Upload folder using huggingface_hub

40_128_e5_3e-5/checkpoint-1062/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

40_128_e5_3e-5/checkpoint-1062/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": [
+    "down_proj",
+    "o_proj",
+    "v_proj",
+    "up_proj",
+    "gate_proj",
+    "k_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

40_128_e5_3e-5/checkpoint-1062/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5bae4240d5b438ceb57501e59cabd7ec6985787073fb519adfa387cdce33004b
+size 791751704

40_128_e5_3e-5/checkpoint-1062/latest

@@ -0,0 +1 @@
+global_step1062

40_128_e5_3e-5/checkpoint-1062/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02d3192872c0cfc4ca4e076129a5976f8104781f63086bc76a068d5894d86bc2
+size 15920

40_128_e5_3e-5/checkpoint-1062/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c8df365b8759445acc5a214f59745b9c36a0f87af28fc29279dc4a6f9bd5537
+size 15920

40_128_e5_3e-5/checkpoint-1062/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1613243690fb54f72e045ca74ad508c35d00adce175a016696f2bab7621b9faa
+size 15920

40_128_e5_3e-5/checkpoint-1062/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9520961f9bbc2b36f9e0665cd016902c17830a513e14f888d270265fa9d595be
+size 15920

40_128_e5_3e-5/checkpoint-1062/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84e735593696b89c4596525d419c51e26a4efa75aa23720f58a7e76703248410
+size 15920

40_128_e5_3e-5/checkpoint-1062/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07d8f6efbff8752c747a29c90b1e95142701485775785ba0a56beba152aae688
+size 15920

40_128_e5_3e-5/checkpoint-1062/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10325c7ddab5e91bc991060103197efa267c131a83ecbe4fcaa29f0358a945c4
+size 15920

40_128_e5_3e-5/checkpoint-1062/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50db5b38d8784219fd97ca597164a3206b50fc685d471b0d1852c813834e747b
+size 15920

40_128_e5_3e-5/checkpoint-1062/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20be17243af5f1b444f317c4bf40ad545fcfcba6e5a5858e42d0f912953304da
+size 1064

40_128_e5_3e-5/checkpoint-1062/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
+  }
+}

40_128_e5_3e-5/checkpoint-1062/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
+}

40_128_e5_3e-5/checkpoint-1062/trainer_state.json

@@ -0,0 +1,1518 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1062,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.014144271570014143,
+      "grad_norm": 1.1091066598892212,
+      "learning_rate": 1.348314606741573e-06,
+      "loss": 1.3703,
+      "step": 5
+    },
+    {
+      "epoch": 0.028288543140028287,
+      "grad_norm": 1.1845866441726685,
+      "learning_rate": 3.033707865168539e-06,
+      "loss": 1.3488,
+      "step": 10
+    },
+    {
+      "epoch": 0.042432814710042434,
+      "grad_norm": 0.6645119190216064,
+      "learning_rate": 4.719101123595506e-06,
+      "loss": 1.3309,
+      "step": 15
+    },
+    {
+      "epoch": 0.056577086280056574,
+      "grad_norm": 0.6867905855178833,
+      "learning_rate": 6.404494382022472e-06,
+      "loss": 1.3852,
+      "step": 20
+    },
+    {
+      "epoch": 0.07072135785007072,
+      "grad_norm": 0.8682529330253601,
+      "learning_rate": 8.089887640449438e-06,
+      "loss": 1.2832,
+      "step": 25
+    },
+    {
+      "epoch": 0.08486562942008487,
+      "grad_norm": 0.5214539766311646,
+      "learning_rate": 9.775280898876405e-06,
+      "loss": 1.351,
+      "step": 30
+    },
+    {
+      "epoch": 0.09900990099009901,
+      "grad_norm": 0.5842329859733582,
+      "learning_rate": 1.146067415730337e-05,
+      "loss": 1.199,
+      "step": 35
+    },
+    {
+      "epoch": 0.11315417256011315,
+      "grad_norm": 0.5440396070480347,
+      "learning_rate": 1.3146067415730338e-05,
+      "loss": 1.2351,
+      "step": 40
+    },
+    {
+      "epoch": 0.1272984441301273,
+      "grad_norm": 0.5273528695106506,
+      "learning_rate": 1.4831460674157303e-05,
+      "loss": 1.2773,
+      "step": 45
+    },
+    {
+      "epoch": 0.14144271570014144,
+      "grad_norm": 0.5239897966384888,
+      "learning_rate": 1.651685393258427e-05,
+      "loss": 1.2557,
+      "step": 50
+    },
+    {
+      "epoch": 0.15558698727015557,
+      "grad_norm": 0.5751009583473206,
+      "learning_rate": 1.8202247191011237e-05,
+      "loss": 1.2738,
+      "step": 55
+    },
+    {
+      "epoch": 0.16973125884016974,
+      "grad_norm": 0.5215653777122498,
+      "learning_rate": 1.98876404494382e-05,
+      "loss": 1.2258,
+      "step": 60
+    },
+    {
+      "epoch": 0.18387553041018387,
+      "grad_norm": 0.5515840649604797,
+      "learning_rate": 2.1573033707865168e-05,
+      "loss": 1.1885,
+      "step": 65
+    },
+    {
+      "epoch": 0.19801980198019803,
+      "grad_norm": 0.5996809601783752,
+      "learning_rate": 2.3258426966292135e-05,
+      "loss": 1.1447,
+      "step": 70
+    },
+    {
+      "epoch": 0.21216407355021216,
+      "grad_norm": 0.6658637523651123,
+      "learning_rate": 2.4943820224719103e-05,
+      "loss": 1.1789,
+      "step": 75
+    },
+    {
+      "epoch": 0.2263083451202263,
+      "grad_norm": 0.5400289297103882,
+      "learning_rate": 2.6629213483146066e-05,
+      "loss": 1.1788,
+      "step": 80
+    },
+    {
+      "epoch": 0.24045261669024046,
+      "grad_norm": 0.6226291060447693,
+      "learning_rate": 2.8314606741573034e-05,
+      "loss": 1.1713,
+      "step": 85
+    },
+    {
+      "epoch": 0.2545968882602546,
+      "grad_norm": 0.6412995457649231,
+      "learning_rate": 3e-05,
+      "loss": 1.1693,
+      "step": 90
+    },
+    {
+      "epoch": 0.26874115983026875,
+      "grad_norm": 0.6276025176048279,
+      "learning_rate": 2.9999345118939752e-05,
+      "loss": 1.1233,
+      "step": 95
+    },
+    {
+      "epoch": 0.2828854314002829,
+      "grad_norm": 0.5257118344306946,
+      "learning_rate": 2.9997380532941557e-05,
+      "loss": 1.0805,
+      "step": 100
+    },
+    {
+      "epoch": 0.297029702970297,
+      "grad_norm": 0.6102027893066406,
+      "learning_rate": 2.999410641354812e-05,
+      "loss": 1.1336,
+      "step": 105
+    },
+    {
+      "epoch": 0.31117397454031115,
+      "grad_norm": 0.6746109127998352,
+      "learning_rate": 2.9989523046647264e-05,
+      "loss": 1.1515,
+      "step": 110
+    },
+    {
+      "epoch": 0.32531824611032534,
+      "grad_norm": 0.6818143129348755,
+      "learning_rate": 2.9983630832447017e-05,
+      "loss": 1.0251,
+      "step": 115
+    },
+    {
+      "epoch": 0.33946251768033947,
+      "grad_norm": 0.6401700973510742,
+      "learning_rate": 2.9976430285440642e-05,
+      "loss": 1.0817,
+      "step": 120
+    },
+    {
+      "epoch": 0.3536067892503536,
+      "grad_norm": 0.6563226580619812,
+      "learning_rate": 2.9967922034361726e-05,
+      "loss": 1.0144,
+      "step": 125
+    },
+    {
+      "epoch": 0.36775106082036774,
+      "grad_norm": 0.7533389925956726,
+      "learning_rate": 2.995810682212926e-05,
+      "loss": 1.0501,
+      "step": 130
+    },
+    {
+      "epoch": 0.38189533239038187,
+      "grad_norm": 0.6909534335136414,
+      "learning_rate": 2.9946985505782792e-05,
+      "loss": 1.0284,
+      "step": 135
+    },
+    {
+      "epoch": 0.39603960396039606,
+      "grad_norm": 0.639701783657074,
+      "learning_rate": 2.993455905640758e-05,
+      "loss": 1.0045,
+      "step": 140
+    },
+    {
+      "epoch": 0.4101838755304102,
+      "grad_norm": 1.8686423301696777,
+      "learning_rate": 2.9920828559049805e-05,
+      "loss": 0.9764,
+      "step": 145
+    },
+    {
+      "epoch": 0.4243281471004243,
+      "grad_norm": 0.7100444436073303,
+      "learning_rate": 2.9905795212621825e-05,
+      "loss": 1.0063,
+      "step": 150
+    },
+    {
+      "epoch": 0.43847241867043846,
+      "grad_norm": 0.8259682059288025,
+      "learning_rate": 2.9889460329797482e-05,
+      "loss": 0.9937,
+      "step": 155
+    },
+    {
+      "epoch": 0.4526166902404526,
+      "grad_norm": 0.6790755391120911,
+      "learning_rate": 2.9871825336897495e-05,
+      "loss": 0.9834,
+      "step": 160
+    },
+    {
+      "epoch": 0.4667609618104668,
+      "grad_norm": 0.7154070734977722,
+      "learning_rate": 2.985289177376491e-05,
+      "loss": 0.9921,
+      "step": 165
+    },
+    {
+      "epoch": 0.4809052333804809,
+      "grad_norm": 0.8498982787132263,
+      "learning_rate": 2.983266129363065e-05,
+      "loss": 0.9274,
+      "step": 170
+    },
+    {
+      "epoch": 0.49504950495049505,
+      "grad_norm": 0.7261009216308594,
+      "learning_rate": 2.9811135662969146e-05,
+      "loss": 0.9445,
+      "step": 175
+    },
+    {
+      "epoch": 0.5091937765205092,
+      "grad_norm": 1.0621588230133057,
+      "learning_rate": 2.978831676134411e-05,
+      "loss": 0.9137,
+      "step": 180
+    },
+    {
+      "epoch": 0.5233380480905233,
+      "grad_norm": 0.9448386430740356,
+      "learning_rate": 2.9764206581244415e-05,
+      "loss": 0.9553,
+      "step": 185
+    },
+    {
+      "epoch": 0.5374823196605375,
+      "grad_norm": 0.8367416858673096,
+      "learning_rate": 2.973880722791009e-05,
+      "loss": 0.8912,
+      "step": 190
+    },
+    {
+      "epoch": 0.5516265912305516,
+      "grad_norm": 0.8499005436897278,
+      "learning_rate": 2.971212091914854e-05,
+      "loss": 0.9632,
+      "step": 195
+    },
+    {
+      "epoch": 0.5657708628005658,
+      "grad_norm": 0.7935090065002441,
+      "learning_rate": 2.968414998514085e-05,
+      "loss": 0.8724,
+      "step": 200
+    },
+    {
+      "epoch": 0.57991513437058,
+      "grad_norm": 0.8054782152175903,
+      "learning_rate": 2.9654896868238334e-05,
+      "loss": 0.8849,
+      "step": 205
+    },
+    {
+      "epoch": 0.594059405940594,
+      "grad_norm": 0.9245705008506775,
+      "learning_rate": 2.96243641227493e-05,
+      "loss": 0.8934,
+      "step": 210
+    },
+    {
+      "epoch": 0.6082036775106082,
+      "grad_norm": 0.851338803768158,
+      "learning_rate": 2.959255441471597e-05,
+      "loss": 0.8839,
+      "step": 215
+    },
+    {
+      "epoch": 0.6223479490806223,
+      "grad_norm": 0.9376785159111023,
+      "learning_rate": 2.9559470521681726e-05,
+      "loss": 0.8592,
+      "step": 220
+    },
+    {
+      "epoch": 0.6364922206506365,
+      "grad_norm": 0.8601529002189636,
+      "learning_rate": 2.9525115332448557e-05,
+      "loss": 0.8879,
+      "step": 225
+    },
+    {
+      "epoch": 0.6506364922206507,
+      "grad_norm": 0.9360470175743103,
+      "learning_rate": 2.948949184682484e-05,
+      "loss": 0.8337,
+      "step": 230
+    },
+    {
+      "epoch": 0.6647807637906648,
+      "grad_norm": 0.9480540156364441,
+      "learning_rate": 2.9452603175363364e-05,
+      "loss": 0.9082,
+      "step": 235
+    },
+    {
+      "epoch": 0.6789250353606789,
+      "grad_norm": 0.9365689158439636,
+      "learning_rate": 2.941445253908978e-05,
+      "loss": 0.8773,
+      "step": 240
+    },
+    {
+      "epoch": 0.693069306930693,
+      "grad_norm": 0.9173905253410339,
+      "learning_rate": 2.9375043269221294e-05,
+      "loss": 0.81,
+      "step": 245
+    },
+    {
+      "epoch": 0.7072135785007072,
+      "grad_norm": 0.9756169319152832,
+      "learning_rate": 2.9334378806875838e-05,
+      "loss": 0.8046,
+      "step": 250
+    },
+    {
+      "epoch": 0.7213578500707214,
+      "grad_norm": 0.9777851104736328,
+      "learning_rate": 2.929246270277157e-05,
+      "loss": 0.8452,
+      "step": 255
+    },
+    {
+      "epoch": 0.7355021216407355,
+      "grad_norm": 0.9113903045654297,
+      "learning_rate": 2.9249298616916856e-05,
+      "loss": 0.8244,
+      "step": 260
+    },
+    {
+      "epoch": 0.7496463932107497,
+      "grad_norm": 0.906170666217804,
+      "learning_rate": 2.9204890318290666e-05,
+      "loss": 0.7866,
+      "step": 265
+    },
+    {
+      "epoch": 0.7637906647807637,
+      "grad_norm": 0.9952336549758911,
+      "learning_rate": 2.915924168451349e-05,
+      "loss": 0.817,
+      "step": 270
+    },
+    {
+      "epoch": 0.7779349363507779,
+      "grad_norm": 1.1563059091567993,
+      "learning_rate": 2.9112356701508756e-05,
+      "loss": 0.8101,
+      "step": 275
+    },
+    {
+      "epoch": 0.7920792079207921,
+      "grad_norm": 1.035338282585144,
+      "learning_rate": 2.9064239463154782e-05,
+      "loss": 0.7379,
+      "step": 280
+    },
+    {
+      "epoch": 0.8062234794908062,
+      "grad_norm": 1.0756317377090454,
+      "learning_rate": 2.9014894170927308e-05,
+      "loss": 0.7709,
+      "step": 285
+    },
+    {
+      "epoch": 0.8203677510608204,
+      "grad_norm": 1.1403799057006836,
+      "learning_rate": 2.8964325133532638e-05,
+      "loss": 0.8136,
+      "step": 290
+    },
+    {
+      "epoch": 0.8345120226308345,
+      "grad_norm": 1.111162543296814,
+      "learning_rate": 2.8912536766531424e-05,
+      "loss": 0.7563,
+      "step": 295
+    },
+    {
+      "epoch": 0.8486562942008486,
+      "grad_norm": 0.9991361498832703,
+      "learning_rate": 2.885953359195308e-05,
+      "loss": 0.74,
+      "step": 300
+    },
+    {
+      "epoch": 0.8628005657708628,
+      "grad_norm": 0.9173119068145752,
+      "learning_rate": 2.8805320237900965e-05,
+      "loss": 0.7599,
+      "step": 305
+    },
+    {
+      "epoch": 0.8769448373408769,
+      "grad_norm": 1.1988530158996582,
+      "learning_rate": 2.874990143814826e-05,
+      "loss": 0.7256,
+      "step": 310
+    },
+    {
+      "epoch": 0.8910891089108911,
+      "grad_norm": 1.0119719505310059,
+      "learning_rate": 2.8693282031724592e-05,
+      "loss": 0.7472,
+      "step": 315
+    },
+    {
+      "epoch": 0.9052333804809052,
+      "grad_norm": 1.0434908866882324,
+      "learning_rate": 2.8635466962493564e-05,
+      "loss": 0.7512,
+      "step": 320
+    },
+    {
+      "epoch": 0.9193776520509194,
+      "grad_norm": 1.0775619745254517,
+      "learning_rate": 2.8576461278721016e-05,
+      "loss": 0.6982,
+      "step": 325
+    },
+    {
+      "epoch": 0.9335219236209336,
+      "grad_norm": 0.9957691431045532,
+      "learning_rate": 2.8516270132634244e-05,
+      "loss": 0.6556,
+      "step": 330
+    },
+    {
+      "epoch": 0.9476661951909476,
+      "grad_norm": 1.0233315229415894,
+      "learning_rate": 2.845489877997213e-05,
+      "loss": 0.7431,
+      "step": 335
+    },
+    {
+      "epoch": 0.9618104667609618,
+      "grad_norm": 0.9704013466835022,
+      "learning_rate": 2.8392352579526205e-05,
+      "loss": 0.7121,
+      "step": 340
+    },
+    {
+      "epoch": 0.9759547383309759,
+      "grad_norm": 1.0681896209716797,
+      "learning_rate": 2.832863699267274e-05,
+      "loss": 0.716,
+      "step": 345
+    },
+    {
+      "epoch": 0.9900990099009901,
+      "grad_norm": 1.088863492012024,
+      "learning_rate": 2.8263757582895888e-05,
+      "loss": 0.699,
+      "step": 350
+    },
+    {
+      "epoch": 1.002828854314003,
+      "grad_norm": 1.075731873512268,
+      "learning_rate": 2.819772001530186e-05,
+      "loss": 0.647,
+      "step": 355
+    },
+    {
+      "epoch": 1.016973125884017,
+      "grad_norm": 1.180709958076477,
+      "learning_rate": 2.8130530056124298e-05,
+      "loss": 0.6268,
+      "step": 360
+    },
+    {
+      "epoch": 1.031117397454031,
+      "grad_norm": 1.0877044200897217,
+      "learning_rate": 2.8062193572220763e-05,
+      "loss": 0.638,
+      "step": 365
+    },
+    {
+      "epoch": 1.0452616690240453,
+      "grad_norm": 1.0629935264587402,
+      "learning_rate": 2.799271653056046e-05,
+      "loss": 0.6209,
+      "step": 370
+    },
+    {
+      "epoch": 1.0594059405940595,
+      "grad_norm": 1.149365782737732,
+      "learning_rate": 2.7922104997703214e-05,
+      "loss": 0.5914,
+      "step": 375
+    },
+    {
+      "epoch": 1.0735502121640736,
+      "grad_norm": 1.2020652294158936,
+      "learning_rate": 2.785036513926976e-05,
+      "loss": 0.6314,
+      "step": 380
+    },
+    {
+      "epoch": 1.0876944837340876,
+      "grad_norm": 1.1636576652526855,
+      "learning_rate": 2.7777503219403374e-05,
+      "loss": 0.6092,
+      "step": 385
+    },
+    {
+      "epoch": 1.1018387553041018,
+      "grad_norm": 1.1622604131698608,
+      "learning_rate": 2.7703525600222896e-05,
+      "loss": 0.5938,
+      "step": 390
+    },
+    {
+      "epoch": 1.115983026874116,
+      "grad_norm": 1.1172856092453003,
+      "learning_rate": 2.7628438741267218e-05,
+      "loss": 0.5954,
+      "step": 395
+    },
+    {
+      "epoch": 1.1301272984441302,
+      "grad_norm": 1.2672319412231445,
+      "learning_rate": 2.7552249198931255e-05,
+      "loss": 0.5924,
+      "step": 400
+    },
+    {
+      "epoch": 1.1442715700141444,
+      "grad_norm": 1.1092612743377686,
+      "learning_rate": 2.7474963625893433e-05,
+      "loss": 0.5816,
+      "step": 405
+    },
+    {
+      "epoch": 1.1584158415841583,
+      "grad_norm": 1.2071410417556763,
+      "learning_rate": 2.7396588770534822e-05,
+      "loss": 0.5221,
+      "step": 410
+    },
+    {
+      "epoch": 1.1725601131541725,
+      "grad_norm": 1.2543182373046875,
+      "learning_rate": 2.7317131476349874e-05,
+      "loss": 0.5432,
+      "step": 415
+    },
+    {
+      "epoch": 1.1867043847241867,
+      "grad_norm": 1.193509578704834,
+      "learning_rate": 2.723659868134886e-05,
+      "loss": 0.6123,
+      "step": 420
+    },
+    {
+      "epoch": 1.200848656294201,
+      "grad_norm": 1.1950852870941162,
+      "learning_rate": 2.715499741745208e-05,
+      "loss": 0.5481,
+      "step": 425
+    },
+    {
+      "epoch": 1.214992927864215,
+      "grad_norm": 1.1583995819091797,
+      "learning_rate": 2.707233480987582e-05,
+      "loss": 0.5594,
+      "step": 430
+    },
+    {
+      "epoch": 1.229137199434229,
+      "grad_norm": 1.2848454713821411,
+      "learning_rate": 2.6988618076510227e-05,
+      "loss": 0.5634,
+      "step": 435
+    },
+    {
+      "epoch": 1.2432814710042432,
+      "grad_norm": 1.0936193466186523,
+      "learning_rate": 2.6903854527289047e-05,
+      "loss": 0.5672,
+      "step": 440
+    },
+    {
+      "epoch": 1.2574257425742574,
+      "grad_norm": 1.2483489513397217,
+      "learning_rate": 2.6818051563551355e-05,
+      "loss": 0.5401,
+      "step": 445
+    },
+    {
+      "epoch": 1.2715700141442716,
+      "grad_norm": 1.2003945112228394,
+      "learning_rate": 2.6731216677395253e-05,
+      "loss": 0.5686,
+      "step": 450
+    },
+    {
+      "epoch": 1.2857142857142856,
+      "grad_norm": 1.4688889980316162,
+      "learning_rate": 2.6643357451023725e-05,
+      "loss": 0.5277,
+      "step": 455
+    },
+    {
+      "epoch": 1.2998585572842998,
+      "grad_norm": 1.1361063718795776,
+      "learning_rate": 2.6554481556082543e-05,
+      "loss": 0.5116,
+      "step": 460
+    },
+    {
+      "epoch": 1.314002828854314,
+      "grad_norm": 1.3163748979568481,
+      "learning_rate": 2.646459675299042e-05,
+      "loss": 0.5612,
+      "step": 465
+    },
+    {
+      "epoch": 1.3281471004243282,
+      "grad_norm": 1.1666852235794067,
+      "learning_rate": 2.637371089026137e-05,
+      "loss": 0.5716,
+      "step": 470
+    },
+    {
+      "epoch": 1.3422913719943423,
+      "grad_norm": 1.1531682014465332,
+      "learning_rate": 2.6281831903819413e-05,
+      "loss": 0.5386,
+      "step": 475
+    },
+    {
+      "epoch": 1.3564356435643563,
+      "grad_norm": 1.0326910018920898,
+      "learning_rate": 2.6188967816305624e-05,
+      "loss": 0.5494,
+      "step": 480
+    },
+    {
+      "epoch": 1.3705799151343705,
+      "grad_norm": 1.110970377922058,
+      "learning_rate": 2.6095126736377618e-05,
+      "loss": 0.487,
+      "step": 485
+    },
+    {
+      "epoch": 1.3847241867043847,
+      "grad_norm": 1.253021001815796,
+      "learning_rate": 2.6000316858001512e-05,
+      "loss": 0.5005,
+      "step": 490
+    },
+    {
+      "epoch": 1.3988684582743989,
+      "grad_norm": 1.1914869546890259,
+      "learning_rate": 2.5904546459736468e-05,
+      "loss": 0.4999,
+      "step": 495
+    },
+    {
+      "epoch": 1.413012729844413,
+      "grad_norm": 1.1897658109664917,
+      "learning_rate": 2.5807823904011803e-05,
+      "loss": 0.518,
+      "step": 500
+    },
+    {
+      "epoch": 1.427157001414427,
+      "grad_norm": 1.1284512281417847,
+      "learning_rate": 2.5710157636396845e-05,
+      "loss": 0.4999,
+      "step": 505
+    },
+    {
+      "epoch": 1.4413012729844414,
+      "grad_norm": 1.2429921627044678,
+      "learning_rate": 2.5611556184863437e-05,
+      "loss": 0.5009,
+      "step": 510
+    },
+    {
+      "epoch": 1.4554455445544554,
+      "grad_norm": 1.276069164276123,
+      "learning_rate": 2.5512028159041328e-05,
+      "loss": 0.4511,
+      "step": 515
+    },
+    {
+      "epoch": 1.4695898161244696,
+      "grad_norm": 1.2455278635025024,
+      "learning_rate": 2.54115822494664e-05,
+      "loss": 0.483,
+      "step": 520
+    },
+    {
+      "epoch": 1.4837340876944838,
+      "grad_norm": 1.2925827503204346,
+      "learning_rate": 2.5310227226821814e-05,
+      "loss": 0.4746,
+      "step": 525
+    },
+    {
+      "epoch": 1.4978783592644977,
+      "grad_norm": 1.246673345565796,
+      "learning_rate": 2.5207971941172193e-05,
+      "loss": 0.4656,
+      "step": 530
+    },
+    {
+      "epoch": 1.5120226308345122,
+      "grad_norm": 1.27567720413208,
+      "learning_rate": 2.510482532119087e-05,
+      "loss": 0.4722,
+      "step": 535
+    },
+    {
+      "epoch": 1.5261669024045261,
+      "grad_norm": 1.2516908645629883,
+      "learning_rate": 2.500079637338021e-05,
+      "loss": 0.4791,
+      "step": 540
+    },
+    {
+      "epoch": 1.5403111739745403,
+      "grad_norm": 1.4572430849075317,
+      "learning_rate": 2.489589418128524e-05,
+      "loss": 0.4871,
+      "step": 545
+    },
+    {
+      "epoch": 1.5544554455445545,
+      "grad_norm": 1.2278729677200317,
+      "learning_rate": 2.479012790470046e-05,
+      "loss": 0.4845,
+      "step": 550
+    },
+    {
+      "epoch": 1.5685997171145685,
+      "grad_norm": 1.2259896993637085,
+      "learning_rate": 2.468350677887005e-05,
+      "loss": 0.4505,
+      "step": 555
+    },
+    {
+      "epoch": 1.5827439886845829,
+      "grad_norm": 1.0976181030273438,
+      "learning_rate": 2.457604011368147e-05,
+      "loss": 0.439,
+      "step": 560
+    },
+    {
+      "epoch": 1.5968882602545968,
+      "grad_norm": 1.3767074346542358,
+      "learning_rate": 2.4467737292852542e-05,
+      "loss": 0.4443,
+      "step": 565
+    },
+    {
+      "epoch": 1.611032531824611,
+      "grad_norm": 1.2550078630447388,
+      "learning_rate": 2.4358607773112074e-05,
+      "loss": 0.4749,
+      "step": 570
+    },
+    {
+      "epoch": 1.6251768033946252,
+      "grad_norm": 1.2857836484909058,
+      "learning_rate": 2.424866108337415e-05,
+      "loss": 0.4499,
+      "step": 575
+    },
+    {
+      "epoch": 1.6393210749646392,
+      "grad_norm": 1.2040728330612183,
+      "learning_rate": 2.413790682390607e-05,
+      "loss": 0.454,
+      "step": 580
+    },
+    {
+      "epoch": 1.6534653465346536,
+      "grad_norm": 1.22360360622406,
+      "learning_rate": 2.402635466549008e-05,
+      "loss": 0.4122,
+      "step": 585
+    },
+    {
+      "epoch": 1.6676096181046676,
+      "grad_norm": 1.181302785873413,
+      "learning_rate": 2.3914014348578957e-05,
+      "loss": 0.4265,
+      "step": 590
+    },
+    {
+      "epoch": 1.6817538896746818,
+      "grad_norm": 1.2027560472488403,
+      "learning_rate": 2.3800895682445475e-05,
+      "loss": 0.4536,
+      "step": 595
+    },
+    {
+      "epoch": 1.695898161244696,
+      "grad_norm": 1.2701138257980347,
+      "learning_rate": 2.3687008544325904e-05,
+      "loss": 0.4209,
+      "step": 600
+    },
+    {
+      "epoch": 1.71004243281471,
+      "grad_norm": 1.1245002746582031,
+      "learning_rate": 2.3572362878557546e-05,
+      "loss": 0.4382,
+      "step": 605
+    },
+    {
+      "epoch": 1.7241867043847243,
+      "grad_norm": 1.1181449890136719,
+      "learning_rate": 2.3456968695710415e-05,
+      "loss": 0.4125,
+      "step": 610
+    },
+    {
+      "epoch": 1.7383309759547383,
+      "grad_norm": 1.2436251640319824,
+      "learning_rate": 2.3340836071713164e-05,
+      "loss": 0.3983,
+      "step": 615
+    },
+    {
+      "epoch": 1.7524752475247525,
+      "grad_norm": 1.3499740362167358,
+      "learning_rate": 2.322397514697324e-05,
+      "loss": 0.4585,
+      "step": 620
+    },
+    {
+      "epoch": 1.7666195190947667,
+      "grad_norm": 1.1281145811080933,
+      "learning_rate": 2.310639612549149e-05,
+      "loss": 0.377,
+      "step": 625
+    },
+    {
+      "epoch": 1.7807637906647806,
+      "grad_norm": 1.2632691860198975,
+      "learning_rate": 2.298810927397114e-05,
+      "loss": 0.4788,
+      "step": 630
+    },
+    {
+      "epoch": 1.794908062234795,
+      "grad_norm": 1.3099757432937622,
+      "learning_rate": 2.286912492092136e-05,
+      "loss": 0.3682,
+      "step": 635
+    },
+    {
+      "epoch": 1.809052333804809,
+      "grad_norm": 1.3557870388031006,
+      "learning_rate": 2.2749453455755392e-05,
+      "loss": 0.3754,
+      "step": 640
+    },
+    {
+      "epoch": 1.8231966053748232,
+      "grad_norm": 1.1247097253799438,
+      "learning_rate": 2.2629105327883354e-05,
+      "loss": 0.3961,
+      "step": 645
+    },
+    {
+      "epoch": 1.8373408769448374,
+      "grad_norm": 1.1884528398513794,
+      "learning_rate": 2.250809104579987e-05,
+      "loss": 0.3828,
+      "step": 650
+    },
+    {
+      "epoch": 1.8514851485148514,
+      "grad_norm": 1.119057297706604,
+      "learning_rate": 2.2386421176166445e-05,
+      "loss": 0.404,
+      "step": 655
+    },
+    {
+      "epoch": 1.8656294200848658,
+      "grad_norm": 1.1483960151672363,
+      "learning_rate": 2.2264106342888853e-05,
+      "loss": 0.3872,
+      "step": 660
+    },
+    {
+      "epoch": 1.8797736916548797,
+      "grad_norm": 1.416157603263855,
+      "learning_rate": 2.2141157226189442e-05,
+      "loss": 0.3667,
+      "step": 665
+    },
+    {
+      "epoch": 1.893917963224894,
+      "grad_norm": 1.1745518445968628,
+      "learning_rate": 2.2017584561674605e-05,
+      "loss": 0.4214,
+      "step": 670
+    },
+    {
+      "epoch": 1.908062234794908,
+      "grad_norm": 1.1854530572891235,
+      "learning_rate": 2.1893399139397347e-05,
+      "loss": 0.411,
+      "step": 675
+    },
+    {
+      "epoch": 1.922206506364922,
+      "grad_norm": 1.1574931144714355,
+      "learning_rate": 2.176861180291514e-05,
+      "loss": 0.3547,
+      "step": 680
+    },
+    {
+      "epoch": 1.9363507779349365,
+      "grad_norm": 1.4279593229293823,
+      "learning_rate": 2.1643233448343077e-05,
+      "loss": 0.4094,
+      "step": 685
+    },
+    {
+      "epoch": 1.9504950495049505,
+      "grad_norm": 1.2585395574569702,
+      "learning_rate": 2.1517275023402463e-05,
+      "loss": 0.3927,
+      "step": 690
+    },
+    {
+      "epoch": 1.9646393210749646,
+      "grad_norm": 1.2487115859985352,
+      "learning_rate": 2.1390747526464885e-05,
+      "loss": 0.3876,
+      "step": 695
+    },
+    {
+      "epoch": 1.9787835926449788,
+      "grad_norm": 1.1419602632522583,
+      "learning_rate": 2.126366200559184e-05,
+      "loss": 0.3809,
+      "step": 700
+    },
+    {
+      "epoch": 1.9929278642149928,
+      "grad_norm": 1.240652084350586,
+      "learning_rate": 2.11360295575701e-05,
+      "loss": 0.3525,
+      "step": 705
+    },
+    {
+      "epoch": 2.005657708628006,
+      "grad_norm": 1.1952710151672363,
+      "learning_rate": 2.100786132694271e-05,
+      "loss": 0.3126,
+      "step": 710
+    },
+    {
+      "epoch": 2.01980198019802,
+      "grad_norm": 1.3939076662063599,
+      "learning_rate": 2.0879168505035898e-05,
+      "loss": 0.2736,
+      "step": 715
+    },
+    {
+      "epoch": 2.033946251768034,
+      "grad_norm": 1.285657525062561,
+      "learning_rate": 2.074996232898189e-05,
+      "loss": 0.3355,
+      "step": 720
+    },
+    {
+      "epoch": 2.048090523338048,
+      "grad_norm": 1.210814356803894,
+      "learning_rate": 2.0620254080737697e-05,
+      "loss": 0.3003,
+      "step": 725
+    },
+    {
+      "epoch": 2.062234794908062,
+      "grad_norm": 1.160455346107483,
+      "learning_rate": 2.0490055086100004e-05,
+      "loss": 0.2567,
+      "step": 730
+    },
+    {
+      "epoch": 2.0763790664780766,
+      "grad_norm": 1.3446382284164429,
+      "learning_rate": 2.0359376713716224e-05,
+      "loss": 0.2943,
+      "step": 735
+    },
+    {
+      "epoch": 2.0905233380480905,
+      "grad_norm": 1.1885087490081787,
+      "learning_rate": 2.022823037409184e-05,
+      "loss": 0.277,
+      "step": 740
+    },
+    {
+      "epoch": 2.1046676096181045,
+      "grad_norm": 1.2343266010284424,
+      "learning_rate": 2.0096627518594035e-05,
+      "loss": 0.2947,
+      "step": 745
+    },
+    {
+      "epoch": 2.118811881188119,
+      "grad_norm": 1.2885913848876953,
+      "learning_rate": 1.9964579638451827e-05,
+      "loss": 0.2829,
+      "step": 750
+    },
+    {
+      "epoch": 2.132956152758133,
+      "grad_norm": 1.389346718788147,
+      "learning_rate": 1.9832098263752637e-05,
+      "loss": 0.3234,
+      "step": 755
+    },
+    {
+      "epoch": 2.1471004243281473,
+      "grad_norm": 1.2736543416976929,
+      "learning_rate": 1.9699194962435555e-05,
+      "loss": 0.2745,
+      "step": 760
+    },
+    {
+      "epoch": 2.1612446958981613,
+      "grad_norm": 1.1569143533706665,
+      "learning_rate": 1.9565881339281232e-05,
+      "loss": 0.2755,
+      "step": 765
+    },
+    {
+      "epoch": 2.1753889674681752,
+      "grad_norm": 1.2544715404510498,
+      "learning_rate": 1.9432169034898583e-05,
+      "loss": 0.274,
+      "step": 770
+    },
+    {
+      "epoch": 2.1895332390381896,
+      "grad_norm": 1.2836891412734985,
+      "learning_rate": 1.9298069724708365e-05,
+      "loss": 0.2877,
+      "step": 775
+    },
+    {
+      "epoch": 2.2036775106082036,
+      "grad_norm": 1.2649877071380615,
+      "learning_rate": 1.9163595117923697e-05,
+      "loss": 0.2787,
+      "step": 780
+    },
+    {
+      "epoch": 2.217821782178218,
+      "grad_norm": 1.4977940320968628,
+      "learning_rate": 1.9028756956527656e-05,
+      "loss": 0.277,
+      "step": 785
+    },
+    {
+      "epoch": 2.231966053748232,
+      "grad_norm": 1.2288274765014648,
+      "learning_rate": 1.8893567014247997e-05,
+      "loss": 0.3017,
+      "step": 790
+    },
+    {
+      "epoch": 2.246110325318246,
+      "grad_norm": 1.256234884262085,
+      "learning_rate": 1.8758037095529074e-05,
+      "loss": 0.2797,
+      "step": 795
+    },
+    {
+      "epoch": 2.2602545968882604,
+      "grad_norm": 1.199599027633667,
+      "learning_rate": 1.8622179034501145e-05,
+      "loss": 0.2632,
+      "step": 800
+    },
+    {
+      "epoch": 2.2743988684582743,
+      "grad_norm": 1.0664340257644653,
+      "learning_rate": 1.848600469394701e-05,
+      "loss": 0.2737,
+      "step": 805
+    },
+    {
+      "epoch": 2.2885431400282887,
+      "grad_norm": 1.302608609199524,
+      "learning_rate": 1.8349525964266206e-05,
+      "loss": 0.2544,
+      "step": 810
+    },
+    {
+      "epoch": 2.3026874115983027,
+      "grad_norm": 1.4746507406234741,
+      "learning_rate": 1.821275476243676e-05,
+      "loss": 0.2653,
+      "step": 815
+    },
+    {
+      "epoch": 2.3168316831683167,
+      "grad_norm": 1.360602855682373,
+      "learning_rate": 1.8075703030974628e-05,
+      "loss": 0.2759,
+      "step": 820
+    },
+    {
+      "epoch": 2.330975954738331,
+      "grad_norm": 1.1771636009216309,
+      "learning_rate": 1.7938382736890903e-05,
+      "loss": 0.2517,
+      "step": 825
+    },
+    {
+      "epoch": 2.345120226308345,
+      "grad_norm": 1.1717984676361084,
+      "learning_rate": 1.7800805870646895e-05,
+      "loss": 0.2662,
+      "step": 830
+    },
+    {
+      "epoch": 2.3592644978783595,
+      "grad_norm": 1.2912698984146118,
+      "learning_rate": 1.7662984445107133e-05,
+      "loss": 0.2536,
+      "step": 835
+    },
+    {
+      "epoch": 2.3734087694483734,
+      "grad_norm": 1.3276891708374023,
+      "learning_rate": 1.752493049449046e-05,
+      "loss": 0.259,
+      "step": 840
+    },
+    {
+      "epoch": 2.3875530410183874,
+      "grad_norm": 1.1467636823654175,
+      "learning_rate": 1.7386656073319216e-05,
+      "loss": 0.2556,
+      "step": 845
+    },
+    {
+      "epoch": 2.401697312588402,
+      "grad_norm": 1.110716700553894,
+      "learning_rate": 1.7248173255366672e-05,
+      "loss": 0.2669,
+      "step": 850
+    },
+    {
+      "epoch": 2.4158415841584158,
+      "grad_norm": 1.1397027969360352,
+      "learning_rate": 1.7109494132602784e-05,
+      "loss": 0.2524,
+      "step": 855
+    },
+    {
+      "epoch": 2.42998585572843,
+      "grad_norm": 1.236080527305603,
+      "learning_rate": 1.6970630814138337e-05,
+      "loss": 0.2581,
+      "step": 860
+    },
+    {
+      "epoch": 2.444130127298444,
+      "grad_norm": 1.2145838737487793,
+      "learning_rate": 1.6831595425167635e-05,
+      "loss": 0.2544,
+      "step": 865
+    },
+    {
+      "epoch": 2.458274398868458,
+      "grad_norm": 1.1833267211914062,
+      "learning_rate": 1.6692400105909737e-05,
+      "loss": 0.2658,
+      "step": 870
+    },
+    {
+      "epoch": 2.4724186704384725,
+      "grad_norm": 1.1189638376235962,
+      "learning_rate": 1.6553057010548408e-05,
+      "loss": 0.2798,
+      "step": 875
+    },
+    {
+      "epoch": 2.4865629420084865,
+      "grad_norm": 1.1513794660568237,
+      "learning_rate": 1.641357830617085e-05,
+      "loss": 0.2461,
+      "step": 880
+    },
+    {
+      "epoch": 2.500707213578501,
+      "grad_norm": 1.157867431640625,
+      "learning_rate": 1.6273976171705313e-05,
+      "loss": 0.2453,
+      "step": 885
+    },
+    {
+      "epoch": 2.514851485148515,
+      "grad_norm": 1.2551815509796143,
+      "learning_rate": 1.6134262796857628e-05,
+      "loss": 0.2473,
+      "step": 890
+    },
+    {
+      "epoch": 2.528995756718529,
+      "grad_norm": 1.3023325204849243,
+      "learning_rate": 1.599445038104688e-05,
+      "loss": 0.2495,
+      "step": 895
+    },
+    {
+      "epoch": 2.5431400282885432,
+      "grad_norm": 1.4266198873519897,
+      "learning_rate": 1.5854551132340145e-05,
+      "loss": 0.2507,
+      "step": 900
+    },
+    {
+      "epoch": 2.557284299858557,
+      "grad_norm": 1.084045171737671,
+      "learning_rate": 1.5714577266386526e-05,
+      "loss": 0.2485,
+      "step": 905
+    },
+    {
+      "epoch": 2.571428571428571,
+      "grad_norm": 1.359621524810791,
+      "learning_rate": 1.557454100535053e-05,
+      "loss": 0.2291,
+      "step": 910
+    },
+    {
+      "epoch": 2.5855728429985856,
+      "grad_norm": 1.2676442861557007,
+      "learning_rate": 1.543445457684484e-05,
+      "loss": 0.2468,
+      "step": 915
+    },
+    {
+      "epoch": 2.5997171145685996,
+      "grad_norm": 1.1858233213424683,
+      "learning_rate": 1.5294330212862624e-05,
+      "loss": 0.2343,
+      "step": 920
+    },
+    {
+      "epoch": 2.613861386138614,
+      "grad_norm": 1.1932803392410278,
+      "learning_rate": 1.5154180148709494e-05,
+      "loss": 0.2596,
+      "step": 925
+    },
+    {
+      "epoch": 2.628005657708628,
+      "grad_norm": 1.235780954360962,
+      "learning_rate": 1.5014016621935127e-05,
+      "loss": 0.2349,
+      "step": 930
+    },
+    {
+      "epoch": 2.6421499292786423,
+      "grad_norm": 1.4623289108276367,
+      "learning_rate": 1.4873851871264731e-05,
+      "loss": 0.2528,
+      "step": 935
+    },
+    {
+      "epoch": 2.6562942008486563,
+      "grad_norm": 1.3628381490707397,
+      "learning_rate": 1.4733698135530376e-05,
+      "loss": 0.2397,
+      "step": 940
+    },
+    {
+      "epoch": 2.6704384724186703,
+      "grad_norm": 1.3325926065444946,
+      "learning_rate": 1.4593567652602331e-05,
+      "loss": 0.2267,
+      "step": 945
+    },
+    {
+      "epoch": 2.6845827439886847,
+      "grad_norm": 1.1442891359329224,
+      "learning_rate": 1.44534726583205e-05,
+      "loss": 0.2417,
+      "step": 950
+    },
+    {
+      "epoch": 2.6987270155586986,
+      "grad_norm": 1.4495691061019897,
+      "learning_rate": 1.4313425385426003e-05,
+      "loss": 0.2246,
+      "step": 955
+    },
+    {
+      "epoch": 2.7128712871287126,
+      "grad_norm": 1.4209120273590088,
+      "learning_rate": 1.417343806249304e-05,
+      "loss": 0.2084,
+      "step": 960
+    },
+    {
+      "epoch": 2.727015558698727,
+      "grad_norm": 1.1167337894439697,
+      "learning_rate": 1.403352291286114e-05,
+      "loss": 0.2375,
+      "step": 965
+    },
+    {
+      "epoch": 2.741159830268741,
+      "grad_norm": 1.3341835737228394,
+      "learning_rate": 1.3893692153567846e-05,
+      "loss": 0.2252,
+      "step": 970
+    },
+    {
+      "epoch": 2.7553041018387554,
+      "grad_norm": 1.3639682531356812,
+      "learning_rate": 1.3753957994281944e-05,
+      "loss": 0.2232,
+      "step": 975
+    },
+    {
+      "epoch": 2.7694483734087694,
+      "grad_norm": 1.3704020977020264,
+      "learning_rate": 1.3614332636237352e-05,
+      "loss": 0.2411,
+      "step": 980
+    },
+    {
+      "epoch": 2.783592644978784,
+      "grad_norm": 1.1877855062484741,
+      "learning_rate": 1.3474828271167735e-05,
+      "loss": 0.2096,
+      "step": 985
+    },
+    {
+      "epoch": 2.7977369165487977,
+      "grad_norm": 1.3705607652664185,
+      "learning_rate": 1.3335457080241966e-05,
+      "loss": 0.2333,
+      "step": 990
+    },
+    {
+      "epoch": 2.8118811881188117,
+      "grad_norm": 1.2100168466567993,
+      "learning_rate": 1.319623123300048e-05,
+      "loss": 0.2126,
+      "step": 995
+    },
+    {
+      "epoch": 2.826025459688826,
+      "grad_norm": 1.3423547744750977,
+      "learning_rate": 1.305716288629267e-05,
+      "loss": 0.2046,
+      "step": 1000
+    },
+    {
+      "epoch": 2.84016973125884,
+      "grad_norm": 1.3659006357192993,
+      "learning_rate": 1.2918264183215384e-05,
+      "loss": 0.1937,
+      "step": 1005
+    },
+    {
+      "epoch": 2.854314002828854,
+      "grad_norm": 1.1460381746292114,
+      "learning_rate": 1.2779547252052611e-05,
+      "loss": 0.2332,
+      "step": 1010
+    },
+    {
+      "epoch": 2.8684582743988685,
+      "grad_norm": 1.2629969120025635,
+      "learning_rate": 1.2641024205216478e-05,
+      "loss": 0.2037,
+      "step": 1015
+    },
+    {
+      "epoch": 2.882602545968883,
+      "grad_norm": 1.171892762184143,
+      "learning_rate": 1.250270713818963e-05,
+      "loss": 0.2328,
+      "step": 1020
+    },
+    {
+      "epoch": 2.896746817538897,
+      "grad_norm": 1.2251631021499634,
+      "learning_rate": 1.2364608128469059e-05,
+      "loss": 0.2081,
+      "step": 1025
+    },
+    {
+      "epoch": 2.910891089108911,
+      "grad_norm": 1.1573634147644043,
+      "learning_rate": 1.2226739234511554e-05,
+      "loss": 0.226,
+      "step": 1030
+    },
+    {
+      "epoch": 2.9250353606789252,
+      "grad_norm": 1.232812762260437,
+      "learning_rate": 1.2089112494680775e-05,
+      "loss": 0.2094,
+      "step": 1035
+    },
+    {
+      "epoch": 2.939179632248939,
+      "grad_norm": 1.2418936491012573,
+      "learning_rate": 1.1951739926196097e-05,
+      "loss": 0.2016,
+      "step": 1040
+    },
+    {
+      "epoch": 2.953323903818953,
+      "grad_norm": 1.1464600563049316,
+      "learning_rate": 1.1814633524083292e-05,
+      "loss": 0.2109,
+      "step": 1045
+    },
+    {
+      "epoch": 2.9674681753889676,
+      "grad_norm": 1.307362675666809,
+      "learning_rate": 1.167780526012716e-05,
+      "loss": 0.205,
+      "step": 1050
+    },
+    {
+      "epoch": 2.9816124469589815,
+      "grad_norm": 1.443464756011963,
+      "learning_rate": 1.1541267081826172e-05,
+      "loss": 0.1853,
+      "step": 1055
+    },
+    {
+      "epoch": 2.9957567185289955,
+      "grad_norm": 1.1249037981033325,
+      "learning_rate": 1.1405030911349266e-05,
+      "loss": 0.1821,
+      "step": 1060
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1770,
+  "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": 1.4714141712024863e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

40_128_e5_3e-5/checkpoint-1062/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8f1a562a572af6f17e99424943e6ee2738b5feac80076391e8a179272d29064
+size 7736

40_128_e5_3e-5/checkpoint-1062/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)

40_128_e5_3e-5/checkpoint-1416/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

40_128_e5_3e-5/checkpoint-1416/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": [
+    "down_proj",
+    "o_proj",
+    "v_proj",
+    "up_proj",
+    "gate_proj",
+    "k_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

40_128_e5_3e-5/checkpoint-1416/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7259780e9c8e048050202cd716e98f85b63d63b2965dc10c6e0f509459315c14
+size 791751704

40_128_e5_3e-5/checkpoint-1416/latest

@@ -0,0 +1 @@
+global_step1416

40_128_e5_3e-5/checkpoint-1416/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1114981e282a8933077345efb28a7586b0fbf50597f73f95bfc04646df278bab
+size 15920

40_128_e5_3e-5/checkpoint-1416/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:377afbae7a5948587d8562ef723c614da8201f687ea7e586f90243e0e85726c5
+size 15920

40_128_e5_3e-5/checkpoint-1416/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f5ec0b7763881f3a9d884790fe6c3f4aacaaeccd4408a507bc1f4468e4da657
+size 15920

40_128_e5_3e-5/checkpoint-1416/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d339c28329327d4da955a0f315c9307e285d69384b5d53d76368ab24764fda84
+size 15920

40_128_e5_3e-5/checkpoint-1416/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a0ef2c406b8e49e25297e61979d2d3bf87861565d0f8bfe926797bfc72a97c9
+size 15920

40_128_e5_3e-5/checkpoint-1416/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ada2df90e2f2c3b930f9c7953805bfd93c3bf66e0f9996c4125068889a99001a
+size 15920

40_128_e5_3e-5/checkpoint-1416/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36406aec7e4998cb5ad10dfe3543d4bffb6f296da01a697840107adf9b5badd0
+size 15920

40_128_e5_3e-5/checkpoint-1416/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4379c856d4e73e648f51f71b0df3b9a367903b9a14a2008001623b2323295371
+size 15920

40_128_e5_3e-5/checkpoint-1416/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e342a04509aa661573d5a4f7abbc043555051dd41965443bf89e020c181dc54e
+size 1064

40_128_e5_3e-5/checkpoint-1416/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
+  }
+}

40_128_e5_3e-5/checkpoint-1416/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
+}

40_128_e5_3e-5/checkpoint-1416/trainer_state.json

@@ -0,0 +1,2015 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1416,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.014144271570014143,
+      "grad_norm": 1.1091066598892212,
+      "learning_rate": 1.348314606741573e-06,
+      "loss": 1.3703,
+      "step": 5
+    },
+    {
+      "epoch": 0.028288543140028287,
+      "grad_norm": 1.1845866441726685,
+      "learning_rate": 3.033707865168539e-06,
+      "loss": 1.3488,
+      "step": 10
+    },
+    {
+      "epoch": 0.042432814710042434,
+      "grad_norm": 0.6645119190216064,
+      "learning_rate": 4.719101123595506e-06,
+      "loss": 1.3309,
+      "step": 15
+    },
+    {
+      "epoch": 0.056577086280056574,
+      "grad_norm": 0.6867905855178833,
+      "learning_rate": 6.404494382022472e-06,
+      "loss": 1.3852,
+      "step": 20
+    },
+    {
+      "epoch": 0.07072135785007072,
+      "grad_norm": 0.8682529330253601,
+      "learning_rate": 8.089887640449438e-06,
+      "loss": 1.2832,
+      "step": 25
+    },
+    {
+      "epoch": 0.08486562942008487,
+      "grad_norm": 0.5214539766311646,
+      "learning_rate": 9.775280898876405e-06,
+      "loss": 1.351,
+      "step": 30
+    },
+    {
+      "epoch": 0.09900990099009901,
+      "grad_norm": 0.5842329859733582,
+      "learning_rate": 1.146067415730337e-05,
+      "loss": 1.199,
+      "step": 35
+    },
+    {
+      "epoch": 0.11315417256011315,
+      "grad_norm": 0.5440396070480347,
+      "learning_rate": 1.3146067415730338e-05,
+      "loss": 1.2351,
+      "step": 40
+    },
+    {
+      "epoch": 0.1272984441301273,
+      "grad_norm": 0.5273528695106506,
+      "learning_rate": 1.4831460674157303e-05,
+      "loss": 1.2773,
+      "step": 45
+    },
+    {
+      "epoch": 0.14144271570014144,
+      "grad_norm": 0.5239897966384888,
+      "learning_rate": 1.651685393258427e-05,
+      "loss": 1.2557,
+      "step": 50
+    },
+    {
+      "epoch": 0.15558698727015557,
+      "grad_norm": 0.5751009583473206,
+      "learning_rate": 1.8202247191011237e-05,
+      "loss": 1.2738,
+      "step": 55
+    },
+    {
+      "epoch": 0.16973125884016974,
+      "grad_norm": 0.5215653777122498,
+      "learning_rate": 1.98876404494382e-05,
+      "loss": 1.2258,
+      "step": 60
+    },
+    {
+      "epoch": 0.18387553041018387,
+      "grad_norm": 0.5515840649604797,
+      "learning_rate": 2.1573033707865168e-05,
+      "loss": 1.1885,
+      "step": 65
+    },
+    {
+      "epoch": 0.19801980198019803,
+      "grad_norm": 0.5996809601783752,
+      "learning_rate": 2.3258426966292135e-05,
+      "loss": 1.1447,
+      "step": 70
+    },
+    {
+      "epoch": 0.21216407355021216,
+      "grad_norm": 0.6658637523651123,
+      "learning_rate": 2.4943820224719103e-05,
+      "loss": 1.1789,
+      "step": 75
+    },
+    {
+      "epoch": 0.2263083451202263,
+      "grad_norm": 0.5400289297103882,
+      "learning_rate": 2.6629213483146066e-05,
+      "loss": 1.1788,
+      "step": 80
+    },
+    {
+      "epoch": 0.24045261669024046,
+      "grad_norm": 0.6226291060447693,
+      "learning_rate": 2.8314606741573034e-05,
+      "loss": 1.1713,
+      "step": 85
+    },
+    {
+      "epoch": 0.2545968882602546,
+      "grad_norm": 0.6412995457649231,
+      "learning_rate": 3e-05,
+      "loss": 1.1693,
+      "step": 90
+    },
+    {
+      "epoch": 0.26874115983026875,
+      "grad_norm": 0.6276025176048279,
+      "learning_rate": 2.9999345118939752e-05,
+      "loss": 1.1233,
+      "step": 95
+    },
+    {
+      "epoch": 0.2828854314002829,
+      "grad_norm": 0.5257118344306946,
+      "learning_rate": 2.9997380532941557e-05,
+      "loss": 1.0805,
+      "step": 100
+    },
+    {
+      "epoch": 0.297029702970297,
+      "grad_norm": 0.6102027893066406,
+      "learning_rate": 2.999410641354812e-05,
+      "loss": 1.1336,
+      "step": 105
+    },
+    {
+      "epoch": 0.31117397454031115,
+      "grad_norm": 0.6746109127998352,
+      "learning_rate": 2.9989523046647264e-05,
+      "loss": 1.1515,
+      "step": 110
+    },
+    {
+      "epoch": 0.32531824611032534,
+      "grad_norm": 0.6818143129348755,
+      "learning_rate": 2.9983630832447017e-05,
+      "loss": 1.0251,
+      "step": 115
+    },
+    {
+      "epoch": 0.33946251768033947,
+      "grad_norm": 0.6401700973510742,
+      "learning_rate": 2.9976430285440642e-05,
+      "loss": 1.0817,
+      "step": 120
+    },
+    {
+      "epoch": 0.3536067892503536,
+      "grad_norm": 0.6563226580619812,
+      "learning_rate": 2.9967922034361726e-05,
+      "loss": 1.0144,
+      "step": 125
+    },
+    {
+      "epoch": 0.36775106082036774,
+      "grad_norm": 0.7533389925956726,
+      "learning_rate": 2.995810682212926e-05,
+      "loss": 1.0501,
+      "step": 130
+    },
+    {
+      "epoch": 0.38189533239038187,
+      "grad_norm": 0.6909534335136414,
+      "learning_rate": 2.9946985505782792e-05,
+      "loss": 1.0284,
+      "step": 135
+    },
+    {
+      "epoch": 0.39603960396039606,
+      "grad_norm": 0.639701783657074,
+      "learning_rate": 2.993455905640758e-05,
+      "loss": 1.0045,
+      "step": 140
+    },
+    {
+      "epoch": 0.4101838755304102,
+      "grad_norm": 1.8686423301696777,
+      "learning_rate": 2.9920828559049805e-05,
+      "loss": 0.9764,
+      "step": 145
+    },
+    {
+      "epoch": 0.4243281471004243,
+      "grad_norm": 0.7100444436073303,
+      "learning_rate": 2.9905795212621825e-05,
+      "loss": 1.0063,
+      "step": 150
+    },
+    {
+      "epoch": 0.43847241867043846,
+      "grad_norm": 0.8259682059288025,
+      "learning_rate": 2.9889460329797482e-05,
+      "loss": 0.9937,
+      "step": 155
+    },
+    {
+      "epoch": 0.4526166902404526,
+      "grad_norm": 0.6790755391120911,
+      "learning_rate": 2.9871825336897495e-05,
+      "loss": 0.9834,
+      "step": 160
+    },
+    {
+      "epoch": 0.4667609618104668,
+      "grad_norm": 0.7154070734977722,
+      "learning_rate": 2.985289177376491e-05,
+      "loss": 0.9921,
+      "step": 165
+    },
+    {
+      "epoch": 0.4809052333804809,
+      "grad_norm": 0.8498982787132263,
+      "learning_rate": 2.983266129363065e-05,
+      "loss": 0.9274,
+      "step": 170
+    },
+    {
+      "epoch": 0.49504950495049505,
+      "grad_norm": 0.7261009216308594,
+      "learning_rate": 2.9811135662969146e-05,
+      "loss": 0.9445,
+      "step": 175
+    },
+    {
+      "epoch": 0.5091937765205092,
+      "grad_norm": 1.0621588230133057,
+      "learning_rate": 2.978831676134411e-05,
+      "loss": 0.9137,
+      "step": 180
+    },
+    {
+      "epoch": 0.5233380480905233,
+      "grad_norm": 0.9448386430740356,
+      "learning_rate": 2.9764206581244415e-05,
+      "loss": 0.9553,
+      "step": 185
+    },
+    {
+      "epoch": 0.5374823196605375,
+      "grad_norm": 0.8367416858673096,
+      "learning_rate": 2.973880722791009e-05,
+      "loss": 0.8912,
+      "step": 190
+    },
+    {
+      "epoch": 0.5516265912305516,
+      "grad_norm": 0.8499005436897278,
+      "learning_rate": 2.971212091914854e-05,
+      "loss": 0.9632,
+      "step": 195
+    },
+    {
+      "epoch": 0.5657708628005658,
+      "grad_norm": 0.7935090065002441,
+      "learning_rate": 2.968414998514085e-05,
+      "loss": 0.8724,
+      "step": 200
+    },
+    {
+      "epoch": 0.57991513437058,
+      "grad_norm": 0.8054782152175903,
+      "learning_rate": 2.9654896868238334e-05,
+      "loss": 0.8849,
+      "step": 205
+    },
+    {
+      "epoch": 0.594059405940594,
+      "grad_norm": 0.9245705008506775,
+      "learning_rate": 2.96243641227493e-05,
+      "loss": 0.8934,
+      "step": 210
+    },
+    {
+      "epoch": 0.6082036775106082,
+      "grad_norm": 0.851338803768158,
+      "learning_rate": 2.959255441471597e-05,
+      "loss": 0.8839,
+      "step": 215
+    },
+    {
+      "epoch": 0.6223479490806223,
+      "grad_norm": 0.9376785159111023,
+      "learning_rate": 2.9559470521681726e-05,
+      "loss": 0.8592,
+      "step": 220
+    },
+    {
+      "epoch": 0.6364922206506365,
+      "grad_norm": 0.8601529002189636,
+      "learning_rate": 2.9525115332448557e-05,
+      "loss": 0.8879,
+      "step": 225
+    },
+    {
+      "epoch": 0.6506364922206507,
+      "grad_norm": 0.9360470175743103,
+      "learning_rate": 2.948949184682484e-05,
+      "loss": 0.8337,
+      "step": 230
+    },
+    {
+      "epoch": 0.6647807637906648,
+      "grad_norm": 0.9480540156364441,
+      "learning_rate": 2.9452603175363364e-05,
+      "loss": 0.9082,
+      "step": 235
+    },
+    {
+      "epoch": 0.6789250353606789,
+      "grad_norm": 0.9365689158439636,
+      "learning_rate": 2.941445253908978e-05,
+      "loss": 0.8773,
+      "step": 240
+    },
+    {
+      "epoch": 0.693069306930693,
+      "grad_norm": 0.9173905253410339,
+      "learning_rate": 2.9375043269221294e-05,
+      "loss": 0.81,
+      "step": 245
+    },
+    {
+      "epoch": 0.7072135785007072,
+      "grad_norm": 0.9756169319152832,
+      "learning_rate": 2.9334378806875838e-05,
+      "loss": 0.8046,
+      "step": 250
+    },
+    {
+      "epoch": 0.7213578500707214,
+      "grad_norm": 0.9777851104736328,
+      "learning_rate": 2.929246270277157e-05,
+      "loss": 0.8452,
+      "step": 255
+    },
+    {
+      "epoch": 0.7355021216407355,
+      "grad_norm": 0.9113903045654297,
+      "learning_rate": 2.9249298616916856e-05,
+      "loss": 0.8244,
+      "step": 260
+    },
+    {
+      "epoch": 0.7496463932107497,
+      "grad_norm": 0.906170666217804,
+      "learning_rate": 2.9204890318290666e-05,
+      "loss": 0.7866,
+      "step": 265
+    },
+    {
+      "epoch": 0.7637906647807637,
+      "grad_norm": 0.9952336549758911,
+      "learning_rate": 2.915924168451349e-05,
+      "loss": 0.817,
+      "step": 270
+    },
+    {
+      "epoch": 0.7779349363507779,
+      "grad_norm": 1.1563059091567993,
+      "learning_rate": 2.9112356701508756e-05,
+      "loss": 0.8101,
+      "step": 275
+    },
+    {
+      "epoch": 0.7920792079207921,
+      "grad_norm": 1.035338282585144,
+      "learning_rate": 2.9064239463154782e-05,
+      "loss": 0.7379,
+      "step": 280
+    },
+    {
+      "epoch": 0.8062234794908062,
+      "grad_norm": 1.0756317377090454,
+      "learning_rate": 2.9014894170927308e-05,
+      "loss": 0.7709,
+      "step": 285
+    },
+    {
+      "epoch": 0.8203677510608204,
+      "grad_norm": 1.1403799057006836,
+      "learning_rate": 2.8964325133532638e-05,
+      "loss": 0.8136,
+      "step": 290
+    },
+    {
+      "epoch": 0.8345120226308345,
+      "grad_norm": 1.111162543296814,
+      "learning_rate": 2.8912536766531424e-05,
+      "loss": 0.7563,
+      "step": 295
+    },
+    {
+      "epoch": 0.8486562942008486,
+      "grad_norm": 0.9991361498832703,
+      "learning_rate": 2.885953359195308e-05,
+      "loss": 0.74,
+      "step": 300
+    },
+    {
+      "epoch": 0.8628005657708628,
+      "grad_norm": 0.9173119068145752,
+      "learning_rate": 2.8805320237900965e-05,
+      "loss": 0.7599,
+      "step": 305
+    },
+    {
+      "epoch": 0.8769448373408769,
+      "grad_norm": 1.1988530158996582,
+      "learning_rate": 2.874990143814826e-05,
+      "loss": 0.7256,
+      "step": 310
+    },
+    {
+      "epoch": 0.8910891089108911,
+      "grad_norm": 1.0119719505310059,
+      "learning_rate": 2.8693282031724592e-05,
+      "loss": 0.7472,
+      "step": 315
+    },
+    {
+      "epoch": 0.9052333804809052,
+      "grad_norm": 1.0434908866882324,
+      "learning_rate": 2.8635466962493564e-05,
+      "loss": 0.7512,
+      "step": 320
+    },
+    {
+      "epoch": 0.9193776520509194,
+      "grad_norm": 1.0775619745254517,
+      "learning_rate": 2.8576461278721016e-05,
+      "loss": 0.6982,
+      "step": 325
+    },
+    {
+      "epoch": 0.9335219236209336,
+      "grad_norm": 0.9957691431045532,
+      "learning_rate": 2.8516270132634244e-05,
+      "loss": 0.6556,
+      "step": 330
+    },
+    {
+      "epoch": 0.9476661951909476,
+      "grad_norm": 1.0233315229415894,
+      "learning_rate": 2.845489877997213e-05,
+      "loss": 0.7431,
+      "step": 335
+    },
+    {
+      "epoch": 0.9618104667609618,
+      "grad_norm": 0.9704013466835022,
+      "learning_rate": 2.8392352579526205e-05,
+      "loss": 0.7121,
+      "step": 340
+    },
+    {
+      "epoch": 0.9759547383309759,
+      "grad_norm": 1.0681896209716797,
+      "learning_rate": 2.832863699267274e-05,
+      "loss": 0.716,
+      "step": 345
+    },
+    {
+      "epoch": 0.9900990099009901,
+      "grad_norm": 1.088863492012024,
+      "learning_rate": 2.8263757582895888e-05,
+      "loss": 0.699,
+      "step": 350
+    },
+    {
+      "epoch": 1.002828854314003,
+      "grad_norm": 1.075731873512268,
+      "learning_rate": 2.819772001530186e-05,
+      "loss": 0.647,
+      "step": 355
+    },
+    {
+      "epoch": 1.016973125884017,
+      "grad_norm": 1.180709958076477,
+      "learning_rate": 2.8130530056124298e-05,
+      "loss": 0.6268,
+      "step": 360
+    },
+    {
+      "epoch": 1.031117397454031,
+      "grad_norm": 1.0877044200897217,
+      "learning_rate": 2.8062193572220763e-05,
+      "loss": 0.638,
+      "step": 365
+    },
+    {
+      "epoch": 1.0452616690240453,
+      "grad_norm": 1.0629935264587402,
+      "learning_rate": 2.799271653056046e-05,
+      "loss": 0.6209,
+      "step": 370
+    },
+    {
+      "epoch": 1.0594059405940595,
+      "grad_norm": 1.149365782737732,
+      "learning_rate": 2.7922104997703214e-05,
+      "loss": 0.5914,
+      "step": 375
+    },
+    {
+      "epoch": 1.0735502121640736,
+      "grad_norm": 1.2020652294158936,
+      "learning_rate": 2.785036513926976e-05,
+      "loss": 0.6314,
+      "step": 380
+    },
+    {
+      "epoch": 1.0876944837340876,
+      "grad_norm": 1.1636576652526855,
+      "learning_rate": 2.7777503219403374e-05,
+      "loss": 0.6092,
+      "step": 385
+    },
+    {
+      "epoch": 1.1018387553041018,
+      "grad_norm": 1.1622604131698608,
+      "learning_rate": 2.7703525600222896e-05,
+      "loss": 0.5938,
+      "step": 390
+    },
+    {
+      "epoch": 1.115983026874116,
+      "grad_norm": 1.1172856092453003,
+      "learning_rate": 2.7628438741267218e-05,
+      "loss": 0.5954,
+      "step": 395
+    },
+    {
+      "epoch": 1.1301272984441302,
+      "grad_norm": 1.2672319412231445,
+      "learning_rate": 2.7552249198931255e-05,
+      "loss": 0.5924,
+      "step": 400
+    },
+    {
+      "epoch": 1.1442715700141444,
+      "grad_norm": 1.1092612743377686,
+      "learning_rate": 2.7474963625893433e-05,
+      "loss": 0.5816,
+      "step": 405
+    },
+    {
+      "epoch": 1.1584158415841583,
+      "grad_norm": 1.2071410417556763,
+      "learning_rate": 2.7396588770534822e-05,
+      "loss": 0.5221,
+      "step": 410
+    },
+    {
+      "epoch": 1.1725601131541725,
+      "grad_norm": 1.2543182373046875,
+      "learning_rate": 2.7317131476349874e-05,
+      "loss": 0.5432,
+      "step": 415
+    },
+    {
+      "epoch": 1.1867043847241867,
+      "grad_norm": 1.193509578704834,
+      "learning_rate": 2.723659868134886e-05,
+      "loss": 0.6123,
+      "step": 420
+    },
+    {
+      "epoch": 1.200848656294201,
+      "grad_norm": 1.1950852870941162,
+      "learning_rate": 2.715499741745208e-05,
+      "loss": 0.5481,
+      "step": 425
+    },
+    {
+      "epoch": 1.214992927864215,
+      "grad_norm": 1.1583995819091797,
+      "learning_rate": 2.707233480987582e-05,
+      "loss": 0.5594,
+      "step": 430
+    },
+    {
+      "epoch": 1.229137199434229,
+      "grad_norm": 1.2848454713821411,
+      "learning_rate": 2.6988618076510227e-05,
+      "loss": 0.5634,
+      "step": 435
+    },
+    {
+      "epoch": 1.2432814710042432,
+      "grad_norm": 1.0936193466186523,
+      "learning_rate": 2.6903854527289047e-05,
+      "loss": 0.5672,
+      "step": 440
+    },
+    {
+      "epoch": 1.2574257425742574,
+      "grad_norm": 1.2483489513397217,
+      "learning_rate": 2.6818051563551355e-05,
+      "loss": 0.5401,
+      "step": 445
+    },
+    {
+      "epoch": 1.2715700141442716,
+      "grad_norm": 1.2003945112228394,
+      "learning_rate": 2.6731216677395253e-05,
+      "loss": 0.5686,
+      "step": 450
+    },
+    {
+      "epoch": 1.2857142857142856,
+      "grad_norm": 1.4688889980316162,
+      "learning_rate": 2.6643357451023725e-05,
+      "loss": 0.5277,
+      "step": 455
+    },
+    {
+      "epoch": 1.2998585572842998,
+      "grad_norm": 1.1361063718795776,
+      "learning_rate": 2.6554481556082543e-05,
+      "loss": 0.5116,
+      "step": 460
+    },
+    {
+      "epoch": 1.314002828854314,
+      "grad_norm": 1.3163748979568481,
+      "learning_rate": 2.646459675299042e-05,
+      "loss": 0.5612,
+      "step": 465
+    },
+    {
+      "epoch": 1.3281471004243282,
+      "grad_norm": 1.1666852235794067,
+      "learning_rate": 2.637371089026137e-05,
+      "loss": 0.5716,
+      "step": 470
+    },
+    {
+      "epoch": 1.3422913719943423,
+      "grad_norm": 1.1531682014465332,
+      "learning_rate": 2.6281831903819413e-05,
+      "loss": 0.5386,
+      "step": 475
+    },
+    {
+      "epoch": 1.3564356435643563,
+      "grad_norm": 1.0326910018920898,
+      "learning_rate": 2.6188967816305624e-05,
+      "loss": 0.5494,
+      "step": 480
+    },
+    {
+      "epoch": 1.3705799151343705,
+      "grad_norm": 1.110970377922058,
+      "learning_rate": 2.6095126736377618e-05,
+      "loss": 0.487,
+      "step": 485
+    },
+    {
+      "epoch": 1.3847241867043847,
+      "grad_norm": 1.253021001815796,
+      "learning_rate": 2.6000316858001512e-05,
+      "loss": 0.5005,
+      "step": 490
+    },
+    {
+      "epoch": 1.3988684582743989,
+      "grad_norm": 1.1914869546890259,
+      "learning_rate": 2.5904546459736468e-05,
+      "loss": 0.4999,
+      "step": 495
+    },
+    {
+      "epoch": 1.413012729844413,
+      "grad_norm": 1.1897658109664917,
+      "learning_rate": 2.5807823904011803e-05,
+      "loss": 0.518,
+      "step": 500
+    },
+    {
+      "epoch": 1.427157001414427,
+      "grad_norm": 1.1284512281417847,
+      "learning_rate": 2.5710157636396845e-05,
+      "loss": 0.4999,
+      "step": 505
+    },
+    {
+      "epoch": 1.4413012729844414,
+      "grad_norm": 1.2429921627044678,
+      "learning_rate": 2.5611556184863437e-05,
+      "loss": 0.5009,
+      "step": 510
+    },
+    {
+      "epoch": 1.4554455445544554,
+      "grad_norm": 1.276069164276123,
+      "learning_rate": 2.5512028159041328e-05,
+      "loss": 0.4511,
+      "step": 515
+    },
+    {
+      "epoch": 1.4695898161244696,
+      "grad_norm": 1.2455278635025024,
+      "learning_rate": 2.54115822494664e-05,
+      "loss": 0.483,
+      "step": 520
+    },
+    {
+      "epoch": 1.4837340876944838,
+      "grad_norm": 1.2925827503204346,
+      "learning_rate": 2.5310227226821814e-05,
+      "loss": 0.4746,
+      "step": 525
+    },
+    {
+      "epoch": 1.4978783592644977,
+      "grad_norm": 1.246673345565796,
+      "learning_rate": 2.5207971941172193e-05,
+      "loss": 0.4656,
+      "step": 530
+    },
+    {
+      "epoch": 1.5120226308345122,
+      "grad_norm": 1.27567720413208,
+      "learning_rate": 2.510482532119087e-05,
+      "loss": 0.4722,
+      "step": 535
+    },
+    {
+      "epoch": 1.5261669024045261,
+      "grad_norm": 1.2516908645629883,
+      "learning_rate": 2.500079637338021e-05,
+      "loss": 0.4791,
+      "step": 540
+    },
+    {
+      "epoch": 1.5403111739745403,
+      "grad_norm": 1.4572430849075317,
+      "learning_rate": 2.489589418128524e-05,
+      "loss": 0.4871,
+      "step": 545
+    },
+    {
+      "epoch": 1.5544554455445545,
+      "grad_norm": 1.2278729677200317,
+      "learning_rate": 2.479012790470046e-05,
+      "loss": 0.4845,
+      "step": 550
+    },
+    {
+      "epoch": 1.5685997171145685,
+      "grad_norm": 1.2259896993637085,
+      "learning_rate": 2.468350677887005e-05,
+      "loss": 0.4505,
+      "step": 555
+    },
+    {
+      "epoch": 1.5827439886845829,
+      "grad_norm": 1.0976181030273438,
+      "learning_rate": 2.457604011368147e-05,
+      "loss": 0.439,
+      "step": 560
+    },
+    {
+      "epoch": 1.5968882602545968,
+      "grad_norm": 1.3767074346542358,
+      "learning_rate": 2.4467737292852542e-05,
+      "loss": 0.4443,
+      "step": 565
+    },
+    {
+      "epoch": 1.611032531824611,
+      "grad_norm": 1.2550078630447388,
+      "learning_rate": 2.4358607773112074e-05,
+      "loss": 0.4749,
+      "step": 570
+    },
+    {
+      "epoch": 1.6251768033946252,
+      "grad_norm": 1.2857836484909058,
+      "learning_rate": 2.424866108337415e-05,
+      "loss": 0.4499,
+      "step": 575
+    },
+    {
+      "epoch": 1.6393210749646392,
+      "grad_norm": 1.2040728330612183,
+      "learning_rate": 2.413790682390607e-05,
+      "loss": 0.454,
+      "step": 580
+    },
+    {
+      "epoch": 1.6534653465346536,
+      "grad_norm": 1.22360360622406,
+      "learning_rate": 2.402635466549008e-05,
+      "loss": 0.4122,
+      "step": 585
+    },
+    {
+      "epoch": 1.6676096181046676,
+      "grad_norm": 1.181302785873413,
+      "learning_rate": 2.3914014348578957e-05,
+      "loss": 0.4265,
+      "step": 590
+    },
+    {
+      "epoch": 1.6817538896746818,
+      "grad_norm": 1.2027560472488403,
+      "learning_rate": 2.3800895682445475e-05,
+      "loss": 0.4536,
+      "step": 595
+    },
+    {
+      "epoch": 1.695898161244696,
+      "grad_norm": 1.2701138257980347,
+      "learning_rate": 2.3687008544325904e-05,
+      "loss": 0.4209,
+      "step": 600
+    },
+    {
+      "epoch": 1.71004243281471,
+      "grad_norm": 1.1245002746582031,
+      "learning_rate": 2.3572362878557546e-05,
+      "loss": 0.4382,
+      "step": 605
+    },
+    {
+      "epoch": 1.7241867043847243,
+      "grad_norm": 1.1181449890136719,
+      "learning_rate": 2.3456968695710415e-05,
+      "loss": 0.4125,
+      "step": 610
+    },
+    {
+      "epoch": 1.7383309759547383,
+      "grad_norm": 1.2436251640319824,
+      "learning_rate": 2.3340836071713164e-05,
+      "loss": 0.3983,
+      "step": 615
+    },
+    {
+      "epoch": 1.7524752475247525,
+      "grad_norm": 1.3499740362167358,
+      "learning_rate": 2.322397514697324e-05,
+      "loss": 0.4585,
+      "step": 620
+    },
+    {
+      "epoch": 1.7666195190947667,
+      "grad_norm": 1.1281145811080933,
+      "learning_rate": 2.310639612549149e-05,
+      "loss": 0.377,
+      "step": 625
+    },
+    {
+      "epoch": 1.7807637906647806,
+      "grad_norm": 1.2632691860198975,
+      "learning_rate": 2.298810927397114e-05,
+      "loss": 0.4788,
+      "step": 630
+    },
+    {
+      "epoch": 1.794908062234795,
+      "grad_norm": 1.3099757432937622,
+      "learning_rate": 2.286912492092136e-05,
+      "loss": 0.3682,
+      "step": 635
+    },
+    {
+      "epoch": 1.809052333804809,
+      "grad_norm": 1.3557870388031006,
+      "learning_rate": 2.2749453455755392e-05,
+      "loss": 0.3754,
+      "step": 640
+    },
+    {
+      "epoch": 1.8231966053748232,
+      "grad_norm": 1.1247097253799438,
+      "learning_rate": 2.2629105327883354e-05,
+      "loss": 0.3961,
+      "step": 645
+    },
+    {
+      "epoch": 1.8373408769448374,
+      "grad_norm": 1.1884528398513794,
+      "learning_rate": 2.250809104579987e-05,
+      "loss": 0.3828,
+      "step": 650
+    },
+    {
+      "epoch": 1.8514851485148514,
+      "grad_norm": 1.119057297706604,
+      "learning_rate": 2.2386421176166445e-05,
+      "loss": 0.404,
+      "step": 655
+    },
+    {
+      "epoch": 1.8656294200848658,
+      "grad_norm": 1.1483960151672363,
+      "learning_rate": 2.2264106342888853e-05,
+      "loss": 0.3872,
+      "step": 660
+    },
+    {
+      "epoch": 1.8797736916548797,
+      "grad_norm": 1.416157603263855,
+      "learning_rate": 2.2141157226189442e-05,
+      "loss": 0.3667,
+      "step": 665
+    },
+    {
+      "epoch": 1.893917963224894,
+      "grad_norm": 1.1745518445968628,
+      "learning_rate": 2.2017584561674605e-05,
+      "loss": 0.4214,
+      "step": 670
+    },
+    {
+      "epoch": 1.908062234794908,
+      "grad_norm": 1.1854530572891235,
+      "learning_rate": 2.1893399139397347e-05,
+      "loss": 0.411,
+      "step": 675
+    },
+    {
+      "epoch": 1.922206506364922,
+      "grad_norm": 1.1574931144714355,
+      "learning_rate": 2.176861180291514e-05,
+      "loss": 0.3547,
+      "step": 680
+    },
+    {
+      "epoch": 1.9363507779349365,
+      "grad_norm": 1.4279593229293823,
+      "learning_rate": 2.1643233448343077e-05,
+      "loss": 0.4094,
+      "step": 685
+    },
+    {
+      "epoch": 1.9504950495049505,
+      "grad_norm": 1.2585395574569702,
+      "learning_rate": 2.1517275023402463e-05,
+      "loss": 0.3927,
+      "step": 690
+    },
+    {
+      "epoch": 1.9646393210749646,
+      "grad_norm": 1.2487115859985352,
+      "learning_rate": 2.1390747526464885e-05,
+      "loss": 0.3876,
+      "step": 695
+    },
+    {
+      "epoch": 1.9787835926449788,
+      "grad_norm": 1.1419602632522583,
+      "learning_rate": 2.126366200559184e-05,
+      "loss": 0.3809,
+      "step": 700
+    },
+    {
+      "epoch": 1.9929278642149928,
+      "grad_norm": 1.240652084350586,
+      "learning_rate": 2.11360295575701e-05,
+      "loss": 0.3525,
+      "step": 705
+    },
+    {
+      "epoch": 2.005657708628006,
+      "grad_norm": 1.1952710151672363,
+      "learning_rate": 2.100786132694271e-05,
+      "loss": 0.3126,
+      "step": 710
+    },
+    {
+      "epoch": 2.01980198019802,
+      "grad_norm": 1.3939076662063599,
+      "learning_rate": 2.0879168505035898e-05,
+      "loss": 0.2736,
+      "step": 715
+    },
+    {
+      "epoch": 2.033946251768034,
+      "grad_norm": 1.285657525062561,
+      "learning_rate": 2.074996232898189e-05,
+      "loss": 0.3355,
+      "step": 720
+    },
+    {
+      "epoch": 2.048090523338048,
+      "grad_norm": 1.210814356803894,
+      "learning_rate": 2.0620254080737697e-05,
+      "loss": 0.3003,
+      "step": 725
+    },
+    {
+      "epoch": 2.062234794908062,
+      "grad_norm": 1.160455346107483,
+      "learning_rate": 2.0490055086100004e-05,
+      "loss": 0.2567,
+      "step": 730
+    },
+    {
+      "epoch": 2.0763790664780766,
+      "grad_norm": 1.3446382284164429,
+      "learning_rate": 2.0359376713716224e-05,
+      "loss": 0.2943,
+      "step": 735
+    },
+    {
+      "epoch": 2.0905233380480905,
+      "grad_norm": 1.1885087490081787,
+      "learning_rate": 2.022823037409184e-05,
+      "loss": 0.277,
+      "step": 740
+    },
+    {
+      "epoch": 2.1046676096181045,
+      "grad_norm": 1.2343266010284424,
+      "learning_rate": 2.0096627518594035e-05,
+      "loss": 0.2947,
+      "step": 745
+    },
+    {
+      "epoch": 2.118811881188119,
+      "grad_norm": 1.2885913848876953,
+      "learning_rate": 1.9964579638451827e-05,
+      "loss": 0.2829,
+      "step": 750
+    },
+    {
+      "epoch": 2.132956152758133,
+      "grad_norm": 1.389346718788147,
+      "learning_rate": 1.9832098263752637e-05,
+      "loss": 0.3234,
+      "step": 755
+    },
+    {
+      "epoch": 2.1471004243281473,
+      "grad_norm": 1.2736543416976929,
+      "learning_rate": 1.9699194962435555e-05,
+      "loss": 0.2745,
+      "step": 760
+    },
+    {
+      "epoch": 2.1612446958981613,
+      "grad_norm": 1.1569143533706665,
+      "learning_rate": 1.9565881339281232e-05,
+      "loss": 0.2755,
+      "step": 765
+    },
+    {
+      "epoch": 2.1753889674681752,
+      "grad_norm": 1.2544715404510498,
+      "learning_rate": 1.9432169034898583e-05,
+      "loss": 0.274,
+      "step": 770
+    },
+    {
+      "epoch": 2.1895332390381896,
+      "grad_norm": 1.2836891412734985,
+      "learning_rate": 1.9298069724708365e-05,
+      "loss": 0.2877,
+      "step": 775
+    },
+    {
+      "epoch": 2.2036775106082036,
+      "grad_norm": 1.2649877071380615,
+      "learning_rate": 1.9163595117923697e-05,
+      "loss": 0.2787,
+      "step": 780
+    },
+    {
+      "epoch": 2.217821782178218,
+      "grad_norm": 1.4977940320968628,
+      "learning_rate": 1.9028756956527656e-05,
+      "loss": 0.277,
+      "step": 785
+    },
+    {
+      "epoch": 2.231966053748232,
+      "grad_norm": 1.2288274765014648,
+      "learning_rate": 1.8893567014247997e-05,
+      "loss": 0.3017,
+      "step": 790
+    },
+    {
+      "epoch": 2.246110325318246,
+      "grad_norm": 1.256234884262085,
+      "learning_rate": 1.8758037095529074e-05,
+      "loss": 0.2797,
+      "step": 795
+    },
+    {
+      "epoch": 2.2602545968882604,
+      "grad_norm": 1.199599027633667,
+      "learning_rate": 1.8622179034501145e-05,
+      "loss": 0.2632,
+      "step": 800
+    },
+    {
+      "epoch": 2.2743988684582743,
+      "grad_norm": 1.0664340257644653,
+      "learning_rate": 1.848600469394701e-05,
+      "loss": 0.2737,
+      "step": 805
+    },
+    {
+      "epoch": 2.2885431400282887,
+      "grad_norm": 1.302608609199524,
+      "learning_rate": 1.8349525964266206e-05,
+      "loss": 0.2544,
+      "step": 810
+    },
+    {
+      "epoch": 2.3026874115983027,
+      "grad_norm": 1.4746507406234741,
+      "learning_rate": 1.821275476243676e-05,
+      "loss": 0.2653,
+      "step": 815
+    },
+    {
+      "epoch": 2.3168316831683167,
+      "grad_norm": 1.360602855682373,
+      "learning_rate": 1.8075703030974628e-05,
+      "loss": 0.2759,
+      "step": 820
+    },
+    {
+      "epoch": 2.330975954738331,
+      "grad_norm": 1.1771636009216309,
+      "learning_rate": 1.7938382736890903e-05,
+      "loss": 0.2517,
+      "step": 825
+    },
+    {
+      "epoch": 2.345120226308345,
+      "grad_norm": 1.1717984676361084,
+      "learning_rate": 1.7800805870646895e-05,
+      "loss": 0.2662,
+      "step": 830
+    },
+    {
+      "epoch": 2.3592644978783595,
+      "grad_norm": 1.2912698984146118,
+      "learning_rate": 1.7662984445107133e-05,
+      "loss": 0.2536,
+      "step": 835
+    },
+    {
+      "epoch": 2.3734087694483734,
+      "grad_norm": 1.3276891708374023,
+      "learning_rate": 1.752493049449046e-05,
+      "loss": 0.259,
+      "step": 840
+    },
+    {
+      "epoch": 2.3875530410183874,
+      "grad_norm": 1.1467636823654175,
+      "learning_rate": 1.7386656073319216e-05,
+      "loss": 0.2556,
+      "step": 845
+    },
+    {
+      "epoch": 2.401697312588402,
+      "grad_norm": 1.110716700553894,
+      "learning_rate": 1.7248173255366672e-05,
+      "loss": 0.2669,
+      "step": 850
+    },
+    {
+      "epoch": 2.4158415841584158,
+      "grad_norm": 1.1397027969360352,
+      "learning_rate": 1.7109494132602784e-05,
+      "loss": 0.2524,
+      "step": 855
+    },
+    {
+      "epoch": 2.42998585572843,
+      "grad_norm": 1.236080527305603,
+      "learning_rate": 1.6970630814138337e-05,
+      "loss": 0.2581,
+      "step": 860
+    },
+    {
+      "epoch": 2.444130127298444,
+      "grad_norm": 1.2145838737487793,
+      "learning_rate": 1.6831595425167635e-05,
+      "loss": 0.2544,
+      "step": 865
+    },
+    {
+      "epoch": 2.458274398868458,
+      "grad_norm": 1.1833267211914062,
+      "learning_rate": 1.6692400105909737e-05,
+      "loss": 0.2658,
+      "step": 870
+    },
+    {
+      "epoch": 2.4724186704384725,
+      "grad_norm": 1.1189638376235962,
+      "learning_rate": 1.6553057010548408e-05,
+      "loss": 0.2798,
+      "step": 875
+    },
+    {
+      "epoch": 2.4865629420084865,
+      "grad_norm": 1.1513794660568237,
+      "learning_rate": 1.641357830617085e-05,
+      "loss": 0.2461,
+      "step": 880
+    },
+    {
+      "epoch": 2.500707213578501,
+      "grad_norm": 1.157867431640625,
+      "learning_rate": 1.6273976171705313e-05,
+      "loss": 0.2453,
+      "step": 885
+    },
+    {
+      "epoch": 2.514851485148515,
+      "grad_norm": 1.2551815509796143,
+      "learning_rate": 1.6134262796857628e-05,
+      "loss": 0.2473,
+      "step": 890
+    },
+    {
+      "epoch": 2.528995756718529,
+      "grad_norm": 1.3023325204849243,
+      "learning_rate": 1.599445038104688e-05,
+      "loss": 0.2495,
+      "step": 895
+    },
+    {
+      "epoch": 2.5431400282885432,
+      "grad_norm": 1.4266198873519897,
+      "learning_rate": 1.5854551132340145e-05,
+      "loss": 0.2507,
+      "step": 900
+    },
+    {
+      "epoch": 2.557284299858557,
+      "grad_norm": 1.084045171737671,
+      "learning_rate": 1.5714577266386526e-05,
+      "loss": 0.2485,
+      "step": 905
+    },
+    {
+      "epoch": 2.571428571428571,
+      "grad_norm": 1.359621524810791,
+      "learning_rate": 1.557454100535053e-05,
+      "loss": 0.2291,
+      "step": 910
+    },
+    {
+      "epoch": 2.5855728429985856,
+      "grad_norm": 1.2676442861557007,
+      "learning_rate": 1.543445457684484e-05,
+      "loss": 0.2468,
+      "step": 915
+    },
+    {
+      "epoch": 2.5997171145685996,
+      "grad_norm": 1.1858233213424683,
+      "learning_rate": 1.5294330212862624e-05,
+      "loss": 0.2343,
+      "step": 920
+    },
+    {
+      "epoch": 2.613861386138614,
+      "grad_norm": 1.1932803392410278,
+      "learning_rate": 1.5154180148709494e-05,
+      "loss": 0.2596,
+      "step": 925
+    },
+    {
+      "epoch": 2.628005657708628,
+      "grad_norm": 1.235780954360962,
+      "learning_rate": 1.5014016621935127e-05,
+      "loss": 0.2349,
+      "step": 930
+    },
+    {
+      "epoch": 2.6421499292786423,
+      "grad_norm": 1.4623289108276367,
+      "learning_rate": 1.4873851871264731e-05,
+      "loss": 0.2528,
+      "step": 935
+    },
+    {
+      "epoch": 2.6562942008486563,
+      "grad_norm": 1.3628381490707397,
+      "learning_rate": 1.4733698135530376e-05,
+      "loss": 0.2397,
+      "step": 940
+    },
+    {
+      "epoch": 2.6704384724186703,
+      "grad_norm": 1.3325926065444946,
+      "learning_rate": 1.4593567652602331e-05,
+      "loss": 0.2267,
+      "step": 945
+    },
+    {
+      "epoch": 2.6845827439886847,
+      "grad_norm": 1.1442891359329224,
+      "learning_rate": 1.44534726583205e-05,
+      "loss": 0.2417,
+      "step": 950
+    },
+    {
+      "epoch": 2.6987270155586986,
+      "grad_norm": 1.4495691061019897,
+      "learning_rate": 1.4313425385426003e-05,
+      "loss": 0.2246,
+      "step": 955
+    },
+    {
+      "epoch": 2.7128712871287126,
+      "grad_norm": 1.4209120273590088,
+      "learning_rate": 1.417343806249304e-05,
+      "loss": 0.2084,
+      "step": 960
+    },
+    {
+      "epoch": 2.727015558698727,
+      "grad_norm": 1.1167337894439697,
+      "learning_rate": 1.403352291286114e-05,
+      "loss": 0.2375,
+      "step": 965
+    },
+    {
+      "epoch": 2.741159830268741,
+      "grad_norm": 1.3341835737228394,
+      "learning_rate": 1.3893692153567846e-05,
+      "loss": 0.2252,
+      "step": 970
+    },
+    {
+      "epoch": 2.7553041018387554,
+      "grad_norm": 1.3639682531356812,
+      "learning_rate": 1.3753957994281944e-05,
+      "loss": 0.2232,
+      "step": 975
+    },
+    {
+      "epoch": 2.7694483734087694,
+      "grad_norm": 1.3704020977020264,
+      "learning_rate": 1.3614332636237352e-05,
+      "loss": 0.2411,
+      "step": 980
+    },
+    {
+      "epoch": 2.783592644978784,
+      "grad_norm": 1.1877855062484741,
+      "learning_rate": 1.3474828271167735e-05,
+      "loss": 0.2096,
+      "step": 985
+    },
+    {
+      "epoch": 2.7977369165487977,
+      "grad_norm": 1.3705607652664185,
+      "learning_rate": 1.3335457080241966e-05,
+      "loss": 0.2333,
+      "step": 990
+    },
+    {
+      "epoch": 2.8118811881188117,
+      "grad_norm": 1.2100168466567993,
+      "learning_rate": 1.319623123300048e-05,
+      "loss": 0.2126,
+      "step": 995
+    },
+    {
+      "epoch": 2.826025459688826,
+      "grad_norm": 1.3423547744750977,
+      "learning_rate": 1.305716288629267e-05,
+      "loss": 0.2046,
+      "step": 1000
+    },
+    {
+      "epoch": 2.84016973125884,
+      "grad_norm": 1.3659006357192993,
+      "learning_rate": 1.2918264183215384e-05,
+      "loss": 0.1937,
+      "step": 1005
+    },
+    {
+      "epoch": 2.854314002828854,
+      "grad_norm": 1.1460381746292114,
+      "learning_rate": 1.2779547252052611e-05,
+      "loss": 0.2332,
+      "step": 1010
+    },
+    {
+      "epoch": 2.8684582743988685,
+      "grad_norm": 1.2629969120025635,
+      "learning_rate": 1.2641024205216478e-05,
+      "loss": 0.2037,
+      "step": 1015
+    },
+    {
+      "epoch": 2.882602545968883,
+      "grad_norm": 1.171892762184143,
+      "learning_rate": 1.250270713818963e-05,
+      "loss": 0.2328,
+      "step": 1020
+    },
+    {
+      "epoch": 2.896746817538897,
+      "grad_norm": 1.2251631021499634,
+      "learning_rate": 1.2364608128469059e-05,
+      "loss": 0.2081,
+      "step": 1025
+    },
+    {
+      "epoch": 2.910891089108911,
+      "grad_norm": 1.1573634147644043,
+      "learning_rate": 1.2226739234511554e-05,
+      "loss": 0.226,
+      "step": 1030
+    },
+    {
+      "epoch": 2.9250353606789252,
+      "grad_norm": 1.232812762260437,
+      "learning_rate": 1.2089112494680775e-05,
+      "loss": 0.2094,
+      "step": 1035
+    },
+    {
+      "epoch": 2.939179632248939,
+      "grad_norm": 1.2418936491012573,
+      "learning_rate": 1.1951739926196097e-05,
+      "loss": 0.2016,
+      "step": 1040
+    },
+    {
+      "epoch": 2.953323903818953,
+      "grad_norm": 1.1464600563049316,
+      "learning_rate": 1.1814633524083292e-05,
+      "loss": 0.2109,
+      "step": 1045
+    },
+    {
+      "epoch": 2.9674681753889676,
+      "grad_norm": 1.307362675666809,
+      "learning_rate": 1.167780526012716e-05,
+      "loss": 0.205,
+      "step": 1050
+    },
+    {
+      "epoch": 2.9816124469589815,
+      "grad_norm": 1.443464756011963,
+      "learning_rate": 1.1541267081826172e-05,
+      "loss": 0.1853,
+      "step": 1055
+    },
+    {
+      "epoch": 2.9957567185289955,
+      "grad_norm": 1.1249037981033325,
+      "learning_rate": 1.1405030911349266e-05,
+      "loss": 0.1821,
+      "step": 1060
+    },
+    {
+      "epoch": 3.0084865629420086,
+      "grad_norm": 1.3226265907287598,
+      "learning_rate": 1.1269108644494802e-05,
+      "loss": 0.1554,
+      "step": 1065
+    },
+    {
+      "epoch": 3.0226308345120225,
+      "grad_norm": 1.1344726085662842,
+      "learning_rate": 1.1133512149651881e-05,
+      "loss": 0.1485,
+      "step": 1070
+    },
+    {
+      "epoch": 3.036775106082037,
+      "grad_norm": 1.2142800092697144,
+      "learning_rate": 1.0998253266764008e-05,
+      "loss": 0.1832,
+      "step": 1075
+    },
+    {
+      "epoch": 3.050919377652051,
+      "grad_norm": 1.09934401512146,
+      "learning_rate": 1.0863343806295273e-05,
+      "loss": 0.1403,
+      "step": 1080
+    },
+    {
+      "epoch": 3.065063649222065,
+      "grad_norm": 1.1893819570541382,
+      "learning_rate": 1.0728795548199076e-05,
+      "loss": 0.1426,
+      "step": 1085
+    },
+    {
+      "epoch": 3.0792079207920793,
+      "grad_norm": 1.27206552028656,
+      "learning_rate": 1.0594620240889538e-05,
+      "loss": 0.169,
+      "step": 1090
+    },
+    {
+      "epoch": 3.0933521923620932,
+      "grad_norm": 1.2133691310882568,
+      "learning_rate": 1.0460829600215658e-05,
+      "loss": 0.1455,
+      "step": 1095
+    },
+    {
+      "epoch": 3.1074964639321077,
+      "grad_norm": 1.230672001838684,
+      "learning_rate": 1.0327435308438323e-05,
+      "loss": 0.1519,
+      "step": 1100
+    },
+    {
+      "epoch": 3.1216407355021216,
+      "grad_norm": 1.060921549797058,
+      "learning_rate": 1.0194449013210225e-05,
+      "loss": 0.1408,
+      "step": 1105
+    },
+    {
+      "epoch": 3.1357850070721356,
+      "grad_norm": 1.232483983039856,
+      "learning_rate": 1.0061882326558835e-05,
+      "loss": 0.1527,
+      "step": 1110
+    },
+    {
+      "epoch": 3.14992927864215,
+      "grad_norm": 1.1834524869918823,
+      "learning_rate": 9.929746823872461e-06,
+      "loss": 0.1693,
+      "step": 1115
+    },
+    {
+      "epoch": 3.164073550212164,
+      "grad_norm": 1.115011215209961,
+      "learning_rate": 9.798054042889518e-06,
+      "loss": 0.1625,
+      "step": 1120
+    },
+    {
+      "epoch": 3.1782178217821784,
+      "grad_norm": 1.1622068881988525,
+      "learning_rate": 9.666815482691074e-06,
+      "loss": 0.1489,
+      "step": 1125
+    },
+    {
+      "epoch": 3.1923620933521923,
+      "grad_norm": 1.044087529182434,
+      "learning_rate": 9.536042602696791e-06,
+      "loss": 0.1447,
+      "step": 1130
+    },
+    {
+      "epoch": 3.2065063649222063,
+      "grad_norm": 1.3123761415481567,
+      "learning_rate": 9.40574682166431e-06,
+      "loss": 0.1541,
+      "step": 1135
+    },
+    {
+      "epoch": 3.2206506364922207,
+      "grad_norm": 1.2036011219024658,
+      "learning_rate": 9.275939516692192e-06,
+      "loss": 0.1412,
+      "step": 1140
+    },
+    {
+      "epoch": 3.2347949080622347,
+      "grad_norm": 1.1603245735168457,
+      "learning_rate": 9.14663202222651e-06,
+      "loss": 0.1518,
+      "step": 1145
+    },
+    {
+      "epoch": 3.248939179632249,
+      "grad_norm": 1.1296154260635376,
+      "learning_rate": 9.017835629071134e-06,
+      "loss": 0.13,
+      "step": 1150
+    },
+    {
+      "epoch": 3.263083451202263,
+      "grad_norm": 1.2457711696624756,
+      "learning_rate": 8.889561583401868e-06,
+      "loss": 0.1458,
+      "step": 1155
+    },
+    {
+      "epoch": 3.2772277227722775,
+      "grad_norm": 1.111977219581604,
+      "learning_rate": 8.761821085784457e-06,
+      "loss": 0.1319,
+      "step": 1160
+    },
+    {
+      "epoch": 3.2913719943422914,
+      "grad_norm": 1.1754082441329956,
+      "learning_rate": 8.634625290196554e-06,
+      "loss": 0.1418,
+      "step": 1165
+    },
+    {
+      "epoch": 3.3055162659123054,
+      "grad_norm": 1.1999913454055786,
+      "learning_rate": 8.50798530305383e-06,
+      "loss": 0.1622,
+      "step": 1170
+    },
+    {
+      "epoch": 3.31966053748232,
+      "grad_norm": 1.2002378702163696,
+      "learning_rate": 8.381912182240161e-06,
+      "loss": 0.1534,
+      "step": 1175
+    },
+    {
+      "epoch": 3.333804809052334,
+      "grad_norm": 1.193869948387146,
+      "learning_rate": 8.256416936142079e-06,
+      "loss": 0.1523,
+      "step": 1180
+    },
+    {
+      "epoch": 3.3479490806223478,
+      "grad_norm": 1.0911682844161987,
+      "learning_rate": 8.131510522687568e-06,
+      "loss": 0.1328,
+      "step": 1185
+    },
+    {
+      "epoch": 3.362093352192362,
+      "grad_norm": 1.1066834926605225,
+      "learning_rate": 8.00720384838922e-06,
+      "loss": 0.1339,
+      "step": 1190
+    },
+    {
+      "epoch": 3.376237623762376,
+      "grad_norm": 1.225243330001831,
+      "learning_rate": 7.883507767391915e-06,
+      "loss": 0.1312,
+      "step": 1195
+    },
+    {
+      "epoch": 3.3903818953323905,
+      "grad_norm": 0.943169116973877,
+      "learning_rate": 7.760433080525091e-06,
+      "loss": 0.1402,
+      "step": 1200
+    },
+    {
+      "epoch": 3.4045261669024045,
+      "grad_norm": 1.2137426137924194,
+      "learning_rate": 7.637990534359591e-06,
+      "loss": 0.138,
+      "step": 1205
+    },
+    {
+      "epoch": 3.418670438472419,
+      "grad_norm": 1.1949723958969116,
+      "learning_rate": 7.516190820269354e-06,
+      "loss": 0.138,
+      "step": 1210
+    },
+    {
+      "epoch": 3.432814710042433,
+      "grad_norm": 1.0119128227233887,
+      "learning_rate": 7.395044573497823e-06,
+      "loss": 0.1479,
+      "step": 1215
+    },
+    {
+      "epoch": 3.446958981612447,
+      "grad_norm": 1.2574715614318848,
+      "learning_rate": 7.274562372229344e-06,
+      "loss": 0.1241,
+      "step": 1220
+    },
+    {
+      "epoch": 3.4611032531824613,
+      "grad_norm": 1.2767034769058228,
+      "learning_rate": 7.154754736665471e-06,
+      "loss": 0.1436,
+      "step": 1225
+    },
+    {
+      "epoch": 3.4752475247524752,
+      "grad_norm": 1.1389586925506592,
+      "learning_rate": 7.035632128106406e-06,
+      "loss": 0.1468,
+      "step": 1230
+    },
+    {
+      "epoch": 3.489391796322489,
+      "grad_norm": 1.0416852235794067,
+      "learning_rate": 6.917204948037484e-06,
+      "loss": 0.136,
+      "step": 1235
+    },
+    {
+      "epoch": 3.5035360678925036,
+      "grad_norm": 1.3335916996002197,
+      "learning_rate": 6.799483537221022e-06,
+      "loss": 0.1505,
+      "step": 1240
+    },
+    {
+      "epoch": 3.5176803394625176,
+      "grad_norm": 1.1645817756652832,
+      "learning_rate": 6.6824781747933215e-06,
+      "loss": 0.1472,
+      "step": 1245
+    },
+    {
+      "epoch": 3.531824611032532,
+      "grad_norm": 1.2987821102142334,
+      "learning_rate": 6.566199077367166e-06,
+      "loss": 0.1319,
+      "step": 1250
+    },
+    {
+      "epoch": 3.545968882602546,
+      "grad_norm": 1.166772484779358,
+      "learning_rate": 6.450656398139691e-06,
+      "loss": 0.1315,
+      "step": 1255
+    },
+    {
+      "epoch": 3.5601131541725604,
+      "grad_norm": 1.2419538497924805,
+      "learning_rate": 6.335860226005879e-06,
+      "loss": 0.1107,
+      "step": 1260
+    },
+    {
+      "epoch": 3.5742574257425743,
+      "grad_norm": 1.0474148988723755,
+      "learning_rate": 6.221820584677575e-06,
+      "loss": 0.1175,
+      "step": 1265
+    },
+    {
+      "epoch": 3.5884016973125883,
+      "grad_norm": 1.244253158569336,
+      "learning_rate": 6.108547431808286e-06,
+      "loss": 0.1237,
+      "step": 1270
+    },
+    {
+      "epoch": 3.6025459688826027,
+      "grad_norm": 1.1174659729003906,
+      "learning_rate": 5.996050658123662e-06,
+      "loss": 0.141,
+      "step": 1275
+    },
+    {
+      "epoch": 3.6166902404526167,
+      "grad_norm": 1.0612963438034058,
+      "learning_rate": 5.884340086557901e-06,
+      "loss": 0.1126,
+      "step": 1280
+    },
+    {
+      "epoch": 3.6308345120226306,
+      "grad_norm": 1.4294148683547974,
+      "learning_rate": 5.7734254713960025e-06,
+      "loss": 0.1193,
+      "step": 1285
+    },
+    {
+      "epoch": 3.644978783592645,
+      "grad_norm": 1.0756280422210693,
+      "learning_rate": 5.663316497422076e-06,
+      "loss": 0.1207,
+      "step": 1290
+    },
+    {
+      "epoch": 3.659123055162659,
+      "grad_norm": 1.0577157735824585,
+      "learning_rate": 5.554022779073665e-06,
+      "loss": 0.1177,
+      "step": 1295
+    },
+    {
+      "epoch": 3.6732673267326734,
+      "grad_norm": 1.1960054636001587,
+      "learning_rate": 5.445553859602268e-06,
+      "loss": 0.1273,
+      "step": 1300
+    },
+    {
+      "epoch": 3.6874115983026874,
+      "grad_norm": 1.262784481048584,
+      "learning_rate": 5.3379192102399985e-06,
+      "loss": 0.1197,
+      "step": 1305
+    },
+    {
+      "epoch": 3.701555869872702,
+      "grad_norm": 0.9401896595954895,
+      "learning_rate": 5.231128229372644e-06,
+      "loss": 0.1177,
+      "step": 1310
+    },
+    {
+      "epoch": 3.7157001414427158,
+      "grad_norm": 1.3273756504058838,
+      "learning_rate": 5.125190241718962e-06,
+      "loss": 0.1372,
+      "step": 1315
+    },
+    {
+      "epoch": 3.7298444130127297,
+      "grad_norm": 1.0845891237258911,
+      "learning_rate": 5.0201144975165215e-06,
+      "loss": 0.1311,
+      "step": 1320
+    },
+    {
+      "epoch": 3.743988684582744,
+      "grad_norm": 1.0286343097686768,
+      "learning_rate": 4.915910171713948e-06,
+      "loss": 0.1222,
+      "step": 1325
+    },
+    {
+      "epoch": 3.758132956152758,
+      "grad_norm": 1.0829781293869019,
+      "learning_rate": 4.812586363169835e-06,
+      "loss": 0.1206,
+      "step": 1330
+    },
+    {
+      "epoch": 3.772277227722772,
+      "grad_norm": 1.1519546508789062,
+      "learning_rate": 4.710152093858209e-06,
+      "loss": 0.1154,
+      "step": 1335
+    },
+    {
+      "epoch": 3.7864214992927865,
+      "grad_norm": 1.09241783618927,
+      "learning_rate": 4.6086163080808e-06,
+      "loss": 0.129,
+      "step": 1340
+    },
+    {
+      "epoch": 3.8005657708628005,
+      "grad_norm": 1.1182514429092407,
+      "learning_rate": 4.507987871686001e-06,
+      "loss": 0.1167,
+      "step": 1345
+    },
+    {
+      "epoch": 3.814710042432815,
+      "grad_norm": 1.0947293043136597,
+      "learning_rate": 4.408275571294776e-06,
+      "loss": 0.1191,
+      "step": 1350
+    },
+    {
+      "epoch": 3.828854314002829,
+      "grad_norm": 1.1459898948669434,
+      "learning_rate": 4.309488113533382e-06,
+      "loss": 0.1277,
+      "step": 1355
+    },
+    {
+      "epoch": 3.8429985855728432,
+      "grad_norm": 0.8812487721443176,
+      "learning_rate": 4.211634124273162e-06,
+      "loss": 0.1149,
+      "step": 1360
+    },
+    {
+      "epoch": 3.857142857142857,
+      "grad_norm": 1.1296443939208984,
+      "learning_rate": 4.114722147877353e-06,
+      "loss": 0.1187,
+      "step": 1365
+    },
+    {
+      "epoch": 3.871287128712871,
+      "grad_norm": 1.1804019212722778,
+      "learning_rate": 4.018760646455e-06,
+      "loss": 0.1203,
+      "step": 1370
+    },
+    {
+      "epoch": 3.8854314002828856,
+      "grad_norm": 1.1915318965911865,
+      "learning_rate": 3.92375799912207e-06,
+      "loss": 0.1198,
+      "step": 1375
+    },
+    {
+      "epoch": 3.8995756718528995,
+      "grad_norm": 1.1128413677215576,
+      "learning_rate": 3.829722501269828e-06,
+      "loss": 0.1257,
+      "step": 1380
+    },
+    {
+      "epoch": 3.9137199434229135,
+      "grad_norm": 0.9256439805030823,
+      "learning_rate": 3.736662363840469e-06,
+      "loss": 0.1179,
+      "step": 1385
+    },
+    {
+      "epoch": 3.927864214992928,
+      "grad_norm": 0.9898888468742371,
+      "learning_rate": 3.644585712610198e-06,
+      "loss": 0.1311,
+      "step": 1390
+    },
+    {
+      "epoch": 3.942008486562942,
+      "grad_norm": 0.9239696264266968,
+      "learning_rate": 3.553500587479673e-06,
+      "loss": 0.1099,
+      "step": 1395
+    },
+    {
+      "epoch": 3.9561527581329563,
+      "grad_norm": 1.023583173751831,
+      "learning_rate": 3.4634149417720007e-06,
+      "loss": 0.1099,
+      "step": 1400
+    },
+    {
+      "epoch": 3.9702970297029703,
+      "grad_norm": 0.992881178855896,
+      "learning_rate": 3.3743366415382804e-06,
+      "loss": 0.1168,
+      "step": 1405
+    },
+    {
+      "epoch": 3.9844413012729847,
+      "grad_norm": 0.9432424902915955,
+      "learning_rate": 3.2862734648707275e-06,
+      "loss": 0.1149,
+      "step": 1410
+    },
+    {
+      "epoch": 3.9985855728429986,
+      "grad_norm": 0.9783270359039307,
+      "learning_rate": 3.1992331012235527e-06,
+      "loss": 0.1263,
+      "step": 1415
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1770,
+  "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": 1.9610793816568627e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

40_128_e5_3e-5/checkpoint-1416/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8f1a562a572af6f17e99424943e6ee2738b5feac80076391e8a179272d29064
+size 7736

40_128_e5_3e-5/checkpoint-1416/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)

40_128_e5_3e-5/checkpoint-1770/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

40_128_e5_3e-5/checkpoint-1770/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": [
+    "down_proj",
+    "o_proj",
+    "v_proj",
+    "up_proj",
+    "gate_proj",
+    "k_proj",
+    "q_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

40_128_e5_3e-5/checkpoint-1770/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85e19a3eb92feeaffa1834f631d214ca040e008f69604fe174c55686cb372c9c
+size 791751704

40_128_e5_3e-5/checkpoint-1770/latest

@@ -0,0 +1 @@
+global_step1770

40_128_e5_3e-5/checkpoint-1770/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08f11608a13cdf43aea57788fce7932b8cfece6f86cabb13758cdcc5174d4ffc
+size 15920

40_128_e5_3e-5/checkpoint-1770/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b277102fb96636958dae754efca533a818f66db0e6815e93e3dd49084706e969
+size 15920

40_128_e5_3e-5/checkpoint-1770/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14bfe7280db56d63963f7ff20353f56b7d0c31e87f2357ec2377826a772a7b5e
+size 15920