,
+      "learning_rate": 2.7531600862419422e-05,
+      "loss": 0.5499,
+      "step": 320
+    },
+    {
+      "epoch": 1.152753108348135,
+      "grad_norm": 1.2388120889663696,
+      "learning_rate": 2.743403272479168e-05,
+      "loss": 0.5154,
+      "step": 325
+    },
+    {
+      "epoch": 1.1705150976909413,
+      "grad_norm": 1.31024169921875,
+      "learning_rate": 2.7334753446841277e-05,
+      "loss": 0.5227,
+      "step": 330
+    },
+    {
+      "epoch": 1.1882770870337478,
+      "grad_norm": 1.1155751943588257,
+      "learning_rate": 2.7233776691132848e-05,
+      "loss": 0.5062,
+      "step": 335
+    },
+    {
+      "epoch": 1.206039076376554,
+      "grad_norm": 1.3883517980575562,
+      "learning_rate": 2.7131116353833633e-05,
+      "loss": 0.5329,
+      "step": 340
+    },
+    {
+      "epoch": 1.2238010657193605,
+      "grad_norm": 1.2181286811828613,
+      "learning_rate": 2.702678656280116e-05,
+      "loss": 0.5268,
+      "step": 345
+    },
+    {
+      "epoch": 1.241563055062167,
+      "grad_norm": 1.1056498289108276,
+      "learning_rate": 2.6920801675638965e-05,
+      "loss": 0.5769,
+      "step": 350
+    },
+    {
+      "epoch": 1.2593250444049733,
+      "grad_norm": 1.0863943099975586,
+      "learning_rate": 2.681317627772076e-05,
+      "loss": 0.5136,
+      "step": 355
+    },
+    {
+      "epoch": 1.2770870337477798,
+      "grad_norm": 1.2052078247070312,
+      "learning_rate": 2.670392518018322e-05,
+      "loss": 0.5167,
+      "step": 360
+    },
+    {
+      "epoch": 1.294849023090586,
+      "grad_norm": 1.2922052145004272,
+      "learning_rate": 2.6593063417887713e-05,
+      "loss": 0.5198,
+      "step": 365
+    },
+    {
+      "epoch": 1.3126110124333925,
+      "grad_norm": 1.155900239944458,
+      "learning_rate": 2.6480606247351247e-05,
+      "loss": 0.5878,
+      "step": 370
+    },
+    {
+      "epoch": 1.330373001776199,
+      "grad_norm": 1.140750765800476,
+      "learning_rate": 2.6366569144646903e-05,
+      "loss": 0.5175,
+      "step": 375
+    },
+    {
+      "epoch": 1.3481349911190053,
+      "grad_norm": 1.2477790117263794,
+      "learning_rate": 2.625096780327404e-05,
+      "loss": 0.496,
+      "step": 380
+    },
+    {
+      "epoch": 1.3658969804618117,
+      "grad_norm": 1.0702998638153076,
+      "learning_rate": 2.6133818131998622e-05,
+      "loss": 0.4407,
+      "step": 385
+    },
+    {
+      "epoch": 1.383658969804618,
+      "grad_norm": 1.0303690433502197,
+      "learning_rate": 2.601513625266388e-05,
+      "loss": 0.484,
+      "step": 390
+    },
+    {
+      "epoch": 1.4014209591474245,
+      "grad_norm": 1.126442313194275,
+      "learning_rate": 2.5894938497971668e-05,
+      "loss": 0.4645,
+      "step": 395
+    },
+    {
+      "epoch": 1.419182948490231,
+      "grad_norm": 1.166357398033142,
+      "learning_rate": 2.5773241409234788e-05,
+      "loss": 0.4727,
+      "step": 400
+    },
+    {
+      "epoch": 1.4369449378330372,
+      "grad_norm": 1.2368241548538208,
+      "learning_rate": 2.5650061734100633e-05,
+      "loss": 0.4924,
+      "step": 405
+    },
+    {
+      "epoch": 1.4547069271758437,
+      "grad_norm": 1.234560489654541,
+      "learning_rate": 2.55254164242464e-05,
+      "loss": 0.4303,
+      "step": 410
+    },
+    {
+      "epoch": 1.47246891651865,
+      "grad_norm": 1.1312541961669922,
+      "learning_rate": 2.5399322633046254e-05,
+      "loss": 0.4484,
+      "step": 415
+    },
+    {
+      "epoch": 1.4902309058614565,
+      "grad_norm": 1.03053617477417,
+      "learning_rate": 2.5271797713210726e-05,
+      "loss": 0.4219,
+      "step": 420
+    },
+    {
+      "epoch": 1.507992895204263,
+      "grad_norm": 1.2007708549499512,
+      "learning_rate": 2.5142859214398666e-05,
+      "loss": 0.4475,
+      "step": 425
+    },
+    {
+      "epoch": 1.5257548845470694,
+      "grad_norm": 1.071448802947998,
+      "learning_rate": 2.5012524880802108e-05,
+      "loss": 0.4191,
+      "step": 430
+    },
+    {
+      "epoch": 1.5435168738898757,
+      "grad_norm": 1.225663185119629,
+      "learning_rate": 2.4880812648704367e-05,
+      "loss": 0.441,
+      "step": 435
+    },
+    {
+      "epoch": 1.561278863232682,
+      "grad_norm": 1.0680769681930542,
+      "learning_rate": 2.4747740644011682e-05,
+      "loss": 0.4541,
+      "step": 440
+    },
+    {
+      "epoch": 1.5790408525754884,
+      "grad_norm": 1.0855811834335327,
+      "learning_rate": 2.4613327179758773e-05,
+      "loss": 0.4428,
+      "step": 445
+    },
+    {
+      "epoch": 1.596802841918295,
+      "grad_norm": 1.197534441947937,
+      "learning_rate": 2.4477590753588656e-05,
+      "loss": 0.4217,
+      "step": 450
+    },
+    {
+      "epoch": 1.6145648312611014,
+      "grad_norm": 1.1745901107788086,
+      "learning_rate": 2.434055004520704e-05,
+      "loss": 0.4236,
+      "step": 455
+    },
+    {
+      "epoch": 1.6323268206039077,
+      "grad_norm": 1.2637349367141724,
+      "learning_rate": 2.420222391381168e-05,
+      "loss": 0.4819,
+      "step": 460
+    },
+    {
+      "epoch": 1.650088809946714,
+      "grad_norm": 1.2748724222183228,
+      "learning_rate": 2.4062631395496993e-05,
+      "loss": 0.4384,
+      "step": 465
+    },
+    {
+      "epoch": 1.6678507992895204,
+      "grad_norm": 1.0431526899337769,
+      "learning_rate": 2.392179170063441e-05,
+      "loss": 0.3863,
+      "step": 470
+    },
+    {
+      "epoch": 1.6856127886323269,
+      "grad_norm": 1.1915509700775146,
+      "learning_rate": 2.3779724211228632e-05,
+      "loss": 0.4026,
+      "step": 475
+    },
+    {
+      "epoch": 1.7033747779751334,
+      "grad_norm": 1.1567018032073975,
+      "learning_rate": 2.3636448478250378e-05,
+      "loss": 0.3837,
+      "step": 480
+    },
+    {
+      "epoch": 1.7211367673179396,
+      "grad_norm": 1.1712560653686523,
+      "learning_rate": 2.3491984218945814e-05,
+      "loss": 0.4064,
+      "step": 485
+    },
+    {
+      "epoch": 1.738898756660746,
+      "grad_norm": 1.1683224439620972,
+      "learning_rate": 2.3346351314123093e-05,
+      "loss": 0.4456,
+      "step": 490
+    },
+    {
+      "epoch": 1.7566607460035524,
+      "grad_norm": 1.2679800987243652,
+      "learning_rate": 2.3199569805416436e-05,
+      "loss": 0.4393,
+      "step": 495
+    },
+    {
+      "epoch": 1.7744227353463589,
+      "grad_norm": 1.3210681676864624,
+      "learning_rate": 2.305165989252804e-05,
+      "loss": 0.3742,
+      "step": 500
+    },
+    {
+      "epoch": 1.7921847246891653,
+      "grad_norm": 1.0676344633102417,
+      "learning_rate": 2.2902641930448253e-05,
+      "loss": 0.3477,
+      "step": 505
+    },
+    {
+      "epoch": 1.8099467140319716,
+      "grad_norm": 1.1912447214126587,
+      "learning_rate": 2.2752536426654363e-05,
+      "loss": 0.3905,
+      "step": 510
+    },
+    {
+      "epoch": 1.8277087033747779,
+      "grad_norm": 1.1013745069503784,
+      "learning_rate": 2.260136403828842e-05,
+      "loss": 0.3606,
+      "step": 515
+    },
+    {
+      "epoch": 1.8454706927175843,
+      "grad_norm": 1.0575093030929565,
+      "learning_rate": 2.2449145569314437e-05,
+      "loss": 0.3727,
+      "step": 520
+    },
+    {
+      "epoch": 1.8632326820603908,
+      "grad_norm": 1.0612674951553345,
+      "learning_rate": 2.229590196765541e-05,
+      "loss": 0.3739,
+      "step": 525
+    },
+    {
+      "epoch": 1.8809946714031973,
+      "grad_norm": 1.2063333988189697,
+      "learning_rate": 2.2141654322310534e-05,
+      "loss": 0.3485,
+      "step": 530
+    },
+    {
+      "epoch": 1.8987566607460036,
+      "grad_norm": 1.1345680952072144,
+      "learning_rate": 2.1986423860452957e-05,
+      "loss": 0.3975,
+      "step": 535
+    },
+    {
+      "epoch": 1.9165186500888098,
+      "grad_norm": 1.2932173013687134,
+      "learning_rate": 2.183023194450857e-05,
+      "loss": 0.3558,
+      "step": 540
+    },
+    {
+      "epoch": 1.9342806394316163,
+      "grad_norm": 1.1968685388565063,
+      "learning_rate": 2.167310006921615e-05,
+      "loss": 0.3387,
+      "step": 545
+    },
+    {
+      "epoch": 1.9520426287744228,
+      "grad_norm": 1.1290210485458374,
+      "learning_rate": 2.151504985866933e-05,
+      "loss": 0.3449,
+      "step": 550
+    },
+    {
+      "epoch": 1.9698046181172293,
+      "grad_norm": 1.1544890403747559,
+      "learning_rate": 2.1356103063340716e-05,
+      "loss": 0.3698,
+      "step": 555
+    },
+    {
+      "epoch": 1.9875666074600356,
+      "grad_norm": 1.199317216873169,
+      "learning_rate": 2.1196281557088674e-05,
+      "loss": 0.3192,
+      "step": 560
+    },
+    {
+      "epoch": 2.003552397868561,
+      "grad_norm": 1.2987078428268433,
+      "learning_rate": 2.1035607334147083e-05,
+      "loss": 0.3253,
+      "step": 565
+    },
+    {
+      "epoch": 2.0213143872113677,
+      "grad_norm": 1.2794301509857178,
+      "learning_rate": 2.087410250609854e-05,
+      "loss": 0.2469,
+      "step": 570
+    },
+    {
+      "epoch": 2.039076376554174,
+      "grad_norm": 1.2110157012939453,
+      "learning_rate": 2.0711789298831435e-05,
+      "loss": 0.2784,
+      "step": 575
+    },
+    {
+      "epoch": 2.0568383658969807,
+      "grad_norm": 1.1234550476074219,
+      "learning_rate": 2.0548690049481274e-05,
+      "loss": 0.2782,
+      "step": 580
+    },
+    {
+      "epoch": 2.0746003552397867,
+      "grad_norm": 1.2034980058670044,
+      "learning_rate": 2.0384827203356702e-05,
+      "loss": 0.2768,
+      "step": 585
+    },
+    {
+      "epoch": 2.092362344582593,
+      "grad_norm": 1.1906369924545288,
+      "learning_rate": 2.0220223310850625e-05,
+      "loss": 0.2756,
+      "step": 590
+    },
+    {
+      "epoch": 2.1101243339253997,
+      "grad_norm": 1.1176049709320068,
+      "learning_rate": 2.0054901024336896e-05,
+      "loss": 0.262,
+      "step": 595
+    },
+    {
+      "epoch": 2.127886323268206,
+      "grad_norm": 1.0588326454162598,
+      "learning_rate": 1.988888309505293e-05,
+      "loss": 0.2821,
+      "step": 600
+    },
+    {
+      "epoch": 2.1456483126110126,
+      "grad_norm": 1.1338000297546387,
+      "learning_rate": 1.9722192369968776e-05,
+      "loss": 0.2613,
+      "step": 605
+    },
+    {
+      "epoch": 2.1634103019538187,
+      "grad_norm": 1.1807286739349365,
+      "learning_rate": 1.9554851788642928e-05,
+      "loss": 0.2339,
+      "step": 610
+    },
+    {
+      "epoch": 2.181172291296625,
+      "grad_norm": 1.0721327066421509,
+      "learning_rate": 1.9386884380065474e-05,
+      "loss": 0.2644,
+      "step": 615
+    },
+    {
+      "epoch": 2.1989342806394316,
+      "grad_norm": 1.0247159004211426,
+      "learning_rate": 1.9218313259488883e-05,
+      "loss": 0.2627,
+      "step": 620
+    },
+    {
+      "epoch": 2.216696269982238,
+      "grad_norm": 1.3054232597351074,
+      "learning_rate": 1.9049161625246962e-05,
+      "loss": 0.2381,
+      "step": 625
+    },
+    {
+      "epoch": 2.2344582593250446,
+      "grad_norm": 1.1615649461746216,
+      "learning_rate": 1.8879452755562344e-05,
+      "loss": 0.249,
+      "step": 630
+    },
+    {
+      "epoch": 2.2522202486678506,
+      "grad_norm": 1.5350950956344604,
+      "learning_rate": 1.8709210005343e-05,
+      "loss": 0.242,
+      "step": 635
+    },
+    {
+      "epoch": 2.269982238010657,
+      "grad_norm": 1.1368858814239502,
+      "learning_rate": 1.85384568029682e-05,
+      "loss": 0.236,
+      "step": 640
+    },
+    {
+      "epoch": 2.2877442273534636,
+      "grad_norm": 1.163325309753418,
+      "learning_rate": 1.836721664706437e-05,
+      "loss": 0.2277,
+      "step": 645
+    },
+    {
+      "epoch": 2.30550621669627,
+      "grad_norm": 1.2359697818756104,
+      "learning_rate": 1.819551310327122e-05,
+      "loss": 0.2626,
+      "step": 650
+    },
+    {
+      "epoch": 2.323268206039076,
+      "grad_norm": 1.1182560920715332,
+      "learning_rate": 1.8023369800998776e-05,
+      "loss": 0.2628,
+      "step": 655
+    },
+    {
+      "epoch": 2.3410301953818826,
+      "grad_norm": 1.1834044456481934,
+      "learning_rate": 1.78508104301755e-05,
+      "loss": 0.2552,
+      "step": 660
+    },
+    {
+      "epoch": 2.358792184724689,
+      "grad_norm": 1.1643848419189453,
+      "learning_rate": 1.767785873798817e-05,
+      "loss": 0.2402,
+      "step": 665
+    },
+    {
+      "epoch": 2.3765541740674956,
+      "grad_norm": 1.1760133504867554,
+      "learning_rate": 1.7504538525613833e-05,
+      "loss": 0.236,
+      "step": 670
+    },
+    {
+      "epoch": 2.394316163410302,
+      "grad_norm": 1.2071462869644165,
+      "learning_rate": 1.7330873644944367e-05,
+      "loss": 0.2473,
+      "step": 675
+    },
+    {
+      "epoch": 2.412078152753108,
+      "grad_norm": 1.2645944356918335,
+      "learning_rate": 1.715688799530403e-05,
+      "loss": 0.2245,
+      "step": 680
+    },
+    {
+      "epoch": 2.4298401420959146,
+      "grad_norm": 1.1790143251419067,
+      "learning_rate": 1.6982605520160503e-05,
+      "loss": 0.2484,
+      "step": 685
+    },
+    {
+      "epoch": 2.447602131438721,
+      "grad_norm": 1.1275826692581177,
+      "learning_rate": 1.6808050203829845e-05,
+      "loss": 0.2169,
+      "step": 690
+    },
+    {
+      "epoch": 2.4653641207815276,
+      "grad_norm": 1.1996047496795654,
+      "learning_rate": 1.6633246068175833e-05,
+      "loss": 0.2278,
+      "step": 695
+    },
+    {
+      "epoch": 2.483126110124334,
+      "grad_norm": 1.18192458152771,
+      "learning_rate": 1.6458217169304126e-05,
+      "loss": 0.2351,
+      "step": 700
+    },
+    {
+      "epoch": 2.50088809946714,
+      "grad_norm": 1.0418000221252441,
+      "learning_rate": 1.628298759425175e-05,
+      "loss": 0.2297,
+      "step": 705
+    },
+    {
+      "epoch": 2.5186500888099466,
+      "grad_norm": 1.2126238346099854,
+      "learning_rate": 1.6107581457672267e-05,
+      "loss": 0.2396,
+      "step": 710
+    },
+    {
+      "epoch": 2.536412078152753,
+      "grad_norm": 1.1965978145599365,
+      "learning_rate": 1.5932022898517195e-05,
+      "loss": 0.2124,
+      "step": 715
+    },
+    {
+      "epoch": 2.5541740674955595,
+      "grad_norm": 1.1391175985336304,
+      "learning_rate": 1.5756336076714065e-05,
+      "loss": 0.2021,
+      "step": 720
+    },
+    {
+      "epoch": 2.571936056838366,
+      "grad_norm": 1.0899744033813477,
+      "learning_rate": 1.5580545169841592e-05,
+      "loss": 0.2178,
+      "step": 725
+    },
+    {
+      "epoch": 2.589698046181172,
+      "grad_norm": 1.2015151977539062,
+      "learning_rate": 1.5404674369802425e-05,
+      "loss": 0.2487,
+      "step": 730
+    },
+    {
+      "epoch": 2.6074600355239785,
+      "grad_norm": 1.1883600950241089,
+      "learning_rate": 1.522874787949389e-05,
+      "loss": 0.1964,
+      "step": 735
+    },
+    {
+      "epoch": 2.625222024866785,
+      "grad_norm": 1.173923373222351,
+      "learning_rate": 1.5052789909477286e-05,
+      "loss": 0.2199,
+      "step": 740
+    },
+    {
+      "epoch": 2.6429840142095915,
+      "grad_norm": 1.0762274265289307,
+      "learning_rate": 1.4876824674646062e-05,
+      "loss": 0.2259,
+      "step": 745
+    },
+    {
+      "epoch": 2.660746003552398,
+      "grad_norm": 1.0043290853500366,
+      "learning_rate": 1.4700876390893426e-05,
+      "loss": 0.2139,
+      "step": 750
+    },
+    {
+      "epoch": 2.678507992895204,
+      "grad_norm": 1.175076961517334,
+      "learning_rate": 1.4524969271779844e-05,
+      "loss": 0.1997,
+      "step": 755
+    },
+    {
+      "epoch": 2.6962699822380105,
+      "grad_norm": 1.1845417022705078,
+      "learning_rate": 1.4349127525200783e-05,
+      "loss": 0.226,
+      "step": 760
+    },
+    {
+      "epoch": 2.714031971580817,
+      "grad_norm": 1.1732301712036133,
+      "learning_rate": 1.4173375350055319e-05,
+      "loss": 0.2154,
+      "step": 765
+    },
+    {
+      "epoch": 2.7317939609236235,
+      "grad_norm": 1.0934293270111084,
+      "learning_rate": 1.3997736932915922e-05,
+      "loss": 0.1878,
+      "step": 770
+    },
+    {
+      "epoch": 2.74955595026643,
+      "grad_norm": 1.2918049097061157,
+      "learning_rate": 1.3822236444699991e-05,
+      "loss": 0.2082,
+      "step": 775
+    },
+    {
+      "epoch": 2.767317939609236,
+      "grad_norm": 1.0803861618041992,
+      "learning_rate": 1.3646898037343475e-05,
+      "loss": 0.199,
+      "step": 780
+    },
+    {
+      "epoch": 2.7850799289520425,
+      "grad_norm": 1.2331840991973877,
+      "learning_rate": 1.347174584047717e-05,
+      "loss": 0.1915,
+      "step": 785
+    },
+    {
+      "epoch": 2.802841918294849,
+      "grad_norm": 1.2146940231323242,
+      "learning_rate": 1.3296803958106046e-05,
+      "loss": 0.1903,
+      "step": 790
+    },
+    {
+      "epoch": 2.8206039076376554,
+      "grad_norm": 1.3104547262191772,
+      "learning_rate": 1.3122096465292132e-05,
+      "loss": 0.1997,
+      "step": 795
+    },
+    {
+      "epoch": 2.838365896980462,
+      "grad_norm": 1.0410480499267578,
+      "learning_rate": 1.294764740484135e-05,
+      "loss": 0.1848,
+      "step": 800
+    },
+    {
+      "epoch": 2.856127886323268,
+      "grad_norm": 1.2668956518173218,
+      "learning_rate": 1.2773480783994805e-05,
+      "loss": 0.1937,
+      "step": 805
+    },
+    {
+      "epoch": 2.8738898756660745,
+      "grad_norm": 1.1789441108703613,
+      "learning_rate": 1.259962057112498e-05,
+      "loss": 0.2104,
+      "step": 810
+    },
+    {
+      "epoch": 2.891651865008881,
+      "grad_norm": 1.1638792753219604,
+      "learning_rate": 1.2426090692437266e-05,
+      "loss": 0.1726,
+      "step": 815
+    },
+    {
+      "epoch": 2.9094138543516874,
+      "grad_norm": 1.0087004899978638,
+      "learning_rate": 1.225291502867728e-05,
+      "loss": 0.187,
+      "step": 820
+    },
+    {
+      "epoch": 2.927175843694494,
+      "grad_norm": 1.1213196516036987,
+      "learning_rate": 1.208011741184449e-05,
+      "loss": 0.1714,
+      "step": 825
+    },
+    {
+      "epoch": 2.9449378330373,
+      "grad_norm": 1.1991121768951416,
+      "learning_rate": 1.1907721621912476e-05,
+      "loss": 0.185,
+      "step": 830
+    },
+    {
+      "epoch": 2.9626998223801064,
+      "grad_norm": 1.0833150148391724,
+      "learning_rate": 1.1735751383556432e-05,
+      "loss": 0.1779,
+      "step": 835
+    },
+    {
+      "epoch": 2.980461811722913,
+      "grad_norm": 1.2620830535888672,
+      "learning_rate": 1.1564230362888195e-05,
+      "loss": 0.205,
+      "step": 840
+    },
+    {
+      "epoch": 2.9982238010657194,
+      "grad_norm": 1.1784212589263916,
+      "learning_rate": 1.1393182164199404e-05,
+      "loss": 0.2045,
+      "step": 845
+    },
+    {
+      "epoch": 3.0142095914742453,
+      "grad_norm": 1.0905832052230835,
+      "learning_rate": 1.1222630326713122e-05,
+      "loss": 0.1293,
+      "step": 850
+    },
+    {
+      "epoch": 3.0319715808170513,
+      "grad_norm": 1.245754599571228,
+      "learning_rate": 1.1052598321344435e-05,
+      "loss": 0.1574,
+      "step": 855
+    },
+    {
+      "epoch": 3.049733570159858,
+      "grad_norm": 0.9812668561935425,
+      "learning_rate": 1.0883109547470435e-05,
+      "loss": 0.1259,
+      "step": 860
+    },
+    {
+      "epoch": 3.0674955595026643,
+      "grad_norm": 1.1846905946731567,
+      "learning_rate": 1.0714187329710066e-05,
+      "loss": 0.1286,
+      "step": 865
+    },
+    {
+      "epoch": 3.0852575488454708,
+      "grad_norm": 1.0834468603134155,
+      "learning_rate": 1.0545854914714233e-05,
+      "loss": 0.1506,
+      "step": 870
+    },
+    {
+      "epoch": 3.1030195381882772,
+      "grad_norm": 1.082939863204956,
+      "learning_rate": 1.0378135467966691e-05,
+      "loss": 0.1557,
+      "step": 875
+    },
+    {
+      "epoch": 3.1207815275310833,
+      "grad_norm": 1.0465799570083618,
+      "learning_rate": 1.021105207059602e-05,
+      "loss": 0.1127,
+      "step": 880
+    },
+    {
+      "epoch": 3.1385435168738898,
+      "grad_norm": 1.0733388662338257,
+      "learning_rate": 1.0044627716199297e-05,
+      "loss": 0.1394,
+      "step": 885
+    },
+    {
+      "epoch": 3.1563055062166963,
+      "grad_norm": 1.2737563848495483,
+      "learning_rate": 9.878885307677756e-06,
+      "loss": 0.1355,
+      "step": 890
+    },
+    {
+      "epoch": 3.1740674955595027,
+      "grad_norm": 1.2421573400497437,
+      "learning_rate": 9.713847654084966e-06,
+      "loss": 0.1396,
+      "step": 895
+    },
+    {
+      "epoch": 3.191829484902309,
+      "grad_norm": 1.1305955648422241,
+      "learning_rate": 9.549537467487884e-06,
+      "loss": 0.1505,
+      "step": 900
+    },
+    {
+      "epoch": 3.2095914742451153,
+      "grad_norm": 0.9503009915351868,
+      "learning_rate": 9.385977359841287e-06,
+      "loss": 0.1492,
+      "step": 905
+    },
+    {
+      "epoch": 3.2273534635879217,
+      "grad_norm": 1.04369056224823,
+      "learning_rate": 9.22318983987597e-06,
+      "loss": 0.1422,
+      "step": 910
+    },
+    {
+      "epoch": 3.2451154529307282,
+      "grad_norm": 1.144344449043274,
+      "learning_rate": 9.061197310001176e-06,
+      "loss": 0.1164,
+      "step": 915
+    },
+    {
+      "epoch": 3.2628774422735347,
+      "grad_norm": 1.3119343519210815,
+      "learning_rate": 8.900022063221582e-06,
+      "loss": 0.1344,
+      "step": 920
+    },
+    {
+      "epoch": 3.280639431616341,
+      "grad_norm": 1.1138067245483398,
+      "learning_rate": 8.739686280069435e-06,
+      "loss": 0.1319,
+      "step": 925
+    },
+    {
+      "epoch": 3.2984014209591472,
+      "grad_norm": 1.103514313697815,
+      "learning_rate": 8.580212025552108e-06,
+      "loss": 0.1404,
+      "step": 930
+    },
+    {
+      "epoch": 3.3161634103019537,
+      "grad_norm": 1.3191359043121338,
+      "learning_rate": 8.421621246115565e-06,
+      "loss": 0.1382,
+      "step": 935
+    },
+    {
+      "epoch": 3.33392539964476,
+      "grad_norm": 0.9715001583099365,
+      "learning_rate": 8.263935766624145e-06,
+      "loss": 0.1267,
+      "step": 940
+    },
+    {
+      "epoch": 3.3516873889875667,
+      "grad_norm": 1.1174882650375366,
+      "learning_rate": 8.107177287357104e-06,
+      "loss": 0.1309,
+      "step": 945
+    },
+    {
+      "epoch": 3.369449378330373,
+      "grad_norm": 1.1527106761932373,
+      "learning_rate": 7.951367381022243e-06,
+      "loss": 0.1152,
+      "step": 950
+    },
+    {
+      "epoch": 3.387211367673179,
+      "grad_norm": 1.0667994022369385,
+      "learning_rate": 7.79652748978714e-06,
+      "loss": 0.1183,
+      "step": 955
+    },
+    {
+      "epoch": 3.4049733570159857,
+      "grad_norm": 0.9557608366012573,
+      "learning_rate": 7.642678922328335e-06,
+      "loss": 0.1242,
+      "step": 960
+    },
+    {
+      "epoch": 3.422735346358792,
+      "grad_norm": 0.9632545709609985,
+      "learning_rate": 7.489842850898886e-06,
+      "loss": 0.1224,
+      "step": 965
+    },
+    {
+      "epoch": 3.4404973357015987,
+      "grad_norm": 0.9741002321243286,
+      "learning_rate": 7.338040308414697e-06,
+      "loss": 0.1135,
+      "step": 970
+    },
+    {
+      "epoch": 3.458259325044405,
+      "grad_norm": 0.9288985729217529,
+      "learning_rate": 7.187292185560008e-06,
+      "loss": 0.1168,
+      "step": 975
+    },
+    {
+      "epoch": 3.476021314387211,
+      "grad_norm": 1.246943473815918,
+      "learning_rate": 7.037619227912484e-06,
+      "loss": 0.1204,
+      "step": 980
+    },
+    {
+      "epoch": 3.4937833037300177,
+      "grad_norm": 0.9657527804374695,
+      "learning_rate": 6.889042033088256e-06,
+      "loss": 0.1146,
+      "step": 985
+    },
+    {
+      "epoch": 3.511545293072824,
+      "grad_norm": 1.037055492401123,
+      "learning_rate": 6.741581047907329e-06,
+      "loss": 0.1127,
+      "step": 990
+    },
+    {
+      "epoch": 3.5293072824156306,
+      "grad_norm": 0.9712250232696533,
+      "learning_rate": 6.5952565655797535e-06,
+      "loss": 0.1154,
+      "step": 995
+    },
+    {
+      "epoch": 3.5470692717584367,
+      "grad_norm": 1.075392723083496,
+      "learning_rate": 6.450088722912902e-06,
+      "loss": 0.1187,
+      "step": 1000
+    },
+    {
+      "epoch": 3.5648312611012436,
+      "grad_norm": 0.8246014714241028,
+      "learning_rate": 6.3060974975403115e-06,
+      "loss": 0.1035,
+      "step": 1005
+    },
+    {
+      "epoch": 3.5825932504440496,
+      "grad_norm": 0.9024150371551514,
+      "learning_rate": 6.1633027051724024e-06,
+      "loss": 0.1285,
+      "step": 1010
+    },
+    {
+      "epoch": 3.600355239786856,
+      "grad_norm": 0.8948878049850464,
+      "learning_rate": 6.0217239968694875e-06,
+      "loss": 0.1219,
+      "step": 1015
+    },
+    {
+      "epoch": 3.6181172291296626,
+      "grad_norm": 1.0379226207733154,
+      "learning_rate": 5.881380856337466e-06,
+      "loss": 0.1186,
+      "step": 1020
+    },
+    {
+      "epoch": 3.6358792184724686,
+      "grad_norm": 0.9066540598869324,
+      "learning_rate": 5.742292597246491e-06,
+      "loss": 0.1026,
+      "step": 1025
+    },
+    {
+      "epoch": 3.6536412078152756,
+      "grad_norm": 1.2567459344863892,
+      "learning_rate": 5.604478360573089e-06,
+      "loss": 0.1164,
+      "step": 1030
+    },
+    {
+      "epoch": 3.6714031971580816,
+      "grad_norm": 0.9321945905685425,
+      "learning_rate": 5.467957111966023e-06,
+      "loss": 0.1016,
+      "step": 1035
+    },
+    {
+      "epoch": 3.689165186500888,
+      "grad_norm": 0.9881175756454468,
+      "learning_rate": 5.3327476391363015e-06,
+      "loss": 0.1046,
+      "step": 1040
+    },
+    {
+      "epoch": 3.7069271758436946,
+      "grad_norm": 1.0275404453277588,
+      "learning_rate": 5.198868549271635e-06,
+      "loss": 0.1162,
+      "step": 1045
+    },
+    {
+      "epoch": 3.7246891651865006,
+      "grad_norm": 1.2424781322479248,
+      "learning_rate": 5.06633826647578e-06,
+      "loss": 0.1221,
+      "step": 1050
+    },
+    {
+      "epoch": 3.7424511545293075,
+      "grad_norm": 1.0425161123275757,
+      "learning_rate": 4.935175029233048e-06,
+      "loss": 0.1044,
+      "step": 1055
+    },
+    {
+      "epoch": 3.7602131438721136,
+      "grad_norm": 1.0624982118606567,
+      "learning_rate": 4.805396887898397e-06,
+      "loss": 0.1207,
+      "step": 1060
+    },
+    {
+      "epoch": 3.77797513321492,
+      "grad_norm": 0.8982229232788086,
+      "learning_rate": 4.677021702213353e-06,
+      "loss": 0.103,
+      "step": 1065
+    },
+    {
+      "epoch": 3.7957371225577266,
+      "grad_norm": 0.9792070984840393,
+      "learning_rate": 4.550067138848213e-06,
+      "loss": 0.1056,
+      "step": 1070
+    },
+    {
+      "epoch": 3.8134991119005326,
+      "grad_norm": 1.0115687847137451,
+      "learning_rate": 4.424550668970803e-06,
+      "loss": 0.098,
+      "step": 1075
+    },
+    {
+      "epoch": 3.8312611012433395,
+      "grad_norm": 0.7825543880462646,
+      "learning_rate": 4.30048956584215e-06,
+      "loss": 0.0953,
+      "step": 1080
+    },
+    {
+      "epoch": 3.8490230905861456,
+      "grad_norm": 0.8458123803138733,
+      "learning_rate": 4.1779009024393555e-06,
+      "loss": 0.0967,
+      "step": 1085
+    },
+    {
+      "epoch": 3.866785079928952,
+      "grad_norm": 0.8293980360031128,
+      "learning_rate": 4.056801549106072e-06,
+      "loss": 0.1042,
+      "step": 1090
+    },
+    {
+      "epoch": 3.8845470692717585,
+      "grad_norm": 0.8098350167274475,
+      "learning_rate": 3.937208171230838e-06,
+      "loss": 0.0948,
+      "step": 1095
+    },
+    {
+      "epoch": 3.9023090586145646,
+      "grad_norm": 0.9925907850265503,
+      "learning_rate": 3.819137226953651e-06,
+      "loss": 0.0996,
+      "step": 1100
+    },
+    {
+      "epoch": 3.9200710479573715,
+      "grad_norm": 0.8489020466804504,
+      "learning_rate": 3.702604964901011e-06,
+      "loss": 0.1068,
+      "step": 1105
+    },
+    {
+      "epoch": 3.9378330373001775,
+      "grad_norm": 0.8919208645820618,
+      "learning_rate": 3.5876274219498462e-06,
+      "loss": 0.1048,
+      "step": 1110
+    },
+    {
+      "epoch": 3.955595026642984,
+      "grad_norm": 0.9406099915504456,
+      "learning_rate": 3.4742204210205397e-06,
+      "loss": 0.0923,
+      "step": 1115
+    },
+    {
+      "epoch": 3.9733570159857905,
+      "grad_norm": 0.8420392870903015,
+      "learning_rate": 3.362399568899447e-06,
+      "loss": 0.103,
+      "step": 1120
+    },
+    {
+      "epoch": 3.9911190053285965,
+      "grad_norm": 0.8659491539001465,
+      "learning_rate": 3.2521802540910998e-06,
+      "loss": 0.1165,
+      "step": 1125
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1410,
+  "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": 1.6092741551511306e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

25_128_e5_3e-5/checkpoint-1128/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96b7946d1d78b1719e782e6e2e941d5b802d64ac2133b6cbedc8dc90ff75572e
+size 7736

25_128_e5_3e-5/checkpoint-1128/zero_to_fp32.py

@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)

25_128_e5_3e-5/checkpoint-1410/README.md

@@ -0,0 +1,202 @@
+---
+base_model: ibm-granite/granite-3.3-8b-base
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.15.2

25_128_e5_3e-5/checkpoint-1410/adapter_config.json

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

25_128_e5_3e-5/checkpoint-1410/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:904b3b97dfeb2191597688f9119b9be2643f005900c6ff78cacc47a33fdc15f9
+size 791751704

25_128_e5_3e-5/checkpoint-1410/latest

@@ -0,0 +1 @@
+global_step1410

25_128_e5_3e-5/checkpoint-1410/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19983129e9859ae200e12e3efa2afd77091cb308e5c0d72b67e9be2dc3abd5e8
+size 15920

25_128_e5_3e-5/checkpoint-1410/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f725c3c1053d88d27275103a3e8828aae59dce839c9e6ac1b580d86ca9ce0837
+size 15920

25_128_e5_3e-5/checkpoint-1410/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04787040b5f7f45e78b754d20a66da73f6951af9a1a1b211fbeb607eebc0094d
+size 15920

25_128_e5_3e-5/checkpoint-1410/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00559ee20324f97789a62f992a031b84aaae856f55e161cc7834614ec6d4a845
+size 15920

25_128_e5_3e-5/checkpoint-1410/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d23b6e3bb9a387c1bae205ac811479ec114ca5231d84f816cb971eea03f4b35
+size 15920

25_128_e5_3e-5/checkpoint-1410/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51a81833599a081f0c808d7725f456c5f9ead44c5d0b19a851c3346cca5f23a7
+size 15920

25_128_e5_3e-5/checkpoint-1410/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a0b442570665ee9378d9b94f6b4b8a35e5c153f1b40c5b104024eddeb40f440
+size 15920

25_128_e5_3e-5/checkpoint-1410/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:644c99374968f9ccb6594c0b4b6ee89cb1a24f9b0dcf3efc3be663149382e32b
+size 15920

25_128_e5_3e-5/checkpoint-1410/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:887fc311d7224c6ac0c71e24fe25428259f0f37455f26a0b80c50ad0805050ea
+size 1064

25_128_e5_3e-5/checkpoint-1410/special_tokens_map.json

@@ -0,0 +1,45 @@
+{
+  "additional_special_tokens": [
+    "<|endoftext|>",
+    "<fim_prefix>",
+    "<fim_middle>",
+    "<fim_suffix>",
+    "<fim_pad>",
+    "<filename>",
+    "<gh_stars>",
+    "<issue_start>",
+    "<issue_comment>",
+    "<issue_closed>",
+    "<jupyter_start>",
+    "<jupyter_text>",
+    "<jupyter_code>",
+    "<jupyter_output>",
+    "<empty_output>",
+    "<commit_before>",
+    "<commit_msg>",
+    "<commit_after>",
+    "<reponame>"
+  ],
+  "bos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "<reponame>",
+  "unk_token": {
+    "content": "<|endoftext|>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

25_128_e5_3e-5/checkpoint-1410/tokenizer_config.json

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

25_128_e5_3e-5/checkpoint-1410/trainer_state.json

@@ -0,0 +1,2008 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 1410,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.017761989342806393,
+      "grad_norm": 1.3582063913345337,
+      "learning_rate": 1.6901408450704227e-06,
+      "loss": 1.3155,
+      "step": 5
+    },
+    {
+      "epoch": 0.035523978685612786,
+      "grad_norm": 0.8357166647911072,
+      "learning_rate": 3.8028169014084508e-06,
+      "loss": 1.3269,
+      "step": 10
+    },
+    {
+      "epoch": 0.05328596802841918,
+      "grad_norm": 0.627669095993042,
+      "learning_rate": 5.915492957746478e-06,
+      "loss": 1.2938,
+      "step": 15
+    },
+    {
+      "epoch": 0.07104795737122557,
+      "grad_norm": 0.6404774785041809,
+      "learning_rate": 8.028169014084507e-06,
+      "loss": 1.3265,
+      "step": 20
+    },
+    {
+      "epoch": 0.08880994671403197,
+      "grad_norm": 0.5879626274108887,
+      "learning_rate": 1.0140845070422535e-05,
+      "loss": 1.2527,
+      "step": 25
+    },
+    {
+      "epoch": 0.10657193605683836,
+      "grad_norm": 0.5322818756103516,
+      "learning_rate": 1.2253521126760564e-05,
+      "loss": 1.2349,
+      "step": 30
+    },
+    {
+      "epoch": 0.12433392539964476,
+      "grad_norm": 0.4639137387275696,
+      "learning_rate": 1.436619718309859e-05,
+      "loss": 1.2609,
+      "step": 35
+    },
+    {
+      "epoch": 0.14209591474245115,
+      "grad_norm": 0.44531938433647156,
+      "learning_rate": 1.6478873239436623e-05,
+      "loss": 1.1992,
+      "step": 40
+    },
+    {
+      "epoch": 0.15985790408525755,
+      "grad_norm": 0.5321743488311768,
+      "learning_rate": 1.8591549295774646e-05,
+      "loss": 1.1719,
+      "step": 45
+    },
+    {
+      "epoch": 0.17761989342806395,
+      "grad_norm": 0.4335007071495056,
+      "learning_rate": 2.0704225352112676e-05,
+      "loss": 1.1815,
+      "step": 50
+    },
+    {
+      "epoch": 0.19538188277087035,
+      "grad_norm": 0.6063855290412903,
+      "learning_rate": 2.2816901408450703e-05,
+      "loss": 1.2039,
+      "step": 55
+    },
+    {
+      "epoch": 0.21314387211367672,
+      "grad_norm": 0.4855096936225891,
+      "learning_rate": 2.4929577464788733e-05,
+      "loss": 1.1404,
+      "step": 60
+    },
+    {
+      "epoch": 0.23090586145648312,
+      "grad_norm": 0.46861350536346436,
+      "learning_rate": 2.7042253521126763e-05,
+      "loss": 1.1783,
+      "step": 65
+    },
+    {
+      "epoch": 0.24866785079928952,
+      "grad_norm": 0.5658775568008423,
+      "learning_rate": 2.915492957746479e-05,
+      "loss": 1.1626,
+      "step": 70
+    },
+    {
+      "epoch": 0.2664298401420959,
+      "grad_norm": 0.49945878982543945,
+      "learning_rate": 2.9999628430060982e-05,
+      "loss": 1.1247,
+      "step": 75
+    },
+    {
+      "epoch": 0.2841918294849023,
+      "grad_norm": 0.5159095525741577,
+      "learning_rate": 2.9997357791542522e-05,
+      "loss": 1.1237,
+      "step": 80
+    },
+    {
+      "epoch": 0.3019538188277087,
+      "grad_norm": 0.5034985542297363,
+      "learning_rate": 2.9993023254350274e-05,
+      "loss": 1.135,
+      "step": 85
+    },
+    {
+      "epoch": 0.3197158081705151,
+      "grad_norm": 0.5856865644454956,
+      "learning_rate": 2.9986625414992358e-05,
+      "loss": 1.0438,
+      "step": 90
+    },
+    {
+      "epoch": 0.33747779751332146,
+      "grad_norm": 0.5520426630973816,
+      "learning_rate": 2.9978165153923333e-05,
+      "loss": 1.0886,
+      "step": 95
+    },
+    {
+      "epoch": 0.3552397868561279,
+      "grad_norm": 0.49657881259918213,
+      "learning_rate": 2.996764363542307e-05,
+      "loss": 1.0517,
+      "step": 100
+    },
+    {
+      "epoch": 0.37300177619893427,
+      "grad_norm": 0.5683771967887878,
+      "learning_rate": 2.995506230743648e-05,
+      "loss": 1.0431,
+      "step": 105
+    },
+    {
+      "epoch": 0.3907637655417407,
+      "grad_norm": 0.5750526785850525,
+      "learning_rate": 2.9940422901374294e-05,
+      "loss": 1.0203,
+      "step": 110
+    },
+    {
+      "epoch": 0.40852575488454707,
+      "grad_norm": 0.5807237029075623,
+      "learning_rate": 2.992372743187477e-05,
+      "loss": 0.9762,
+      "step": 115
+    },
+    {
+      "epoch": 0.42628774422735344,
+      "grad_norm": 0.5997951030731201,
+      "learning_rate": 2.9904978196526445e-05,
+      "loss": 1.0097,
+      "step": 120
+    },
+    {
+      "epoch": 0.44404973357015987,
+      "grad_norm": 0.5774429440498352,
+      "learning_rate": 2.988417777555195e-05,
+      "loss": 0.9927,
+      "step": 125
+    },
+    {
+      "epoch": 0.46181172291296624,
+      "grad_norm": 0.6737991571426392,
+      "learning_rate": 2.986132903145292e-05,
+      "loss": 0.9825,
+      "step": 130
+    },
+    {
+      "epoch": 0.47957371225577267,
+      "grad_norm": 0.7121424078941345,
+      "learning_rate": 2.9836435108616075e-05,
+      "loss": 0.9792,
+      "step": 135
+    },
+    {
+      "epoch": 0.49733570159857904,
+      "grad_norm": 0.8093393445014954,
+      "learning_rate": 2.9809499432880494e-05,
+      "loss": 0.8921,
+      "step": 140
+    },
+    {
+      "epoch": 0.5150976909413855,
+      "grad_norm": 0.8331164121627808,
+      "learning_rate": 2.9780525711066147e-05,
+      "loss": 0.8977,
+      "step": 145
+    },
+    {
+      "epoch": 0.5328596802841918,
+      "grad_norm": 0.710951566696167,
+      "learning_rate": 2.9749517930463785e-05,
+      "loss": 0.942,
+      "step": 150
+    },
+    {
+      "epoch": 0.5506216696269982,
+      "grad_norm": 0.7496849298477173,
+      "learning_rate": 2.971648035828622e-05,
+      "loss": 0.8894,
+      "step": 155
+    },
+    {
+      "epoch": 0.5683836589698046,
+      "grad_norm": 0.7852967381477356,
+      "learning_rate": 2.9681417541081072e-05,
+      "loss": 0.9124,
+      "step": 160
+    },
+    {
+      "epoch": 0.5861456483126111,
+      "grad_norm": 0.8863868713378906,
+      "learning_rate": 2.9644334304105092e-05,
+      "loss": 0.9476,
+      "step": 165
+    },
+    {
+      "epoch": 0.6039076376554174,
+      "grad_norm": 1.0187339782714844,
+      "learning_rate": 2.960523575066012e-05,
+      "loss": 0.8799,
+      "step": 170
+    },
+    {
+      "epoch": 0.6216696269982238,
+      "grad_norm": 0.941071093082428,
+      "learning_rate": 2.956412726139078e-05,
+      "loss": 0.8753,
+      "step": 175
+    },
+    {
+      "epoch": 0.6394316163410302,
+      "grad_norm": 0.7954825758934021,
+      "learning_rate": 2.9521014493544017e-05,
+      "loss": 0.8491,
+      "step": 180
+    },
+    {
+      "epoch": 0.6571936056838366,
+      "grad_norm": 0.837969183921814,
+      "learning_rate": 2.9475903380190547e-05,
+      "loss": 0.8343,
+      "step": 185
+    },
+    {
+      "epoch": 0.6749555950266429,
+      "grad_norm": 0.971666157245636,
+      "learning_rate": 2.9428800129408374e-05,
+      "loss": 0.8167,
+      "step": 190
+    },
+    {
+      "epoch": 0.6927175843694494,
+      "grad_norm": 0.8235210180282593,
+      "learning_rate": 2.9379711223428446e-05,
+      "loss": 0.8028,
+      "step": 195
+    },
+    {
+      "epoch": 0.7104795737122558,
+      "grad_norm": 0.8552044034004211,
+      "learning_rate": 2.9328643417742592e-05,
+      "loss": 0.8244,
+      "step": 200
+    },
+    {
+      "epoch": 0.7282415630550622,
+      "grad_norm": 0.92862468957901,
+      "learning_rate": 2.9275603740173835e-05,
+      "loss": 0.8089,
+      "step": 205
+    },
+    {
+      "epoch": 0.7460035523978685,
+      "grad_norm": 0.9003951549530029,
+      "learning_rate": 2.9220599489909254e-05,
+      "loss": 0.7295,
+      "step": 210
+    },
+    {
+      "epoch": 0.7637655417406749,
+      "grad_norm": 1.0426762104034424,
+      "learning_rate": 2.916363823649548e-05,
+      "loss": 0.7462,
+      "step": 215
+    },
+    {
+      "epoch": 0.7815275310834814,
+      "grad_norm": 0.8161042332649231,
+      "learning_rate": 2.9104727818797e-05,
+      "loss": 0.7455,
+      "step": 220
+    },
+    {
+      "epoch": 0.7992895204262878,
+      "grad_norm": 0.9472765326499939,
+      "learning_rate": 2.904387634391739e-05,
+      "loss": 0.752,
+      "step": 225
+    },
+    {
+      "epoch": 0.8170515097690941,
+      "grad_norm": 0.9386636018753052,
+      "learning_rate": 2.8981092186083632e-05,
+      "loss": 0.7525,
+      "step": 230
+    },
+    {
+      "epoch": 0.8348134991119005,
+      "grad_norm": 1.1699973344802856,
+      "learning_rate": 2.8916383985493676e-05,
+      "loss": 0.7123,
+      "step": 235
+    },
+    {
+      "epoch": 0.8525754884547069,
+      "grad_norm": 0.8852489590644836,
+      "learning_rate": 2.884976064712741e-05,
+      "loss": 0.698,
+      "step": 240
+    },
+    {
+      "epoch": 0.8703374777975134,
+      "grad_norm": 1.0662732124328613,
+      "learning_rate": 2.878123133952115e-05,
+      "loss": 0.7178,
+      "step": 245
+    },
+    {
+      "epoch": 0.8880994671403197,
+      "grad_norm": 1.2624017000198364,
+      "learning_rate": 2.8710805493505922e-05,
+      "loss": 0.6831,
+      "step": 250
+    },
+    {
+      "epoch": 0.9058614564831261,
+      "grad_norm": 1.024356484413147,
+      "learning_rate": 2.863849280090959e-05,
+      "loss": 0.6938,
+      "step": 255
+    },
+    {
+      "epoch": 0.9236234458259325,
+      "grad_norm": 1.0030966997146606,
+      "learning_rate": 2.856430321322311e-05,
+      "loss": 0.6796,
+      "step": 260
+    },
+    {
+      "epoch": 0.9413854351687388,
+      "grad_norm": 1.0371718406677246,
+      "learning_rate": 2.848824694023102e-05,
+      "loss": 0.6981,
+      "step": 265
+    },
+    {
+      "epoch": 0.9591474245115453,
+      "grad_norm": 0.970425546169281,
+      "learning_rate": 2.8410334448606377e-05,
+      "loss": 0.6895,
+      "step": 270
+    },
+    {
+      "epoch": 0.9769094138543517,
+      "grad_norm": 0.9296903014183044,
+      "learning_rate": 2.8330576460470407e-05,
+      "loss": 0.6514,
+      "step": 275
+    },
+    {
+      "epoch": 0.9946714031971581,
+      "grad_norm": 0.9770685434341431,
+      "learning_rate": 2.824898395191691e-05,
+      "loss": 0.6336,
+      "step": 280
+    },
+    {
+      "epoch": 1.0106571936056838,
+      "grad_norm": 1.1167672872543335,
+      "learning_rate": 2.816556815150179e-05,
+      "loss": 0.5954,
+      "step": 285
+    },
+    {
+      "epoch": 1.0284191829484903,
+      "grad_norm": 1.0073131322860718,
+      "learning_rate": 2.808034053869776e-05,
+      "loss": 0.5953,
+      "step": 290
+    },
+    {
+      "epoch": 1.0461811722912966,
+      "grad_norm": 1.1674365997314453,
+      "learning_rate": 2.7993312842314636e-05,
+      "loss": 0.5708,
+      "step": 295
+    },
+    {
+      "epoch": 1.063943161634103,
+      "grad_norm": 1.1259663105010986,
+      "learning_rate": 2.7904497038885184e-05,
+      "loss": 0.5943,
+      "step": 300
+    },
+    {
+      "epoch": 1.0817051509769093,
+      "grad_norm": 1.0642991065979004,
+      "learning_rate": 2.7813905351016984e-05,
+      "loss": 0.5721,
+      "step": 305
+    },
+    {
+      "epoch": 1.0994671403197158,
+      "grad_norm": 1.2667790651321411,
+      "learning_rate": 2.7721550245710362e-05,
+      "loss": 0.577,
+      "step": 310
+    },
+    {
+      "epoch": 1.1172291296625223,
+      "grad_norm": 1.0688438415527344,
+      "learning_rate": 2.7627444432642708e-05,
+      "loss": 0.5548,
+      "step": 315
+    },
+    {
+      "epoch": 1.1349911190053286,
+      "grad_norm": 1.0694912672042847,
+      "learning_rate": 2.7531600862419422e-05,
+      "loss": 0.5499,
+      "step": 320
+    },
+    {
+      "epoch": 1.152753108348135,
+      "grad_norm": 1.2388120889663696,
+      "learning_rate": 2.743403272479168e-05,
+      "loss": 0.5154,
+      "step": 325
+    },
+    {
+      "epoch": 1.1705150976909413,
+      "grad_norm": 1.31024169921875,
+      "learning_rate": 2.7334753446841277e-05,
+      "loss": 0.5227,
+      "step": 330
+    },
+    {
+      "epoch": 1.1882770870337478,
+      "grad_norm": 1.1155751943588257,
+      "learning_rate": 2.7233776691132848e-05,
+      "loss": 0.5062,
+      "step": 335
+    },
+    {
+      "epoch": 1.206039076376554,
+      "grad_norm": 1.3883517980575562,
+      "learning_rate": 2.7131116353833633e-05,
+      "loss": 0.5329,
+      "step": 340
+    },
+    {
+      "epoch": 1.2238010657193605,
+      "grad_norm": 1.2181286811828613,
+      "learning_rate": 2.702678656280116e-05,
+      "loss": 0.5268,
+      "step": 345
+    },
+    {
+      "epoch": 1.241563055062167,
+      "grad_norm": 1.1056498289108276,
+      "learning_rate": 2.6920801675638965e-05,
+      "loss": 0.5769,
+      "step": 350
+    },
+    {
+      "epoch": 1.2593250444049733,
+      "grad_norm": 1.0863943099975586,
+      "learning_rate": 2.681317627772076e-05,
+      "loss": 0.5136,
+      "step": 355
+    },
+    {
+      "epoch": 1.2770870337477798,
+      "grad_norm": 1.2052078247070312,
+      "learning_rate": 2.670392518018322e-05,
+      "loss": 0.5167,
+      "step": 360
+    },
+    {
+      "epoch": 1.294849023090586,
+      "grad_norm": 1.2922052145004272,
+      "learning_rate": 2.6593063417887713e-05,
+      "loss": 0.5198,
+      "step": 365
+    },
+    {
+      "epoch": 1.3126110124333925,
+      "grad_norm": 1.155900239944458,
+      "learning_rate": 2.6480606247351247e-05,
+      "loss": 0.5878,
+      "step": 370
+    },
+    {
+      "epoch": 1.330373001776199,
+      "grad_norm": 1.140750765800476,
+      "learning_rate": 2.6366569144646903e-05,
+      "loss": 0.5175,
+      "step": 375
+    },
+    {
+      "epoch": 1.3481349911190053,
+      "grad_norm": 1.2477790117263794,
+      "learning_rate": 2.625096780327404e-05,
+      "loss": 0.496,
+      "step": 380
+    },
+    {
+      "epoch": 1.3658969804618117,
+      "grad_norm": 1.0702998638153076,
+      "learning_rate": 2.6133818131998622e-05,
+      "loss": 0.4407,
+      "step": 385
+    },
+    {
+      "epoch": 1.383658969804618,
+      "grad_norm": 1.0303690433502197,
+      "learning_rate": 2.601513625266388e-05,
+      "loss": 0.484,
+      "step": 390
+    },
+    {
+      "epoch": 1.4014209591474245,
+      "grad_norm": 1.126442313194275,
+      "learning_rate": 2.5894938497971668e-05,
+      "loss": 0.4645,
+      "step": 395
+    },
+    {
+      "epoch": 1.419182948490231,
+      "grad_norm": 1.166357398033142,
+      "learning_rate": 2.5773241409234788e-05,
+      "loss": 0.4727,
+      "step": 400
+    },
+    {
+      "epoch": 1.4369449378330372,
+      "grad_norm": 1.2368241548538208,
+      "learning_rate": 2.5650061734100633e-05,
+      "loss": 0.4924,
+      "step": 405
+    },
+    {
+      "epoch": 1.4547069271758437,
+      "grad_norm": 1.234560489654541,
+      "learning_rate": 2.55254164242464e-05,
+      "loss": 0.4303,
+      "step": 410
+    },
+    {
+      "epoch": 1.47246891651865,
+      "grad_norm": 1.1312541961669922,
+      "learning_rate": 2.5399322633046254e-05,
+      "loss": 0.4484,
+      "step": 415
+    },
+    {
+      "epoch": 1.4902309058614565,
+      "grad_norm": 1.03053617477417,
+      "learning_rate": 2.5271797713210726e-05,
+      "loss": 0.4219,
+      "step": 420
+    },
+    {
+      "epoch": 1.507992895204263,
+      "grad_norm": 1.2007708549499512,
+      "learning_rate": 2.5142859214398666e-05,
+      "loss": 0.4475,
+      "step": 425
+    },
+    {
+      "epoch": 1.5257548845470694,
+      "grad_norm": 1.071448802947998,
+      "learning_rate": 2.5012524880802108e-05,
+      "loss": 0.4191,
+      "step": 430
+    },
+    {
+      "epoch": 1.5435168738898757,
+      "grad_norm": 1.225663185119629,
+      "learning_rate": 2.4880812648704367e-05,
+      "loss": 0.441,
+      "step": 435
+    },
+    {
+      "epoch": 1.561278863232682,
+      "grad_norm": 1.0680769681930542,
+      "learning_rate": 2.4747740644011682e-05,
+      "loss": 0.4541,
+      "step": 440
+    },
+    {
+      "epoch": 1.5790408525754884,
+      "grad_norm": 1.0855811834335327,
+      "learning_rate": 2.4613327179758773e-05,
+      "loss": 0.4428,
+      "step": 445
+    },
+    {
+      "epoch": 1.596802841918295,
+      "grad_norm": 1.197534441947937,
+      "learning_rate": 2.4477590753588656e-05,
+      "loss": 0.4217,
+      "step": 450
+    },
+    {
+      "epoch": 1.6145648312611014,
+      "grad_norm": 1.1745901107788086,
+      "learning_rate": 2.434055004520704e-05,
+      "loss": 0.4236,
+      "step": 455
+    },
+    {
+      "epoch": 1.6323268206039077,
+      "grad_norm": 1.2637349367141724,
+      "learning_rate": 2.420222391381168e-05,
+      "loss": 0.4819,
+      "step": 460
+    },
+    {
+      "epoch": 1.650088809946714,
+      "grad_norm": 1.2748724222183228,
+      "learning_rate": 2.4062631395496993e-05,
+      "loss": 0.4384,
+      "step": 465
+    },
+    {
+      "epoch": 1.6678507992895204,
+      "grad_norm": 1.0431526899337769,
+      "learning_rate": 2.392179170063441e-05,
+      "loss": 0.3863,
+      "step": 470
+    },
+    {
+      "epoch": 1.6856127886323269,
+      "grad_norm": 1.1915509700775146,
+      "learning_rate": 2.3779724211228632e-05,
+      "loss": 0.4026,
+      "step": 475
+    },
+    {
+      "epoch": 1.7033747779751334,
+      "grad_norm": 1.1567018032073975,
+      "learning_rate": 2.3636448478250378e-05,
+      "loss": 0.3837,
+      "step": 480
+    },
+    {
+      "epoch": 1.7211367673179396,
+      "grad_norm": 1.1712560653686523,
+      "learning_rate": 2.3491984218945814e-05,
+      "loss": 0.4064,
+      "step": 485
+    },
+    {
+      "epoch": 1.738898756660746,
+      "grad_norm": 1.1683224439620972,
+      "learning_rate": 2.3346351314123093e-05,
+      "loss": 0.4456,
+      "step": 490
+    },
+    {
+      "epoch": 1.7566607460035524,
+      "grad_norm": 1.2679800987243652,
+      "learning_rate": 2.3199569805416436e-05,
+      "loss": 0.4393,
+      "step": 495
+    },
+    {
+      "epoch": 1.7744227353463589,
+      "grad_norm": 1.3210681676864624,
+      "learning_rate": 2.305165989252804e-05,
+      "loss": 0.3742,
+      "step": 500
+    },
+    {
+      "epoch": 1.7921847246891653,
+      "grad_norm": 1.0676344633102417,
+      "learning_rate": 2.2902641930448253e-05,
+      "loss": 0.3477,
+      "step": 505
+    },
+    {
+      "epoch": 1.8099467140319716,
+      "grad_norm": 1.1912447214126587,
+      "learning_rate": 2.2752536426654363e-05,
+      "loss": 0.3905,
+      "step": 510
+    },
+    {
+      "epoch": 1.8277087033747779,
+      "grad_norm": 1.1013745069503784,
+      "learning_rate": 2.260136403828842e-05,
+      "loss": 0.3606,
+      "step": 515
+    },
+    {
+      "epoch": 1.8454706927175843,
+      "grad_norm": 1.0575093030929565,
+      "learning_rate": 2.2449145569314437e-05,
+      "loss": 0.3727,
+      "step": 520
+    },
+    {
+      "epoch": 1.8632326820603908,
+      "grad_norm": 1.0612674951553345,
+      "learning_rate": 2.229590196765541e-05,
+      "loss": 0.3739,
+      "step": 525
+    },
+    {
+      "epoch": 1.8809946714031973,
+      "grad_norm": 1.2063333988189697,
+      "learning_rate": 2.2141654322310534e-05,
+      "loss": 0.3485,
+      "step": 530
+    },
+    {
+      "epoch": 1.8987566607460036,
+      "grad_norm": 1.1345680952072144,
+      "learning_rate": 2.1986423860452957e-05,
+      "loss": 0.3975,
+      "step": 535
+    },
+    {
+      "epoch": 1.9165186500888098,
+      "grad_norm": 1.2932173013687134,
+      "learning_rate": 2.183023194450857e-05,
+      "loss": 0.3558,
+      "step": 540
+    },
+    {
+      "epoch": 1.9342806394316163,
+      "grad_norm": 1.1968685388565063,
+      "learning_rate": 2.167310006921615e-05,
+      "loss": 0.3387,
+      "step": 545
+    },
+    {
+      "epoch": 1.9520426287744228,
+      "grad_norm": 1.1290210485458374,
+      "learning_rate": 2.151504985866933e-05,
+      "loss": 0.3449,
+      "step": 550
+    },
+    {
+      "epoch": 1.9698046181172293,
+      "grad_norm": 1.1544890403747559,
+      "learning_rate": 2.1356103063340716e-05,
+      "loss": 0.3698,
+      "step": 555
+    },
+    {
+      "epoch": 1.9875666074600356,
+      "grad_norm": 1.199317216873169,
+      "learning_rate": 2.1196281557088674e-05,
+      "loss": 0.3192,
+      "step": 560
+    },
+    {
+      "epoch": 2.003552397868561,
+      "grad_norm": 1.2987078428268433,
+      "learning_rate": 2.1035607334147083e-05,
+      "loss": 0.3253,
+      "step": 565
+    },
+    {
+      "epoch": 2.0213143872113677,
+      "grad_norm": 1.2794301509857178,
+      "learning_rate": 2.087410250609854e-05,
+      "loss": 0.2469,
+      "step": 570
+    },
+    {
+      "epoch": 2.039076376554174,
+      "grad_norm": 1.2110157012939453,
+      "learning_rate": 2.0711789298831435e-05,
+      "loss": 0.2784,
+      "step": 575
+    },
+    {
+      "epoch": 2.0568383658969807,
+      "grad_norm": 1.1234550476074219,
+      "learning_rate": 2.0548690049481274e-05,
+      "loss": 0.2782,
+      "step": 580
+    },
+    {
+      "epoch": 2.0746003552397867,
+      "grad_norm": 1.2034980058670044,
+      "learning_rate": 2.0384827203356702e-05,
+      "loss": 0.2768,
+      "step": 585
+    },
+    {
+      "epoch": 2.092362344582593,
+      "grad_norm": 1.1906369924545288,
+      "learning_rate": 2.0220223310850625e-05,
+      "loss": 0.2756,
+      "step": 590
+    },
+    {
+      "epoch": 2.1101243339253997,
+      "grad_norm": 1.1176049709320068,
+      "learning_rate": 2.0054901024336896e-05,
+      "loss": 0.262,
+      "step": 595
+    },
+    {
+      "epoch": 2.127886323268206,
+      "grad_norm": 1.0588326454162598,
+      "learning_rate": 1.988888309505293e-05,
+      "loss": 0.2821,
+      "step": 600
+    },
+    {
+      "epoch": 2.1456483126110126,
+      "grad_norm": 1.1338000297546387,
+      "learning_rate": 1.9722192369968776e-05,
+      "loss": 0.2613,
+      "step": 605
+    },
+    {
+      "epoch": 2.1634103019538187,
+      "grad_norm": 1.1807286739349365,
+      "learning_rate": 1.9554851788642928e-05,
+      "loss": 0.2339,
+      "step": 610
+    },
+    {
+      "epoch": 2.181172291296625,
+      "grad_norm": 1.0721327066421509,
+      "learning_rate": 1.9386884380065474e-05,
+      "loss": 0.2644,
+      "step": 615
+    },
+    {
+      "epoch": 2.1989342806394316,
+      "grad_norm": 1.0247159004211426,
+      "learning_rate": 1.9218313259488883e-05,
+      "loss": 0.2627,
+      "step": 620
+    },
+    {
+      "epoch": 2.216696269982238,
+      "grad_norm": 1.3054232597351074,
+      "learning_rate": 1.9049161625246962e-05,
+      "loss": 0.2381,
+      "step": 625
+    },
+    {
+      "epoch": 2.2344582593250446,
+      "grad_norm": 1.1615649461746216,
+      "learning_rate": 1.8879452755562344e-05,
+      "loss": 0.249,
+      "step": 630
+    },
+    {
+      "epoch": 2.2522202486678506,
+      "grad_norm": 1.5350950956344604,
+      "learning_rate": 1.8709210005343e-05,
+      "loss": 0.242,
+      "step": 635
+    },
+    {
+      "epoch": 2.269982238010657,
+      "grad_norm": 1.1368858814239502,
+      "learning_rate": 1.85384568029682e-05,
+      "loss": 0.236,
+      "step": 640
+    },
+    {
+      "epoch": 2.2877442273534636,
+      "grad_norm": 1.163325309753418,
+      "learning_rate": 1.836721664706437e-05,
+      "loss": 0.2277,
+      "step": 645
+    },
+    {
+      "epoch": 2.30550621669627,
+      "grad_norm": 1.2359697818756104,
+      "learning_rate": 1.819551310327122e-05,
+      "loss": 0.2626,
+      "step": 650
+    },
+    {
+      "epoch": 2.323268206039076,
+      "grad_norm": 1.1182560920715332,
+      "learning_rate": 1.8023369800998776e-05,
+      "loss": 0.2628,
+      "step": 655
+    },
+    {
+      "epoch": 2.3410301953818826,
+      "grad_norm": 1.1834044456481934,
+      "learning_rate": 1.78508104301755e-05,
+      "loss": 0.2552,
+      "step": 660
+    },
+    {
+      "epoch": 2.358792184724689,
+      "grad_norm": 1.1643848419189453,
+      "learning_rate": 1.767785873798817e-05,
+      "loss": 0.2402,
+      "step": 665
+    },
+    {
+      "epoch": 2.3765541740674956,
+      "grad_norm": 1.1760133504867554,
+      "learning_rate": 1.7504538525613833e-05,
+      "loss": 0.236,
+      "step": 670
+    },
+    {
+      "epoch": 2.394316163410302,
+      "grad_norm": 1.2071462869644165,
+      "learning_rate": 1.7330873644944367e-05,
+      "loss": 0.2473,
+      "step": 675
+    },
+    {
+      "epoch": 2.412078152753108,
+      "grad_norm": 1.2645944356918335,
+      "learning_rate": 1.715688799530403e-05,
+      "loss": 0.2245,
+      "step": 680
+    },
+    {
+      "epoch": 2.4298401420959146,
+      "grad_norm": 1.1790143251419067,
+      "learning_rate": 1.6982605520160503e-05,
+      "loss": 0.2484,
+      "step": 685
+    },
+    {
+      "epoch": 2.447602131438721,
+      "grad_norm": 1.1275826692581177,
+      "learning_rate": 1.6808050203829845e-05,
+      "loss": 0.2169,
+      "step": 690
+    },
+    {
+      "epoch": 2.4653641207815276,
+      "grad_norm": 1.1996047496795654,
+      "learning_rate": 1.6633246068175833e-05,
+      "loss": 0.2278,
+      "step": 695
+    },
+    {
+      "epoch": 2.483126110124334,
+      "grad_norm": 1.18192458152771,
+      "learning_rate": 1.6458217169304126e-05,
+      "loss": 0.2351,
+      "step": 700
+    },
+    {
+      "epoch": 2.50088809946714,
+      "grad_norm": 1.0418000221252441,
+      "learning_rate": 1.628298759425175e-05,
+      "loss": 0.2297,
+      "step": 705
+    },
+    {
+      "epoch": 2.5186500888099466,
+      "grad_norm": 1.2126238346099854,
+      "learning_rate": 1.6107581457672267e-05,
+      "loss": 0.2396,
+      "step": 710
+    },
+    {
+      "epoch": 2.536412078152753,
+      "grad_norm": 1.1965978145599365,
+      "learning_rate": 1.5932022898517195e-05,
+      "loss": 0.2124,
+      "step": 715
+    },
+    {
+      "epoch": 2.5541740674955595,
+      "grad_norm": 1.1391175985336304,
+      "learning_rate": 1.5756336076714065e-05,
+      "loss": 0.2021,
+      "step": 720
+    },
+    {
+      "epoch": 2.571936056838366,
+      "grad_norm": 1.0899744033813477,
+      "learning_rate": 1.5580545169841592e-05,
+      "loss": 0.2178,
+      "step": 725
+    },
+    {
+      "epoch": 2.589698046181172,
+      "grad_norm": 1.2015151977539062,
+      "learning_rate": 1.5404674369802425e-05,
+      "loss": 0.2487,
+      "step": 730
+    },
+    {
+      "epoch": 2.6074600355239785,
+      "grad_norm": 1.1883600950241089,
+      "learning_rate": 1.522874787949389e-05,
+      "loss": 0.1964,
+      "step": 735
+    },
+    {
+      "epoch": 2.625222024866785,
+      "grad_norm": 1.173923373222351,
+      "learning_rate": 1.5052789909477286e-05,
+      "loss": 0.2199,
+      "step": 740
+    },
+    {
+      "epoch": 2.6429840142095915,
+      "grad_norm": 1.0762274265289307,
+      "learning_rate": 1.4876824674646062e-05,
+      "loss": 0.2259,
+      "step": 745
+    },
+    {
+      "epoch": 2.660746003552398,
+      "grad_norm": 1.0043290853500366,
+      "learning_rate": 1.4700876390893426e-05,
+      "loss": 0.2139,
+      "step": 750
+    },
+    {
+      "epoch": 2.678507992895204,
+      "grad_norm": 1.175076961517334,
+      "learning_rate": 1.4524969271779844e-05,
+      "loss": 0.1997,
+      "step": 755
+    },
+    {
+      "epoch": 2.6962699822380105,
+      "grad_norm": 1.1845417022705078,
+      "learning_rate": 1.4349127525200783e-05,
+      "loss": 0.226,
+      "step": 760
+    },
+    {
+      "epoch": 2.714031971580817,
+      "grad_norm": 1.1732301712036133,
+      "learning_rate": 1.4173375350055319e-05,
+      "loss": 0.2154,
+      "step": 765
+    },
+    {
+      "epoch": 2.7317939609236235,
+      "grad_norm": 1.0934293270111084,
+      "learning_rate": 1.3997736932915922e-05,
+      "loss": 0.1878,
+      "step": 770
+    },
+    {
+      "epoch": 2.74955595026643,
+      "grad_norm": 1.2918049097061157,
+      "learning_rate": 1.3822236444699991e-05,
+      "loss": 0.2082,
+      "step": 775
+    },
+    {
+      "epoch": 2.767317939609236,
+      "grad_norm": 1.0803861618041992,
+      "learning_rate": 1.3646898037343475e-05,
+      "loss": 0.199,
+      "step": 780
+    },
+    {
+      "epoch": 2.7850799289520425,
+      "grad_norm": 1.2331840991973877,
+      "learning_rate": 1.347174584047717e-05,
+      "loss": 0.1915,
+      "step": 785
+    },
+    {
+      "epoch": 2.802841918294849,
+      "grad_norm": 1.2146940231323242,
+      "learning_rate": 1.3296803958106046e-05,
+      "loss": 0.1903,
+      "step": 790
+    },
+    {
+      "epoch": 2.8206039076376554,
+      "grad_norm": 1.3104547262191772,
+      "learning_rate": 1.3122096465292132e-05,
+      "loss": 0.1997,
+      "step": 795
+    },
+    {
+      "epoch": 2.838365896980462,
+      "grad_norm": 1.0410480499267578,
+      "learning_rate": 1.294764740484135e-05,
+      "loss": 0.1848,
+      "step": 800
+    },
+    {
+      "epoch": 2.856127886323268,
+      "grad_norm": 1.2668956518173218,
+      "learning_rate": 1.2773480783994805e-05,
+      "loss": 0.1937,
+      "step": 805
+    },
+    {
+      "epoch": 2.8738898756660745,
+      "grad_norm": 1.1789441108703613,
+      "learning_rate": 1.259962057112498e-05,
+      "loss": 0.2104,
+      "step": 810
+    },
+    {
+      "epoch": 2.891651865008881,
+      "grad_norm": 1.1638792753219604,
+      "learning_rate": 1.2426090692437266e-05,
+      "loss": 0.1726,
+      "step": 815
+    },
+    {
+      "epoch": 2.9094138543516874,
+      "grad_norm": 1.0087004899978638,
+      "learning_rate": 1.225291502867728e-05,
+      "loss": 0.187,
+      "step": 820
+    },
+    {
+      "epoch": 2.927175843694494,
+      "grad_norm": 1.1213196516036987,
+      "learning_rate": 1.208011741184449e-05,
+      "loss": 0.1714,
+      "step": 825
+    },
+    {
+      "epoch": 2.9449378330373,
+      "grad_norm": 1.1991121768951416,
+      "learning_rate": 1.1907721621912476e-05,
+      "loss": 0.185,
+      "step": 830
+    },
+    {
+      "epoch": 2.9626998223801064,
+      "grad_norm": 1.0833150148391724,
+      "learning_rate": 1.1735751383556432e-05,
+      "loss": 0.1779,
+      "step": 835
+    },
+    {
+      "epoch": 2.980461811722913,
+      "grad_norm": 1.2620830535888672,
+      "learning_rate": 1.1564230362888195e-05,
+      "loss": 0.205,
+      "step": 840
+    },
+    {
+      "epoch": 2.9982238010657194,
+      "grad_norm": 1.1784212589263916,
+      "learning_rate": 1.1393182164199404e-05,
+      "loss": 0.2045,
+      "step": 845
+    },
+    {
+      "epoch": 3.0142095914742453,
+      "grad_norm": 1.0905832052230835,
+      "learning_rate": 1.1222630326713122e-05,
+      "loss": 0.1293,
+      "step": 850
+    },
+    {
+      "epoch": 3.0319715808170513,
+      "grad_norm": 1.245754599571228,
+      "learning_rate": 1.1052598321344435e-05,
+      "loss": 0.1574,
+      "step": 855
+    },
+    {
+      "epoch": 3.049733570159858,
+      "grad_norm": 0.9812668561935425,
+      "learning_rate": 1.0883109547470435e-05,
+      "loss": 0.1259,
+      "step": 860
+    },
+    {
+      "epoch": 3.0674955595026643,
+      "grad_norm": 1.1846905946731567,
+      "learning_rate": 1.0714187329710066e-05,
+      "loss": 0.1286,
+      "step": 865
+    },
+    {
+      "epoch": 3.0852575488454708,
+      "grad_norm": 1.0834468603134155,
+      "learning_rate": 1.0545854914714233e-05,
+      "loss": 0.1506,
+      "step": 870
+    },
+    {
+      "epoch": 3.1030195381882772,
+      "grad_norm": 1.082939863204956,
+      "learning_rate": 1.0378135467966691e-05,
+      "loss": 0.1557,
+      "step": 875
+    },
+    {
+      "epoch": 3.1207815275310833,
+      "grad_norm": 1.0465799570083618,
+      "learning_rate": 1.021105207059602e-05,
+      "loss": 0.1127,
+      "step": 880
+    },
+    {
+      "epoch": 3.1385435168738898,
+      "grad_norm": 1.0733388662338257,
+      "learning_rate": 1.0044627716199297e-05,
+      "loss": 0.1394,
+      "step": 885
+    },
+    {
+      "epoch": 3.1563055062166963,
+      "grad_norm": 1.2737563848495483,
+      "learning_rate": 9.878885307677756e-06,
+      "loss": 0.1355,
+      "step": 890
+    },
+    {
+      "epoch": 3.1740674955595027,
+      "grad_norm": 1.2421573400497437,
+      "learning_rate": 9.713847654084966e-06,
+      "loss": 0.1396,
+      "step": 895
+    },
+    {
+      "epoch": 3.191829484902309,
+      "grad_norm": 1.1305955648422241,
+      "learning_rate": 9.549537467487884e-06,
+      "loss": 0.1505,
+      "step": 900
+    },
+    {
+      "epoch": 3.2095914742451153,
+      "grad_norm": 0.9503009915351868,
+      "learning_rate": 9.385977359841287e-06,
+      "loss": 0.1492,
+      "step": 905
+    },
+    {
+      "epoch": 3.2273534635879217,
+      "grad_norm": 1.04369056224823,
+      "learning_rate": 9.22318983987597e-06,
+      "loss": 0.1422,
+      "step": 910
+    },
+    {
+      "epoch": 3.2451154529307282,
+      "grad_norm": 1.144344449043274,
+      "learning_rate": 9.061197310001176e-06,
+      "loss": 0.1164,
+      "step": 915
+    },
+    {
+      "epoch": 3.2628774422735347,
+      "grad_norm": 1.3119343519210815,
+      "learning_rate": 8.900022063221582e-06,
+      "loss": 0.1344,
+      "step": 920
+    },
+    {
+      "epoch": 3.280639431616341,
+      "grad_norm": 1.1138067245483398,
+      "learning_rate": 8.739686280069435e-06,
+      "loss": 0.1319,
+      "step": 925
+    },
+    {
+      "epoch": 3.2984014209591472,
+      "grad_norm": 1.103514313697815,
+      "learning_rate": 8.580212025552108e-06,
+      "loss": 0.1404,
+      "step": 930
+    },
+    {
+      "epoch": 3.3161634103019537,
+      "grad_norm": 1.3191359043121338,
+      "learning_rate": 8.421621246115565e-06,
+      "loss": 0.1382,
+      "step": 935
+    },
+    {
+      "epoch": 3.33392539964476,
+      "grad_norm": 0.9715001583099365,
+      "learning_rate": 8.263935766624145e-06,
+      "loss": 0.1267,
+      "step": 940
+    },
+    {
+      "epoch": 3.3516873889875667,
+      "grad_norm": 1.1174882650375366,
+      "learning_rate": 8.107177287357104e-06,
+      "loss": 0.1309,
+      "step": 945
+    },
+    {
+      "epoch": 3.369449378330373,
+      "grad_norm": 1.1527106761932373,
+      "learning_rate": 7.951367381022243e-06,
+      "loss": 0.1152,
+      "step": 950
+    },
+    {
+      "epoch": 3.387211367673179,
+      "grad_norm": 1.0667994022369385,
+      "learning_rate": 7.79652748978714e-06,
+      "loss": 0.1183,
+      "step": 955
+    },
+    {
+      "epoch": 3.4049733570159857,
+      "grad_norm": 0.9557608366012573,
+      "learning_rate": 7.642678922328335e-06,
+      "loss": 0.1242,
+      "step": 960
+    },
+    {
+      "epoch": 3.422735346358792,
+      "grad_norm": 0.9632545709609985,
+      "learning_rate": 7.489842850898886e-06,
+      "loss": 0.1224,
+      "step": 965
+    },
+    {
+      "epoch": 3.4404973357015987,
+      "grad_norm": 0.9741002321243286,
+      "learning_rate": 7.338040308414697e-06,
+      "loss": 0.1135,
+      "step": 970
+    },
+    {
+      "epoch": 3.458259325044405,
+      "grad_norm": 0.9288985729217529,
+      "learning_rate": 7.187292185560008e-06,
+      "loss": 0.1168,
+      "step": 975
+    },
+    {
+      "epoch": 3.476021314387211,
+      "grad_norm": 1.246943473815918,
+      "learning_rate": 7.037619227912484e-06,
+      "loss": 0.1204,
+      "step": 980
+    },
+    {
+      "epoch": 3.4937833037300177,
+      "grad_norm": 0.9657527804374695,
+      "learning_rate": 6.889042033088256e-06,
+      "loss": 0.1146,
+      "step": 985
+    },
+    {
+      "epoch": 3.511545293072824,
+      "grad_norm": 1.037055492401123,
+      "learning_rate": 6.741581047907329e-06,
+      "loss": 0.1127,
+      "step": 990
+    },
+    {
+      "epoch": 3.5293072824156306,
+      "grad_norm": 0.9712250232696533,
+      "learning_rate": 6.5952565655797535e-06,
+      "loss": 0.1154,
+      "step": 995
+    },
+    {
+      "epoch": 3.5470692717584367,
+      "grad_norm": 1.075392723083496,
+      "learning_rate": 6.450088722912902e-06,
+      "loss": 0.1187,
+      "step": 1000
+    },
+    {
+      "epoch": 3.5648312611012436,
+      "grad_norm": 0.8246014714241028,
+      "learning_rate": 6.3060974975403115e-06,
+      "loss": 0.1035,
+      "step": 1005
+    },
+    {
+      "epoch": 3.5825932504440496,
+      "grad_norm": 0.9024150371551514,
+      "learning_rate": 6.1633027051724024e-06,
+      "loss": 0.1285,
+      "step": 1010
+    },
+    {
+      "epoch": 3.600355239786856,
+      "grad_norm": 0.8948878049850464,
+      "learning_rate": 6.0217239968694875e-06,
+      "loss": 0.1219,
+      "step": 1015
+    },
+    {
+      "epoch": 3.6181172291296626,
+      "grad_norm": 1.0379226207733154,
+      "learning_rate": 5.881380856337466e-06,
+      "loss": 0.1186,
+      "step": 1020
+    },
+    {
+      "epoch": 3.6358792184724686,
+      "grad_norm": 0.9066540598869324,
+      "learning_rate": 5.742292597246491e-06,
+      "loss": 0.1026,
+      "step": 1025
+    },
+    {
+      "epoch": 3.6536412078152756,
+      "grad_norm": 1.2567459344863892,
+      "learning_rate": 5.604478360573089e-06,
+      "loss": 0.1164,
+      "step": 1030
+    },
+    {
+      "epoch": 3.6714031971580816,
+      "grad_norm": 0.9321945905685425,
+      "learning_rate": 5.467957111966023e-06,
+      "loss": 0.1016,
+      "step": 1035
+    },
+    {
+      "epoch": 3.689165186500888,
+      "grad_norm": 0.9881175756454468,
+      "learning_rate": 5.3327476391363015e-06,
+      "loss": 0.1046,
+      "step": 1040
+    },
+    {
+      "epoch": 3.7069271758436946,
+      "grad_norm": 1.0275404453277588,
+      "learning_rate": 5.198868549271635e-06,
+      "loss": 0.1162,
+      "step": 1045
+    },
+    {
+      "epoch": 3.7246891651865006,
+      "grad_norm": 1.2424781322479248,
+      "learning_rate": 5.06633826647578e-06,
+      "loss": 0.1221,
+      "step": 1050
+    },
+    {
+      "epoch": 3.7424511545293075,
+      "grad_norm": 1.0425161123275757,
+      "learning_rate": 4.935175029233048e-06,
+      "loss": 0.1044,
+      "step": 1055
+    },
+    {
+      "epoch": 3.7602131438721136,
+      "grad_norm": 1.0624982118606567,
+      "learning_rate": 4.805396887898397e-06,
+      "loss": 0.1207,
+      "step": 1060
+    },
+    {
+      "epoch": 3.77797513321492,
+      "grad_norm": 0.8982229232788086,
+      "learning_rate": 4.677021702213353e-06,
+      "loss": 0.103,
+      "step": 1065
+    },
+    {
+      "epoch": 3.7957371225577266,
+      "grad_norm": 0.9792070984840393,
+      "learning_rate": 4.550067138848213e-06,
+      "loss": 0.1056,
+      "step": 1070
+    },
+    {
+      "epoch": 3.8134991119005326,
+      "grad_norm": 1.0115687847137451,
+      "learning_rate": 4.424550668970803e-06,
+      "loss": 0.098,
+      "step": 1075
+    },
+    {
+      "epoch": 3.8312611012433395,
+      "grad_norm": 0.7825543880462646,
+      "learning_rate": 4.30048956584215e-06,
+      "loss": 0.0953,
+      "step": 1080
+    },
+    {
+      "epoch": 3.8490230905861456,
+      "grad_norm": 0.8458123803138733,
+      "learning_rate": 4.1779009024393555e-06,
+      "loss": 0.0967,
+      "step": 1085
+    },
+    {
+      "epoch": 3.866785079928952,
+      "grad_norm": 0.8293980360031128,
+      "learning_rate": 4.056801549106072e-06,
+      "loss": 0.1042,
+      "step": 1090
+    },
+    {
+      "epoch": 3.8845470692717585,
+      "grad_norm": 0.8098350167274475,
+      "learning_rate": 3.937208171230838e-06,
+      "loss": 0.0948,
+      "step": 1095
+    },
+    {
+      "epoch": 3.9023090586145646,
+      "grad_norm": 0.9925907850265503,
+      "learning_rate": 3.819137226953651e-06,
+      "loss": 0.0996,
+      "step": 1100
+    },
+    {
+      "epoch": 3.9200710479573715,
+      "grad_norm": 0.8489020466804504,
+      "learning_rate": 3.702604964901011e-06,
+      "loss": 0.1068,
+      "step": 1105
+    },
+    {
+      "epoch": 3.9378330373001775,
+      "grad_norm": 0.8919208645820618,
+      "learning_rate": 3.5876274219498462e-06,
+      "loss": 0.1048,
+      "step": 1110
+    },
+    {
+      "epoch": 3.955595026642984,
+      "grad_norm": 0.9406099915504456,
+      "learning_rate": 3.4742204210205397e-06,
+      "loss": 0.0923,
+      "step": 1115
+    },
+    {
+      "epoch": 3.9733570159857905,
+      "grad_norm": 0.8420392870903015,
+      "learning_rate": 3.362399568899447e-06,
+      "loss": 0.103,
+      "step": 1120
+    },
+    {
+      "epoch": 3.9911190053285965,
+      "grad_norm": 0.8659491539001465,
+      "learning_rate": 3.2521802540910998e-06,
+      "loss": 0.1165,
+      "step": 1125
+    },
+    {
+      "epoch": 4.007104795737122,
+      "grad_norm": 0.7806524038314819,
+      "learning_rate": 3.143577644700494e-06,
+      "loss": 0.0893,
+      "step": 1130
+    },
+    {
+      "epoch": 4.024866785079929,
+      "grad_norm": 0.7519360184669495,
+      "learning_rate": 3.0366066863456917e-06,
+      "loss": 0.0853,
+      "step": 1135
+    },
+    {
+      "epoch": 4.042628774422735,
+      "grad_norm": 0.8715997934341431,
+      "learning_rate": 2.9312821001010443e-06,
+      "loss": 0.0739,
+      "step": 1140
+    },
+    {
+      "epoch": 4.060390763765541,
+      "grad_norm": 0.820397675037384,
+      "learning_rate": 2.8276183804713074e-06,
+      "loss": 0.0878,
+      "step": 1145
+    },
+    {
+      "epoch": 4.078152753108348,
+      "grad_norm": 0.8093093633651733,
+      "learning_rate": 2.7256297933969533e-06,
+      "loss": 0.0839,
+      "step": 1150
+    },
+    {
+      "epoch": 4.095914742451154,
+      "grad_norm": 0.7562943696975708,
+      "learning_rate": 2.6253303742909134e-06,
+      "loss": 0.0782,
+      "step": 1155
+    },
+    {
+      "epoch": 4.113676731793961,
+      "grad_norm": 0.8195429444313049,
+      "learning_rate": 2.52673392610709e-06,
+      "loss": 0.0793,
+      "step": 1160
+    },
+    {
+      "epoch": 4.131438721136767,
+      "grad_norm": 0.8424890041351318,
+      "learning_rate": 2.4298540174407915e-06,
+      "loss": 0.0718,
+      "step": 1165
+    },
+    {
+      "epoch": 4.149200710479573,
+      "grad_norm": 0.7583281993865967,
+      "learning_rate": 2.3347039806614763e-06,
+      "loss": 0.0716,
+      "step": 1170
+    },
+    {
+      "epoch": 4.16696269982238,
+      "grad_norm": 0.7109741568565369,
+      "learning_rate": 2.241296910077985e-06,
+      "loss": 0.0815,
+      "step": 1175
+    },
+    {
+      "epoch": 4.184724689165186,
+      "grad_norm": 0.8252658247947693,
+      "learning_rate": 2.1496456601365372e-06,
+      "loss": 0.0793,
+      "step": 1180
+    },
+    {
+      "epoch": 4.202486678507993,
+      "grad_norm": 0.7776868939399719,
+      "learning_rate": 2.0597628436517234e-06,
+      "loss": 0.0889,
+      "step": 1185
+    },
+    {
+      "epoch": 4.220248667850799,
+      "grad_norm": 0.7549480199813843,
+      "learning_rate": 1.9716608300707706e-06,
+      "loss": 0.0716,
+      "step": 1190
+    },
+    {
+      "epoch": 4.238010657193605,
+      "grad_norm": 0.8307995200157166,
+      "learning_rate": 1.8853517437712819e-06,
+      "loss": 0.0825,
+      "step": 1195
+    },
+    {
+      "epoch": 4.255772646536412,
+      "grad_norm": 0.8016964793205261,
+      "learning_rate": 1.8008474623927279e-06,
+      "loss": 0.0797,
+      "step": 1200
+    },
+    {
+      "epoch": 4.273534635879218,
+      "grad_norm": 0.7098313570022583,
+      "learning_rate": 1.718159615201853e-06,
+      "loss": 0.0813,
+      "step": 1205
+    },
+    {
+      "epoch": 4.291296625222025,
+      "grad_norm": 0.8292860388755798,
+      "learning_rate": 1.6372995814922981e-06,
+      "loss": 0.0765,
+      "step": 1210
+    },
+    {
+      "epoch": 4.309058614564831,
+      "grad_norm": 0.8883909583091736,
+      "learning_rate": 1.558278489018606e-06,
+      "loss": 0.0793,
+      "step": 1215
+    },
+    {
+      "epoch": 4.326820603907637,
+      "grad_norm": 0.783700168132782,
+      "learning_rate": 1.4811072124648595e-06,
+      "loss": 0.0757,
+      "step": 1220
+    },
+    {
+      "epoch": 4.344582593250444,
+      "grad_norm": 0.8235071897506714,
+      "learning_rate": 1.4057963719481132e-06,
+      "loss": 0.0819,
+      "step": 1225
+    },
+    {
+      "epoch": 4.36234458259325,
+      "grad_norm": 0.9505485892295837,
+      "learning_rate": 1.3323563315568992e-06,
+      "loss": 0.0747,
+      "step": 1230
+    },
+    {
+      "epoch": 4.380106571936057,
+      "grad_norm": 0.748783528804779,
+      "learning_rate": 1.2607971979249328e-06,
+      "loss": 0.0793,
+      "step": 1235
+    },
+    {
+      "epoch": 4.397868561278863,
+      "grad_norm": 0.7575653791427612,
+      "learning_rate": 1.191128818840282e-06,
+      "loss": 0.0776,
+      "step": 1240
+    },
+    {
+      "epoch": 4.415630550621669,
+      "grad_norm": 0.837280809879303,
+      "learning_rate": 1.1233607818901158e-06,
+      "loss": 0.096,
+      "step": 1245
+    },
+    {
+      "epoch": 4.433392539964476,
+      "grad_norm": 0.7313412427902222,
+      "learning_rate": 1.0575024131413018e-06,
+      "loss": 0.0769,
+      "step": 1250
+    },
+    {
+      "epoch": 4.451154529307282,
+      "grad_norm": 0.9059879183769226,
+      "learning_rate": 9.935627758569683e-07,
+      "loss": 0.0911,
+      "step": 1255
+    },
+    {
+      "epoch": 4.468916518650089,
+      "grad_norm": 0.7813154458999634,
+      "learning_rate": 9.315506692492471e-07,
+      "loss": 0.0803,
+      "step": 1260
+    },
+    {
+      "epoch": 4.486678507992895,
+      "grad_norm": 0.9524432420730591,
+      "learning_rate": 8.714746272683416e-07,
+      "loss": 0.0807,
+      "step": 1265
+    },
+    {
+      "epoch": 4.504440497335701,
+      "grad_norm": 0.7138175964355469,
+      "learning_rate": 8.133429174281087e-07,
+      "loss": 0.081,
+      "step": 1270
+    },
+    {
+      "epoch": 4.522202486678508,
+      "grad_norm": 0.7811490893363953,
+      "learning_rate": 7.571635396683057e-07,
+      "loss": 0.0878,
+      "step": 1275
+    },
+    {
+      "epoch": 4.539964476021314,
+      "grad_norm": 0.6768120527267456,
+      "learning_rate": 7.029442252536578e-07,
+      "loss": 0.0787,
+      "step": 1280
+    },
+    {
+      "epoch": 4.557726465364121,
+      "grad_norm": 0.7318595051765442,
+      "learning_rate": 6.506924357098987e-07,
+      "loss": 0.0771,
+      "step": 1285
+    },
+    {
+      "epoch": 4.575488454706927,
+      "grad_norm": 0.866117537021637,
+      "learning_rate": 6.004153617969393e-07,
+      "loss": 0.0747,
+      "step": 1290
+    },
+    {
+      "epoch": 4.593250444049733,
+      "grad_norm": 0.7473618388175964,
+      "learning_rate": 5.521199225192842e-07,
+      "loss": 0.0803,
+      "step": 1295
+    },
+    {
+      "epoch": 4.61101243339254,
+      "grad_norm": 0.7743943333625793,
+      "learning_rate": 5.058127641738658e-07,
+      "loss": 0.0755,
+      "step": 1300
+    },
+    {
+      "epoch": 4.628774422735346,
+      "grad_norm": 0.714055597782135,
+      "learning_rate": 4.61500259435394e-07,
+      "loss": 0.0802,
+      "step": 1305
+    },
+    {
+      "epoch": 4.646536412078152,
+      "grad_norm": 0.7665424942970276,
+      "learning_rate": 4.191885064793649e-07,
+      "loss": 0.0761,
+      "step": 1310
+    },
+    {
+      "epoch": 4.664298401420959,
+      "grad_norm": 0.7264588475227356,
+      "learning_rate": 3.788833281428394e-07,
+      "loss": 0.0838,
+      "step": 1315
+    },
+    {
+      "epoch": 4.682060390763765,
+      "grad_norm": 0.7086548805236816,
+      "learning_rate": 3.405902711231257e-07,
+      "loss": 0.0713,
+      "step": 1320
+    },
+    {
+      "epoch": 4.699822380106572,
+      "grad_norm": 0.7629944086074829,
+      "learning_rate": 3.0431460521446295e-07,
+      "loss": 0.0857,
+      "step": 1325
+    },
+    {
+      "epoch": 4.717584369449378,
+      "grad_norm": 0.7875418066978455,
+      "learning_rate": 2.7006132258278525e-07,
+      "loss": 0.0726,
+      "step": 1330
+    },
+    {
+      "epoch": 4.735346358792185,
+      "grad_norm": 0.7411353588104248,
+      "learning_rate": 2.3783513707873405e-07,
+      "loss": 0.0838,
+      "step": 1335
+    },
+    {
+      "epoch": 4.753108348134991,
+      "grad_norm": 0.7165302038192749,
+      "learning_rate": 2.076404835889356e-07,
+      "loss": 0.0751,
+      "step": 1340
+    },
+    {
+      "epoch": 4.770870337477797,
+      "grad_norm": 0.7152184247970581,
+      "learning_rate": 1.7948151742569085e-07,
+      "loss": 0.0702,
+      "step": 1345
+    },
+    {
+      "epoch": 4.788632326820604,
+      "grad_norm": 0.6724897623062134,
+      "learning_rate": 1.5336211375512776e-07,
+      "loss": 0.08,
+      "step": 1350
+    },
+    {
+      "epoch": 4.80639431616341,
+      "grad_norm": 0.7092737555503845,
+      "learning_rate": 1.2928586706390965e-07,
+      "loss": 0.0718,
+      "step": 1355
+    },
+    {
+      "epoch": 4.824156305506216,
+      "grad_norm": 0.7389643788337708,
+      "learning_rate": 1.072560906645742e-07,
+      "loss": 0.0765,
+      "step": 1360
+    },
+    {
+      "epoch": 4.841918294849023,
+      "grad_norm": 0.7501643896102905,
+      "learning_rate": 8.727581623956816e-08,
+      "loss": 0.0895,
+      "step": 1365
+    },
+    {
+      "epoch": 4.859680284191829,
+      "grad_norm": 0.7356906533241272,
+      "learning_rate": 6.934779342402109e-08,
+      "loss": 0.0765,
+      "step": 1370
+    },
+    {
+      "epoch": 4.877442273534636,
+      "grad_norm": 0.6943116784095764,
+      "learning_rate": 5.347448942735977e-08,
+      "loss": 0.0728,
+      "step": 1375
+    },
+    {
+      "epoch": 4.895204262877442,
+      "grad_norm": 0.7151592373847961,
+      "learning_rate": 3.9658088693773074e-08,
+      "loss": 0.0769,
+      "step": 1380
+    },
+    {
+      "epoch": 4.912966252220249,
+      "grad_norm": 0.6891736388206482,
+      "learning_rate": 2.7900492601592442e-08,
+      "loss": 0.0804,
+      "step": 1385
+    },
+    {
+      "epoch": 4.930728241563055,
+      "grad_norm": 0.7786058187484741,
+      "learning_rate": 1.8203319201632874e-08,
+      "loss": 0.073,
+      "step": 1390
+    },
+    {
+      "epoch": 4.948490230905861,
+      "grad_norm": 0.807967483997345,
+      "learning_rate": 1.0567902994517708e-08,
+      "loss": 0.0796,
+      "step": 1395
+    },
+    {
+      "epoch": 4.966252220248668,
+      "grad_norm": 0.7293710708618164,
+      "learning_rate": 4.995294747028866e-09,
+      "loss": 0.0878,
+      "step": 1400
+    },
+    {
+      "epoch": 4.984014209591474,
+      "grad_norm": 0.7234616875648499,
+      "learning_rate": 1.4862613475008502e-09,
+      "loss": 0.0789,
+      "step": 1405
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 0.764647901058197,
+      "learning_rate": 4.128570029182655e-11,
+      "loss": 0.0743,
+      "step": 1410
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1410,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 2000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.0104339613055713e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

25_128_e5_3e-5/checkpoint-1410/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96b7946d1d78b1719e782e6e2e941d5b802d64ac2133b6cbedc8dc90ff75572e
+size 7736

25_128_e5_3e-5/checkpoint-1410/zero_to_fp32.py

@@ -0,0 +1,604 @@
+#!/usr/bin/env python
+
+# Copyright (c) Microsoft Corporation.
+# SPDX-License-Identifier: Apache-2.0
+
+# DeepSpeed Team
+
+# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets
+# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in
+# the future. Once extracted, the weights don't require DeepSpeed and can be used in any
+# application.
+#
+# example: python zero_to_fp32.py . pytorch_model.bin
+
+import argparse
+import torch
+import glob
+import math
+import os
+import re
+from collections import OrderedDict
+from dataclasses import dataclass
+
+# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with
+# DeepSpeed data structures it has to be available in the current python environment.
+from deepspeed.utils import logger
+from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS,
+                                            FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES,
+                                            FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS)
+
+
+@dataclass
+class zero_model_state:
+    buffers: dict()
+    param_shapes: dict()
+    shared_params: list
+    ds_version: int
+    frozen_param_shapes: dict()
+    frozen_param_fragments: dict()
+
+
+debug = 0
+
+# load to cpu
+device = torch.device('cpu')
+
+
+def atoi(text):
+    return int(text) if text.isdigit() else text
+
+
+def natural_keys(text):
+    '''
+    alist.sort(key=natural_keys) sorts in human order
+    http://nedbatchelder.com/blog/200712/human_sorting.html
+    (See Toothy's implementation in the comments)
+    '''
+    return [atoi(c) for c in re.split(r'(\d+)', text)]
+
+
+def get_model_state_file(checkpoint_dir, zero_stage):
+    if not os.path.isdir(checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist")
+
+    # there should be only one file
+    if zero_stage <= 2:
+        file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt")
+    elif zero_stage == 3:
+        file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt")
+
+    if not os.path.exists(file):
+        raise FileNotFoundError(f"can't find model states file at '{file}'")
+
+    return file
+
+
+def get_checkpoint_files(checkpoint_dir, glob_pattern):
+    # XXX: need to test that this simple glob rule works for multi-node setup too
+    ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys)
+
+    if len(ckpt_files) == 0:
+        raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'")
+
+    return ckpt_files
+
+
+def get_optim_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt")
+
+
+def get_model_state_files(checkpoint_dir):
+    return get_checkpoint_files(checkpoint_dir, "*_model_states.pt")
+
+
+def parse_model_states(files):
+    zero_model_states = []
+    for file in files:
+        state_dict = torch.load(file, map_location=device)
+
+        if BUFFER_NAMES not in state_dict:
+            raise ValueError(f"{file} is not a model state checkpoint")
+        buffer_names = state_dict[BUFFER_NAMES]
+        if debug:
+            print("Found buffers:", buffer_names)
+
+        # recover just the buffers while restoring them to fp32 if they were saved in fp16
+        buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names}
+        param_shapes = state_dict[PARAM_SHAPES]
+
+        # collect parameters that are included in param_shapes
+        param_names = []
+        for s in param_shapes:
+            for name in s.keys():
+                param_names.append(name)
+
+        # update with frozen parameters
+        frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None)
+        if frozen_param_shapes is not None:
+            if debug:
+                print(f"Found frozen_param_shapes: {frozen_param_shapes}")
+            param_names += list(frozen_param_shapes.keys())
+
+        # handle shared params
+        shared_params = [[k, v] for k, v in state_dict["shared_params"].items()]
+
+        ds_version = state_dict.get(DS_VERSION, None)
+
+        frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None)
+
+        z_model_state = zero_model_state(buffers=buffers,
+                                         param_shapes=param_shapes,
+                                         shared_params=shared_params,
+                                         ds_version=ds_version,
+                                         frozen_param_shapes=frozen_param_shapes,
+                                         frozen_param_fragments=frozen_param_fragments)
+        zero_model_states.append(z_model_state)
+
+    return zero_model_states
+
+
+def parse_optim_states(files, ds_checkpoint_dir):
+
+    total_files = len(files)
+    state_dicts = []
+    for f in files:
+        state_dict = torch.load(f, map_location=device)
+        # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights
+        # and also handle the case where it was already removed by another helper script
+        state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None)
+        state_dicts.append(state_dict)
+
+    if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]:
+        raise ValueError(f"{files[0]} is not a zero checkpoint")
+    zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE]
+    world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT]
+
+    # For ZeRO-2 each param group can have different partition_count as data parallelism for expert
+    # parameters can be different from data parallelism for non-expert parameters. So we can just
+    # use the max of the partition_count to get the dp world_size.
+
+    if type(world_size) is list:
+        world_size = max(world_size)
+
+    if world_size != total_files:
+        raise ValueError(
+            f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. "
+            "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes."
+        )
+
+    # the groups are named differently in each stage
+    if zero_stage <= 2:
+        fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS
+    elif zero_stage == 3:
+        fp32_groups_key = FP32_FLAT_GROUPS
+    else:
+        raise ValueError(f"unknown zero stage {zero_stage}")
+
+    if zero_stage <= 2:
+        fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))]
+    elif zero_stage == 3:
+        # if there is more than one param group, there will be multiple flattened tensors - one
+        # flattened tensor per group - for simplicity merge them into a single tensor
+        #
+        # XXX: could make the script more memory efficient for when there are multiple groups - it
+        # will require matching the sub-lists of param_shapes for each param group flattened tensor
+
+        fp32_flat_groups = [
+            torch.cat(state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key], 0) for i in range(len(state_dicts))
+        ]
+
+    return zero_stage, world_size, fp32_flat_groups
+
+
+def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters):
+    """
+    Returns fp32 state_dict reconstructed from ds checkpoint
+
+    Args:
+        - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are)
+
+    """
+    print(f"Processing zero checkpoint '{ds_checkpoint_dir}'")
+
+    optim_files = get_optim_files(ds_checkpoint_dir)
+    zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir)
+    print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}")
+
+    model_files = get_model_state_files(ds_checkpoint_dir)
+
+    zero_model_states = parse_model_states(model_files)
+    print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}')
+
+    if zero_stage <= 2:
+        return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+    elif zero_stage == 3:
+        return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                                          exclude_frozen_parameters)
+
+
+def _zero2_merge_frozen_params(state_dict, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+    frozen_param_fragments = zero_model_states[0].frozen_param_fragments
+
+    if debug:
+        num_elem = sum(s.numel() for s in frozen_param_shapes.values())
+        print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in frozen_param_fragments.values()])
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        state_dict[name] = frozen_param_fragments[name]
+
+        if debug:
+            print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _has_callable(obj, fn):
+    attr = getattr(obj, fn, None)
+    return callable(attr)
+
+
+def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+
+    # Reconstruction protocol:
+    #
+    # XXX: document this
+
+    if debug:
+        for i in range(world_size):
+            for j in range(len(fp32_flat_groups[0])):
+                print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}")
+
+    # XXX: memory usage doubles here (zero2)
+    num_param_groups = len(fp32_flat_groups[0])
+    merged_single_partition_of_fp32_groups = []
+    for i in range(num_param_groups):
+        merged_partitions = [sd[i] for sd in fp32_flat_groups]
+        full_single_fp32_vector = torch.cat(merged_partitions, 0)
+        merged_single_partition_of_fp32_groups.append(full_single_fp32_vector)
+    avail_numel = sum(
+        [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups])
+
+    if debug:
+        wanted_params = sum([len(shapes) for shapes in param_shapes])
+        wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes])
+        # not asserting if there is a mismatch due to possible padding
+        print(f"Have {avail_numel} numels to process.")
+        print(f"Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    total_numel = 0
+    total_params = 0
+    for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups):
+        offset = 0
+        avail_numel = full_single_fp32_vector.numel()
+        for name, shape in shapes.items():
+
+            unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape)
+            total_numel += unpartitioned_numel
+            total_params += 1
+
+            if debug:
+                print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ")
+            state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape)
+            offset += unpartitioned_numel
+
+        # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and
+        # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex
+        # paddings performed in the code it's almost impossible to predict the exact numbers w/o the
+        # live optimizer object, so we are checking that the numbers are within the right range
+        align_to = 2 * world_size
+
+        def zero2_align(x):
+            return align_to * math.ceil(x / align_to)
+
+        if debug:
+            print(f"original offset={offset}, avail_numel={avail_numel}")
+
+        offset = zero2_align(offset)
+        avail_numel = zero2_align(avail_numel)
+
+        if debug:
+            print(f"aligned  offset={offset}, avail_numel={avail_numel}")
+
+        # Sanity check
+        if offset != avail_numel:
+            raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero2_merge_frozen_params(state_dict, zero_model_states)
+
+    _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def zero3_partitioned_param_info(unpartitioned_numel, world_size):
+    remainder = unpartitioned_numel % world_size
+    padding_numel = (world_size - remainder) if remainder else 0
+    partitioned_numel = math.ceil(unpartitioned_numel / world_size)
+    return partitioned_numel, padding_numel
+
+
+def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states):
+    if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0:
+        return
+
+    if debug:
+        for i in range(world_size):
+            num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values())
+            print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}')
+
+        frozen_param_shapes = zero_model_states[0].frozen_param_shapes
+        wanted_params = len(frozen_param_shapes)
+        wanted_numel = sum(s.numel() for s in frozen_param_shapes.values())
+        avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size
+        print(f'Frozen params: Have {avail_numel} numels to process.')
+        print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params')
+
+    total_params = 0
+    total_numel = 0
+    for name, shape in zero_model_states[0].frozen_param_shapes.items():
+        total_params += 1
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+
+        param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states)
+        state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape)
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+    print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states):
+    param_shapes = zero_model_states[0].param_shapes
+    avail_numel = fp32_flat_groups[0].numel() * world_size
+    # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each
+    # param, re-consolidating each param, while dealing with padding if any
+
+    # merge list of dicts, preserving order
+    param_shapes = {k: v for d in param_shapes for k, v in d.items()}
+
+    if debug:
+        for i in range(world_size):
+            print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}")
+
+        wanted_params = len(param_shapes)
+        wanted_numel = sum(shape.numel() for shape in param_shapes.values())
+        # not asserting if there is a mismatch due to possible padding
+        avail_numel = fp32_flat_groups[0].numel() * world_size
+        print(f"Trainable params: Have {avail_numel} numels to process.")
+        print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.")
+
+    # params
+    # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support
+    # out-of-core computing solution
+    offset = 0
+    total_numel = 0
+    total_params = 0
+    for name, shape in param_shapes.items():
+
+        unpartitioned_numel = shape.numel()
+        total_numel += unpartitioned_numel
+        total_params += 1
+
+        partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size)
+
+        if debug:
+            print(
+                f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}"
+            )
+
+        # XXX: memory usage doubles here
+        state_dict[name] = torch.cat(
+            tuple(fp32_flat_groups[i].narrow(0, offset, partitioned_numel) for i in range(world_size)),
+            0).narrow(0, 0, unpartitioned_numel).view(shape)
+        offset += partitioned_numel
+
+    offset *= world_size
+
+    # Sanity check
+    if offset != avail_numel:
+        raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong")
+
+    print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements")
+
+
+def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states,
+                                               exclude_frozen_parameters):
+    state_dict = OrderedDict()
+
+    # buffers
+    buffers = zero_model_states[0].buffers
+    state_dict.update(buffers)
+    if debug:
+        print(f"added {len(buffers)} buffers")
+
+    if not exclude_frozen_parameters:
+        _zero3_merge_frozen_params(state_dict, world_size, zero_model_states)
+
+    _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states)
+
+    # recover shared parameters
+    for pair in zero_model_states[0].shared_params:
+        if pair[1] in state_dict:
+            state_dict[pair[0]] = state_dict[pair[1]]
+
+    return state_dict
+
+
+def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with
+    ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example
+    via a model hub.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+
+    Returns:
+        - pytorch ``state_dict``
+
+    Note: this approach may not work if your application doesn't have sufficient free CPU memory and
+    you may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with
+    the checkpoint.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint
+        # do the training and checkpoint saving
+        state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu
+        model = model.cpu() # move to cpu
+        model.load_state_dict(state_dict)
+        # submit to model hub or save the model to share with others
+
+    In this example the ``model`` will no longer be usable in the deepspeed context of the same
+    application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead.
+
+    """
+    if tag is None:
+        latest_path = os.path.join(checkpoint_dir, 'latest')
+        if os.path.isfile(latest_path):
+            with open(latest_path, 'r') as fd:
+                tag = fd.read().strip()
+        else:
+            raise ValueError(f"Unable to find 'latest' file at {latest_path}")
+
+    ds_checkpoint_dir = os.path.join(checkpoint_dir, tag)
+
+    if not os.path.isdir(ds_checkpoint_dir):
+        raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist")
+
+    return _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters)
+
+
+def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, output_file, tag=None, exclude_frozen_parameters=False):
+    """
+    Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be
+    loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed.
+
+    Args:
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``output_file``: path to the pytorch fp32 state_dict output file (e.g. path/pytorch_model.bin)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+        - ``exclude_frozen_parameters``: exclude frozen parameters
+    """
+
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag, exclude_frozen_parameters)
+    print(f"Saving fp32 state dict to {output_file}")
+    torch.save(state_dict, output_file)
+
+
+def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None):
+    """
+    1. Put the provided model to cpu
+    2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict``
+    3. Load it into the provided model
+
+    Args:
+        - ``model``: the model object to update
+        - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``)
+        - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14``
+
+    Returns:
+        - ``model`: modified model
+
+    Make sure you have plenty of CPU memory available before you call this function. If you don't
+    have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it
+    conveniently placed for you in the checkpoint folder.
+
+    A typical usage might be ::
+
+        from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint
+        model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir)
+        # submit to model hub or save the model to share with others
+
+    Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context
+    of the same application. i.e. you will need to re-initialize the deepspeed engine, since
+    ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it.
+
+    """
+    logger.info(f"Extracting fp32 weights")
+    state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag)
+
+    logger.info(f"Overwriting model with fp32 weights")
+    model = model.cpu()
+    model.load_state_dict(state_dict, strict=False)
+
+    return model
+
+
+if __name__ == "__main__":
+
+    parser = argparse.ArgumentParser()
+    parser.add_argument("checkpoint_dir",
+                        type=str,
+                        help="path to the desired checkpoint folder, e.g., path/checkpoint-12")
+    parser.add_argument(
+        "output_file",
+        type=str,
+        help="path to the pytorch fp32 state_dict output file (e.g. path/checkpoint-12/pytorch_model.bin)")
+    parser.add_argument("-t",
+                        "--tag",
+                        type=str,
+                        default=None,
+                        help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1")
+    parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters")
+    parser.add_argument("-d", "--debug", action='store_true', help="enable debug")
+    args = parser.parse_args()
+
+    debug = args.debug
+
+    convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir,
+                                               args.output_file,
+                                               tag=args.tag,
+                                               exclude_frozen_parameters=args.exclude_frozen_parameters)

25_128_e5_3e-5/checkpoint-282/README.md

@@ -0,0 +1,202 @@
+---
+base_model: ibm-granite/granite-3.3-8b-base
+library_name: peft
+---
+
+# Model Card for Model ID
+
+<!-- Provide a quick summary of what the model is/does. -->
+
+
+
+## Model Details
+
+### Model Description
+
+<!-- Provide a longer summary of what this model is. -->
+
+
+
+- **Developed by:** [More Information Needed]
+- **Funded by [optional]:** [More Information Needed]
+- **Shared by [optional]:** [More Information Needed]
+- **Model type:** [More Information Needed]
+- **Language(s) (NLP):** [More Information Needed]
+- **License:** [More Information Needed]
+- **Finetuned from model [optional]:** [More Information Needed]
+
+### Model Sources [optional]
+
+<!-- Provide the basic links for the model. -->
+
+- **Repository:** [More Information Needed]
+- **Paper [optional]:** [More Information Needed]
+- **Demo [optional]:** [More Information Needed]
+
+## Uses
+
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
+
+### Direct Use
+
+<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
+
+[More Information Needed]
+
+### Downstream Use [optional]
+
+<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
+
+[More Information Needed]
+
+### Out-of-Scope Use
+
+<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
+
+[More Information Needed]
+
+## Bias, Risks, and Limitations
+
+<!-- This section is meant to convey both technical and sociotechnical limitations. -->
+
+[More Information Needed]
+
+### Recommendations
+
+<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
+
+Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
+
+## How to Get Started with the Model
+
+Use the code below to get started with the model.
+
+[More Information Needed]
+
+## Training Details
+
+### Training Data
+
+<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
+
+[More Information Needed]
+
+### Training Procedure
+
+<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
+
+#### Preprocessing [optional]
+
+[More Information Needed]
+
+
+#### Training Hyperparameters
+
+- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
+
+#### Speeds, Sizes, Times [optional]
+
+<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
+
+[More Information Needed]
+
+## Evaluation
+
+<!-- This section describes the evaluation protocols and provides the results. -->
+
+### Testing Data, Factors & Metrics
+
+#### Testing Data
+
+<!-- This should link to a Dataset Card if possible. -->
+
+[More Information Needed]
+
+#### Factors
+
+<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
+
+[More Information Needed]
+
+#### Metrics
+
+<!-- These are the evaluation metrics being used, ideally with a description of why. -->
+
+[More Information Needed]
+
+### Results
+
+[More Information Needed]
+
+#### Summary
+
+
+
+## Model Examination [optional]
+
+<!-- Relevant interpretability work for the model goes here -->
+
+[More Information Needed]
+
+## Environmental Impact
+
+<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
+
+Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
+
+- **Hardware Type:** [More Information Needed]
+- **Hours used:** [More Information Needed]
+- **Cloud Provider:** [More Information Needed]
+- **Compute Region:** [More Information Needed]
+- **Carbon Emitted:** [More Information Needed]
+
+## Technical Specifications [optional]
+
+### Model Architecture and Objective
+
+[More Information Needed]
+
+### Compute Infrastructure
+
+[More Information Needed]
+
+#### Hardware
+
+[More Information Needed]
+
+#### Software
+
+[More Information Needed]
+
+## Citation [optional]
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+[More Information Needed]
+
+**APA:**
+
+[More Information Needed]
+
+## Glossary [optional]
+
+<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
+
+[More Information Needed]
+
+## More Information [optional]
+
+[More Information Needed]
+
+## Model Card Authors [optional]
+
+[More Information Needed]
+
+## Model Card Contact
+
+[More Information Needed]
+### Framework versions
+
+- PEFT 0.15.2

25_128_e5_3e-5/checkpoint-282/adapter_config.json

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

25_128_e5_3e-5/checkpoint-282/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35744a46997848b7e576d66177c0e42504e84750815eaf06783aaf23ff3360e5
+size 791751704

25_128_e5_3e-5/checkpoint-282/latest

@@ -0,0 +1 @@
+global_step282

25_128_e5_3e-5/checkpoint-282/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dcc98dddfdec444146fcf08db64fffaacc91df1213a51acae51845f37eed0559
+size 15920

25_128_e5_3e-5/checkpoint-282/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b428f496a51d652dfacd5c80ace74a2b567beb281193923835f8a72b09fe5463
+size 15920

25_128_e5_3e-5/checkpoint-282/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:969d0ec3bee0c15083d4970b1e269ee0bca038c4f7418622afba5c9feb244b0f
+size 15920