Upload folder using huggingface_hub

37_128_e5_3e-5/checkpoint-1329/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

37_128_e5_3e-5/checkpoint-1329/adapter_config.json

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

37_128_e5_3e-5/checkpoint-1329/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fa02e140a6d29872300c4842b771b461cb2deabd5bd16621a068da0869ff8007
+size 791751704

37_128_e5_3e-5/checkpoint-1329/latest

@@ -0,0 +1 @@
+global_step1329

37_128_e5_3e-5/checkpoint-1329/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b782513af58ec0c496944ec2da95c5de610c3c4974f8a10f9602fabfaa41dd1
+size 15920

37_128_e5_3e-5/checkpoint-1329/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45721e71c69abd23138455012f54240dd75bbdfc42821a583f69ff32cc597a21
+size 15920

37_128_e5_3e-5/checkpoint-1329/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a8a4b42ab895dc49bcd2c6d5c073defa2a30fa2c00f1618785e0ea0a07339c1
+size 15920

37_128_e5_3e-5/checkpoint-1329/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4c789c8616d7307e0dac8599a43ba0239473ff520f9bef35ee488ea58ac2a76c
+size 15920

37_128_e5_3e-5/checkpoint-1329/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a039eaaa69093e3c7df6fbc4e0cb8be7ea03e36e4899ce8d7c4231d1ddc024a
+size 15920

37_128_e5_3e-5/checkpoint-1329/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86069e5818475d2f0d699396d51028e6dd7f30939b94699108b59b527e4d9a0d
+size 15920

37_128_e5_3e-5/checkpoint-1329/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fda23f281762b2712033cb33f0599c707e815346395483594ef0530798763f2
+size 15920

37_128_e5_3e-5/checkpoint-1329/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d7965737fa6e70a391dc91aeac14c1f8a8e9314ed96ad0d5b73f6a5125cdce99
+size 15920

37_128_e5_3e-5/checkpoint-1329/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1220e8cb5d028ec3c04136794e879c499d589e396f16edcdf6a4c1c2bb0079e6
+size 1064

37_128_e5_3e-5/checkpoint-1329/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
+  }
+}

37_128_e5_3e-5/checkpoint-1329/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
+}

37_128_e5_3e-5/checkpoint-1329/trainer_state.json

@@ -0,0 +1,1889 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1329,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011299435028248588,
+      "grad_norm": 1.2254456281661987,
+      "learning_rate": 1.0810810810810812e-06,
+      "loss": 1.3157,
+      "step": 5
+    },
+    {
+      "epoch": 0.022598870056497175,
+      "grad_norm": 1.2222959995269775,
+      "learning_rate": 2.4324324324324325e-06,
+      "loss": 1.3831,
+      "step": 10
+    },
+    {
+      "epoch": 0.03389830508474576,
+      "grad_norm": 0.7486730813980103,
+      "learning_rate": 3.7837837837837835e-06,
+      "loss": 1.3448,
+      "step": 15
+    },
+    {
+      "epoch": 0.04519774011299435,
+      "grad_norm": 0.7303342819213867,
+      "learning_rate": 5.135135135135135e-06,
+      "loss": 1.2883,
+      "step": 20
+    },
+    {
+      "epoch": 0.05649717514124294,
+      "grad_norm": 0.6827901005744934,
+      "learning_rate": 6.486486486486487e-06,
+      "loss": 1.2782,
+      "step": 25
+    },
+    {
+      "epoch": 0.06779661016949153,
+      "grad_norm": 0.6475429534912109,
+      "learning_rate": 7.837837837837838e-06,
+      "loss": 1.2937,
+      "step": 30
+    },
+    {
+      "epoch": 0.07909604519774012,
+      "grad_norm": 0.5939517617225647,
+      "learning_rate": 9.18918918918919e-06,
+      "loss": 1.2553,
+      "step": 35
+    },
+    {
+      "epoch": 0.0903954802259887,
+      "grad_norm": 0.5064879655838013,
+      "learning_rate": 1.0540540540540541e-05,
+      "loss": 1.2067,
+      "step": 40
+    },
+    {
+      "epoch": 0.1016949152542373,
+      "grad_norm": 0.6584354639053345,
+      "learning_rate": 1.1891891891891893e-05,
+      "loss": 1.2246,
+      "step": 45
+    },
+    {
+      "epoch": 0.11299435028248588,
+      "grad_norm": 0.4975661337375641,
+      "learning_rate": 1.3243243243243242e-05,
+      "loss": 1.2558,
+      "step": 50
+    },
+    {
+      "epoch": 0.12429378531073447,
+      "grad_norm": 0.5031336545944214,
+      "learning_rate": 1.4594594594594596e-05,
+      "loss": 1.2726,
+      "step": 55
+    },
+    {
+      "epoch": 0.13559322033898305,
+      "grad_norm": 0.4471903443336487,
+      "learning_rate": 1.5945945945945947e-05,
+      "loss": 1.2089,
+      "step": 60
+    },
+    {
+      "epoch": 0.14689265536723164,
+      "grad_norm": 0.5343024134635925,
+      "learning_rate": 1.72972972972973e-05,
+      "loss": 1.2018,
+      "step": 65
+    },
+    {
+      "epoch": 0.15819209039548024,
+      "grad_norm": 0.47138091921806335,
+      "learning_rate": 1.864864864864865e-05,
+      "loss": 1.2826,
+      "step": 70
+    },
+    {
+      "epoch": 0.1694915254237288,
+      "grad_norm": 0.490500271320343,
+      "learning_rate": 1.9999999999999998e-05,
+      "loss": 1.2116,
+      "step": 75
+    },
+    {
+      "epoch": 0.1807909604519774,
+      "grad_norm": 0.5078205466270447,
+      "learning_rate": 2.135135135135135e-05,
+      "loss": 1.1873,
+      "step": 80
+    },
+    {
+      "epoch": 0.192090395480226,
+      "grad_norm": 1.9143681526184082,
+      "learning_rate": 2.2702702702702705e-05,
+      "loss": 1.2159,
+      "step": 85
+    },
+    {
+      "epoch": 0.2033898305084746,
+      "grad_norm": 0.5685150027275085,
+      "learning_rate": 2.4054054054054056e-05,
+      "loss": 1.1666,
+      "step": 90
+    },
+    {
+      "epoch": 0.21468926553672316,
+      "grad_norm": 0.5490728616714478,
+      "learning_rate": 2.5405405405405408e-05,
+      "loss": 1.1348,
+      "step": 95
+    },
+    {
+      "epoch": 0.22598870056497175,
+      "grad_norm": 0.5205124020576477,
+      "learning_rate": 2.6756756756756756e-05,
+      "loss": 1.1213,
+      "step": 100
+    },
+    {
+      "epoch": 0.23728813559322035,
+      "grad_norm": 0.4488866925239563,
+      "learning_rate": 2.8108108108108107e-05,
+      "loss": 1.1414,
+      "step": 105
+    },
+    {
+      "epoch": 0.24858757062146894,
+      "grad_norm": 0.4913260340690613,
+      "learning_rate": 2.945945945945946e-05,
+      "loss": 1.139,
+      "step": 110
+    },
+    {
+      "epoch": 0.2598870056497175,
+      "grad_norm": 0.5297622084617615,
+      "learning_rate": 2.9999849508725208e-05,
+      "loss": 1.134,
+      "step": 115
+    },
+    {
+      "epoch": 0.2711864406779661,
+      "grad_norm": 0.6094189882278442,
+      "learning_rate": 2.9998929850759165e-05,
+      "loss": 1.1023,
+      "step": 120
+    },
+    {
+      "epoch": 0.2824858757062147,
+      "grad_norm": 0.6322339177131653,
+      "learning_rate": 2.9997174192288006e-05,
+      "loss": 1.1164,
+      "step": 125
+    },
+    {
+      "epoch": 0.2937853107344633,
+      "grad_norm": 0.6201522946357727,
+      "learning_rate": 2.999458263116747e-05,
+      "loss": 1.1188,
+      "step": 130
+    },
+    {
+      "epoch": 0.3050847457627119,
+      "grad_norm": 0.5387688875198364,
+      "learning_rate": 2.9991155311844292e-05,
+      "loss": 1.0651,
+      "step": 135
+    },
+    {
+      "epoch": 0.3163841807909605,
+      "grad_norm": 0.5230715274810791,
+      "learning_rate": 2.998689242534815e-05,
+      "loss": 1.0626,
+      "step": 140
+    },
+    {
+      "epoch": 0.327683615819209,
+      "grad_norm": 0.5703102946281433,
+      "learning_rate": 2.9981794209281003e-05,
+      "loss": 1.0626,
+      "step": 145
+    },
+    {
+      "epoch": 0.3389830508474576,
+      "grad_norm": 0.5561928153038025,
+      "learning_rate": 2.997586094780388e-05,
+      "loss": 1.0494,
+      "step": 150
+    },
+    {
+      "epoch": 0.3502824858757062,
+      "grad_norm": 0.6391568183898926,
+      "learning_rate": 2.9969092971621006e-05,
+      "loss": 1.05,
+      "step": 155
+    },
+    {
+      "epoch": 0.3615819209039548,
+      "grad_norm": 0.6165516376495361,
+      "learning_rate": 2.9961490657961404e-05,
+      "loss": 1.0617,
+      "step": 160
+    },
+    {
+      "epoch": 0.3728813559322034,
+      "grad_norm": 0.6393241286277771,
+      "learning_rate": 2.9953054430557835e-05,
+      "loss": 1.0036,
+      "step": 165
+    },
+    {
+      "epoch": 0.384180790960452,
+      "grad_norm": 0.6466289758682251,
+      "learning_rate": 2.9943784759623205e-05,
+      "loss": 1.0362,
+      "step": 170
+    },
+    {
+      "epoch": 0.3954802259887006,
+      "grad_norm": 0.7020736932754517,
+      "learning_rate": 2.993368216182435e-05,
+      "loss": 1.0052,
+      "step": 175
+    },
+    {
+      "epoch": 0.4067796610169492,
+      "grad_norm": 0.630094051361084,
+      "learning_rate": 2.992274720025323e-05,
+      "loss": 1.0279,
+      "step": 180
+    },
+    {
+      "epoch": 0.4180790960451977,
+      "grad_norm": 0.629647433757782,
+      "learning_rate": 2.9910980484395555e-05,
+      "loss": 1.0309,
+      "step": 185
+    },
+    {
+      "epoch": 0.4293785310734463,
+      "grad_norm": 0.6576105952262878,
+      "learning_rate": 2.989838267009681e-05,
+      "loss": 0.9227,
+      "step": 190
+    },
+    {
+      "epoch": 0.4406779661016949,
+      "grad_norm": 0.6078037023544312,
+      "learning_rate": 2.988495445952569e-05,
+      "loss": 1.0005,
+      "step": 195
+    },
+    {
+      "epoch": 0.4519774011299435,
+      "grad_norm": 0.657957911491394,
+      "learning_rate": 2.9870696601134994e-05,
+      "loss": 0.9715,
+      "step": 200
+    },
+    {
+      "epoch": 0.4632768361581921,
+      "grad_norm": 0.6783279180526733,
+      "learning_rate": 2.9855609889619864e-05,
+      "loss": 0.927,
+      "step": 205
+    },
+    {
+      "epoch": 0.4745762711864407,
+      "grad_norm": 0.6905434727668762,
+      "learning_rate": 2.983969516587352e-05,
+      "loss": 0.9506,
+      "step": 210
+    },
+    {
+      "epoch": 0.4858757062146893,
+      "grad_norm": 0.8056045174598694,
+      "learning_rate": 2.9822953316940393e-05,
+      "loss": 0.9867,
+      "step": 215
+    },
+    {
+      "epoch": 0.4971751412429379,
+      "grad_norm": 0.7994968295097351,
+      "learning_rate": 2.9805385275966658e-05,
+      "loss": 0.9489,
+      "step": 220
+    },
+    {
+      "epoch": 0.5084745762711864,
+      "grad_norm": 0.7940033078193665,
+      "learning_rate": 2.978699202214826e-05,
+      "loss": 0.9538,
+      "step": 225
+    },
+    {
+      "epoch": 0.519774011299435,
+      "grad_norm": 0.7576586604118347,
+      "learning_rate": 2.97677745806763e-05,
+      "loss": 0.9529,
+      "step": 230
+    },
+    {
+      "epoch": 0.5310734463276836,
+      "grad_norm": 0.8559620976448059,
+      "learning_rate": 2.9747734022679913e-05,
+      "loss": 0.9188,
+      "step": 235
+    },
+    {
+      "epoch": 0.5423728813559322,
+      "grad_norm": 0.9041480422019958,
+      "learning_rate": 2.9726871465166567e-05,
+      "loss": 0.9145,
+      "step": 240
+    },
+    {
+      "epoch": 0.5536723163841808,
+      "grad_norm": 0.7606709003448486,
+      "learning_rate": 2.970518807095979e-05,
+      "loss": 0.9179,
+      "step": 245
+    },
+    {
+      "epoch": 0.5649717514124294,
+      "grad_norm": 0.904378354549408,
+      "learning_rate": 2.9682685048634384e-05,
+      "loss": 0.8682,
+      "step": 250
+    },
+    {
+      "epoch": 0.576271186440678,
+      "grad_norm": 0.8670118451118469,
+      "learning_rate": 2.9659363652449027e-05,
+      "loss": 0.9448,
+      "step": 255
+    },
+    {
+      "epoch": 0.5875706214689266,
+      "grad_norm": 0.7508774399757385,
+      "learning_rate": 2.9635225182276387e-05,
+      "loss": 0.9223,
+      "step": 260
+    },
+    {
+      "epoch": 0.5988700564971752,
+      "grad_norm": 0.6919876933097839,
+      "learning_rate": 2.961027098353067e-05,
+      "loss": 0.8971,
+      "step": 265
+    },
+    {
+      "epoch": 0.6101694915254238,
+      "grad_norm": 0.7949540615081787,
+      "learning_rate": 2.9584502447092623e-05,
+      "loss": 0.91,
+      "step": 270
+    },
+    {
+      "epoch": 0.6214689265536724,
+      "grad_norm": 0.7679325938224792,
+      "learning_rate": 2.9557921009232013e-05,
+      "loss": 0.903,
+      "step": 275
+    },
+    {
+      "epoch": 0.632768361581921,
+      "grad_norm": 0.8364932537078857,
+      "learning_rate": 2.953052815152757e-05,
+      "loss": 0.8941,
+      "step": 280
+    },
+    {
+      "epoch": 0.6440677966101694,
+      "grad_norm": 0.8773370981216431,
+      "learning_rate": 2.950232540078442e-05,
+      "loss": 0.8812,
+      "step": 285
+    },
+    {
+      "epoch": 0.655367231638418,
+      "grad_norm": 1.0121004581451416,
+      "learning_rate": 2.947331432894896e-05,
+      "loss": 0.901,
+      "step": 290
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.8700821995735168,
+      "learning_rate": 2.9443496553021268e-05,
+      "loss": 0.8385,
+      "step": 295
+    },
+    {
+      "epoch": 0.6779661016949152,
+      "grad_norm": 0.8492361903190613,
+      "learning_rate": 2.9412873734964973e-05,
+      "loss": 0.8498,
+      "step": 300
+    },
+    {
+      "epoch": 0.6892655367231638,
+      "grad_norm": 0.860744833946228,
+      "learning_rate": 2.938144758161459e-05,
+      "loss": 0.8381,
+      "step": 305
+    },
+    {
+      "epoch": 0.7005649717514124,
+      "grad_norm": 0.8146252036094666,
+      "learning_rate": 2.934921984458043e-05,
+      "loss": 0.8505,
+      "step": 310
+    },
+    {
+      "epoch": 0.711864406779661,
+      "grad_norm": 0.828873336315155,
+      "learning_rate": 2.931619232015094e-05,
+      "loss": 0.8359,
+      "step": 315
+    },
+    {
+      "epoch": 0.7231638418079096,
+      "grad_norm": 0.8928766250610352,
+      "learning_rate": 2.9282366849192596e-05,
+      "loss": 0.8697,
+      "step": 320
+    },
+    {
+      "epoch": 0.7344632768361582,
+      "grad_norm": 0.8543341755867004,
+      "learning_rate": 2.924774531704729e-05,
+      "loss": 0.7961,
+      "step": 325
+    },
+    {
+      "epoch": 0.7457627118644068,
+      "grad_norm": 0.8456112146377563,
+      "learning_rate": 2.921232965342725e-05,
+      "loss": 0.8298,
+      "step": 330
+    },
+    {
+      "epoch": 0.7570621468926554,
+      "grad_norm": 0.8231430053710938,
+      "learning_rate": 2.9176121832307487e-05,
+      "loss": 0.7866,
+      "step": 335
+    },
+    {
+      "epoch": 0.768361581920904,
+      "grad_norm": 0.8562153577804565,
+      "learning_rate": 2.9139123871815762e-05,
+      "loss": 0.7833,
+      "step": 340
+    },
+    {
+      "epoch": 0.7796610169491526,
+      "grad_norm": 0.8552905917167664,
+      "learning_rate": 2.9101337834120113e-05,
+      "loss": 0.7777,
+      "step": 345
+    },
+    {
+      "epoch": 0.7909604519774012,
+      "grad_norm": 0.9859520792961121,
+      "learning_rate": 2.9062765825313887e-05,
+      "loss": 0.8016,
+      "step": 350
+    },
+    {
+      "epoch": 0.8022598870056498,
+      "grad_norm": 0.867014467716217,
+      "learning_rate": 2.9023409995298406e-05,
+      "loss": 0.7895,
+      "step": 355
+    },
+    {
+      "epoch": 0.8135593220338984,
+      "grad_norm": 0.890496015548706,
+      "learning_rate": 2.8983272537663082e-05,
+      "loss": 0.7795,
+      "step": 360
+    },
+    {
+      "epoch": 0.8248587570621468,
+      "grad_norm": 0.9742592573165894,
+      "learning_rate": 2.8942355689563166e-05,
+      "loss": 0.7661,
+      "step": 365
+    },
+    {
+      "epoch": 0.8361581920903954,
+      "grad_norm": 0.855208694934845,
+      "learning_rate": 2.890066173159509e-05,
+      "loss": 0.7863,
+      "step": 370
+    },
+    {
+      "epoch": 0.847457627118644,
+      "grad_norm": 0.9460350871086121,
+      "learning_rate": 2.8858192987669303e-05,
+      "loss": 0.7475,
+      "step": 375
+    },
+    {
+      "epoch": 0.8587570621468926,
+      "grad_norm": 0.9050624966621399,
+      "learning_rate": 2.881495182488077e-05,
+      "loss": 0.739,
+      "step": 380
+    },
+    {
+      "epoch": 0.8700564971751412,
+      "grad_norm": 0.8351825475692749,
+      "learning_rate": 2.8770940653377047e-05,
+      "loss": 0.7322,
+      "step": 385
+    },
+    {
+      "epoch": 0.8813559322033898,
+      "grad_norm": 0.9222843050956726,
+      "learning_rate": 2.8726161926223904e-05,
+      "loss": 0.7435,
+      "step": 390
+    },
+    {
+      "epoch": 0.8926553672316384,
+      "grad_norm": 0.8897649645805359,
+      "learning_rate": 2.8680618139268643e-05,
+      "loss": 0.7229,
+      "step": 395
+    },
+    {
+      "epoch": 0.903954802259887,
+      "grad_norm": 1.0437895059585571,
+      "learning_rate": 2.8634311831000966e-05,
+      "loss": 0.7132,
+      "step": 400
+    },
+    {
+      "epoch": 0.9152542372881356,
+      "grad_norm": 0.9872046113014221,
+      "learning_rate": 2.8587245582411475e-05,
+      "loss": 0.7257,
+      "step": 405
+    },
+    {
+      "epoch": 0.9265536723163842,
+      "grad_norm": 0.9081714749336243,
+      "learning_rate": 2.8539422016847834e-05,
+      "loss": 0.6956,
+      "step": 410
+    },
+    {
+      "epoch": 0.9378531073446328,
+      "grad_norm": 0.9939802289009094,
+      "learning_rate": 2.8490843799868556e-05,
+      "loss": 0.723,
+      "step": 415
+    },
+    {
+      "epoch": 0.9491525423728814,
+      "grad_norm": 0.8449594974517822,
+      "learning_rate": 2.8441513639094396e-05,
+      "loss": 0.7294,
+      "step": 420
+    },
+    {
+      "epoch": 0.96045197740113,
+      "grad_norm": 0.9042068719863892,
+      "learning_rate": 2.8391434284057466e-05,
+      "loss": 0.6565,
+      "step": 425
+    },
+    {
+      "epoch": 0.9717514124293786,
+      "grad_norm": 0.9338545203208923,
+      "learning_rate": 2.8340608526047995e-05,
+      "loss": 0.711,
+      "step": 430
+    },
+    {
+      "epoch": 0.9830508474576272,
+      "grad_norm": 0.9345859885215759,
+      "learning_rate": 2.82890391979587e-05,
+      "loss": 0.7006,
+      "step": 435
+    },
+    {
+      "epoch": 0.9943502824858758,
+      "grad_norm": 0.9804854393005371,
+      "learning_rate": 2.8236729174126948e-05,
+      "loss": 0.7183,
+      "step": 440
+    },
+    {
+      "epoch": 1.0045197740112994,
+      "grad_norm": 0.9410772919654846,
+      "learning_rate": 2.818368137017449e-05,
+      "loss": 0.6665,
+      "step": 445
+    },
+    {
+      "epoch": 1.015819209039548,
+      "grad_norm": 1.156546950340271,
+      "learning_rate": 2.8129898742845006e-05,
+      "loss": 0.616,
+      "step": 450
+    },
+    {
+      "epoch": 1.0271186440677966,
+      "grad_norm": 1.0281895399093628,
+      "learning_rate": 2.8075384289839274e-05,
+      "loss": 0.5688,
+      "step": 455
+    },
+    {
+      "epoch": 1.038418079096045,
+      "grad_norm": 0.9605177044868469,
+      "learning_rate": 2.8020141049648072e-05,
+      "loss": 0.6208,
+      "step": 460
+    },
+    {
+      "epoch": 1.0497175141242938,
+      "grad_norm": 1.0274258852005005,
+      "learning_rate": 2.796417210138287e-05,
+      "loss": 0.578,
+      "step": 465
+    },
+    {
+      "epoch": 1.0610169491525423,
+      "grad_norm": 1.0010178089141846,
+      "learning_rate": 2.7907480564604155e-05,
+      "loss": 0.5861,
+      "step": 470
+    },
+    {
+      "epoch": 1.072316384180791,
+      "grad_norm": 1.5158096551895142,
+      "learning_rate": 2.7850069599147584e-05,
+      "loss": 0.5705,
+      "step": 475
+    },
+    {
+      "epoch": 1.0836158192090395,
+      "grad_norm": 1.0114028453826904,
+      "learning_rate": 2.7791942404947857e-05,
+      "loss": 0.5903,
+      "step": 480
+    },
+    {
+      "epoch": 1.0949152542372882,
+      "grad_norm": 0.9214546084403992,
+      "learning_rate": 2.773310222186037e-05,
+      "loss": 0.5517,
+      "step": 485
+    },
+    {
+      "epoch": 1.1062146892655367,
+      "grad_norm": 1.062912106513977,
+      "learning_rate": 2.7673552329480615e-05,
+      "loss": 0.6101,
+      "step": 490
+    },
+    {
+      "epoch": 1.1175141242937854,
+      "grad_norm": 1.0093401670455933,
+      "learning_rate": 2.76132960469614e-05,
+      "loss": 0.6117,
+      "step": 495
+    },
+    {
+      "epoch": 1.1288135593220339,
+      "grad_norm": 1.0780502557754517,
+      "learning_rate": 2.7552336732827846e-05,
+      "loss": 0.5611,
+      "step": 500
+    },
+    {
+      "epoch": 1.1401129943502826,
+      "grad_norm": 1.2409229278564453,
+      "learning_rate": 2.749067778479017e-05,
+      "loss": 0.5852,
+      "step": 505
+    },
+    {
+      "epoch": 1.151412429378531,
+      "grad_norm": 1.0648524761199951,
+      "learning_rate": 2.742832263955435e-05,
+      "loss": 0.5707,
+      "step": 510
+    },
+    {
+      "epoch": 1.1627118644067798,
+      "grad_norm": 1.1011556386947632,
+      "learning_rate": 2.7365274772630535e-05,
+      "loss": 0.5322,
+      "step": 515
+    },
+    {
+      "epoch": 1.1740112994350282,
+      "grad_norm": 1.0446882247924805,
+      "learning_rate": 2.730153769813933e-05,
+      "loss": 0.5517,
+      "step": 520
+    },
+    {
+      "epoch": 1.1853107344632767,
+      "grad_norm": 1.1494063138961792,
+      "learning_rate": 2.723711496861596e-05,
+      "loss": 0.5803,
+      "step": 525
+    },
+    {
+      "epoch": 1.1966101694915254,
+      "grad_norm": 0.9970608949661255,
+      "learning_rate": 2.7172010174812226e-05,
+      "loss": 0.611,
+      "step": 530
+    },
+    {
+      "epoch": 1.207909604519774,
+      "grad_norm": 1.0200440883636475,
+      "learning_rate": 2.710622694549638e-05,
+      "loss": 0.5742,
+      "step": 535
+    },
+    {
+      "epoch": 1.2192090395480226,
+      "grad_norm": 1.040436863899231,
+      "learning_rate": 2.7039768947250874e-05,
+      "loss": 0.5131,
+      "step": 540
+    },
+    {
+      "epoch": 1.230508474576271,
+      "grad_norm": 0.9933497309684753,
+      "learning_rate": 2.697263988426798e-05,
+      "loss": 0.5384,
+      "step": 545
+    },
+    {
+      "epoch": 1.2418079096045198,
+      "grad_norm": 1.1182832717895508,
+      "learning_rate": 2.6904843498143344e-05,
+      "loss": 0.5149,
+      "step": 550
+    },
+    {
+      "epoch": 1.2531073446327683,
+      "grad_norm": 1.002934217453003,
+      "learning_rate": 2.6836383567667423e-05,
+      "loss": 0.5396,
+      "step": 555
+    },
+    {
+      "epoch": 1.264406779661017,
+      "grad_norm": 1.090778112411499,
+      "learning_rate": 2.6767263908614884e-05,
+      "loss": 0.5593,
+      "step": 560
+    },
+    {
+      "epoch": 1.2757062146892655,
+      "grad_norm": 1.0608378648757935,
+      "learning_rate": 2.6697488373531904e-05,
+      "loss": 0.523,
+      "step": 565
+    },
+    {
+      "epoch": 1.2870056497175142,
+      "grad_norm": 1.0161354541778564,
+      "learning_rate": 2.662706085152144e-05,
+      "loss": 0.56,
+      "step": 570
+    },
+    {
+      "epoch": 1.2983050847457627,
+      "grad_norm": 1.0922820568084717,
+      "learning_rate": 2.655598526802649e-05,
+      "loss": 0.518,
+      "step": 575
+    },
+    {
+      "epoch": 1.3096045197740114,
+      "grad_norm": 0.9997158050537109,
+      "learning_rate": 2.6484265584611258e-05,
+      "loss": 0.5293,
+      "step": 580
+    },
+    {
+      "epoch": 1.3209039548022599,
+      "grad_norm": 1.1040353775024414,
+      "learning_rate": 2.6411905798740374e-05,
+      "loss": 0.5622,
+      "step": 585
+    },
+    {
+      "epoch": 1.3322033898305086,
+      "grad_norm": 0.9653554558753967,
+      "learning_rate": 2.6338909943556083e-05,
+      "loss": 0.482,
+      "step": 590
+    },
+    {
+      "epoch": 1.343502824858757,
+      "grad_norm": 1.0184459686279297,
+      "learning_rate": 2.6265282087653427e-05,
+      "loss": 0.5402,
+      "step": 595
+    },
+    {
+      "epoch": 1.3548022598870055,
+      "grad_norm": 1.1181055307388306,
+      "learning_rate": 2.619102633485352e-05,
+      "loss": 0.5269,
+      "step": 600
+    },
+    {
+      "epoch": 1.3661016949152542,
+      "grad_norm": 1.0076478719711304,
+      "learning_rate": 2.6116146823974747e-05,
+      "loss": 0.496,
+      "step": 605
+    },
+    {
+      "epoch": 1.377401129943503,
+      "grad_norm": 1.2491661310195923,
+      "learning_rate": 2.604064772860213e-05,
+      "loss": 0.5157,
+      "step": 610
+    },
+    {
+      "epoch": 1.3887005649717514,
+      "grad_norm": 1.0203555822372437,
+      "learning_rate": 2.5964533256854686e-05,
+      "loss": 0.4646,
+      "step": 615
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 1.079098105430603,
+      "learning_rate": 2.5887807651150865e-05,
+      "loss": 0.497,
+      "step": 620
+    },
+    {
+      "epoch": 1.4112994350282486,
+      "grad_norm": 1.1061233282089233,
+      "learning_rate": 2.5810475187972113e-05,
+      "loss": 0.4781,
+      "step": 625
+    },
+    {
+      "epoch": 1.422598870056497,
+      "grad_norm": 1.256387710571289,
+      "learning_rate": 2.5732540177624487e-05,
+      "loss": 0.4991,
+      "step": 630
+    },
+    {
+      "epoch": 1.4338983050847458,
+      "grad_norm": 1.09014892578125,
+      "learning_rate": 2.565400696399843e-05,
+      "loss": 0.5003,
+      "step": 635
+    },
+    {
+      "epoch": 1.4451977401129943,
+      "grad_norm": 0.9970406889915466,
+      "learning_rate": 2.557487992432665e-05,
+      "loss": 0.4976,
+      "step": 640
+    },
+    {
+      "epoch": 1.456497175141243,
+      "grad_norm": 0.9552879333496094,
+      "learning_rate": 2.5495163468940135e-05,
+      "loss": 0.4523,
+      "step": 645
+    },
+    {
+      "epoch": 1.4677966101694915,
+      "grad_norm": 1.1708117723464966,
+      "learning_rate": 2.5414862041022336e-05,
+      "loss": 0.5089,
+      "step": 650
+    },
+    {
+      "epoch": 1.4790960451977402,
+      "grad_norm": 1.2075222730636597,
+      "learning_rate": 2.533398011636153e-05,
+      "loss": 0.5079,
+      "step": 655
+    },
+    {
+      "epoch": 1.4903954802259887,
+      "grad_norm": 1.1360461711883545,
+      "learning_rate": 2.5252522203101336e-05,
+      "loss": 0.453,
+      "step": 660
+    },
+    {
+      "epoch": 1.5016949152542374,
+      "grad_norm": 1.0806200504302979,
+      "learning_rate": 2.517049284148944e-05,
+      "loss": 0.454,
+      "step": 665
+    },
+    {
+      "epoch": 1.5129943502824859,
+      "grad_norm": 1.00311279296875,
+      "learning_rate": 2.5087896603624545e-05,
+      "loss": 0.4649,
+      "step": 670
+    },
+    {
+      "epoch": 1.5242937853107343,
+      "grad_norm": 1.0430840253829956,
+      "learning_rate": 2.5004738093201534e-05,
+      "loss": 0.4552,
+      "step": 675
+    },
+    {
+      "epoch": 1.535593220338983,
+      "grad_norm": 1.133973479270935,
+      "learning_rate": 2.4921021945254872e-05,
+      "loss": 0.5055,
+      "step": 680
+    },
+    {
+      "epoch": 1.5468926553672318,
+      "grad_norm": 1.0552384853363037,
+      "learning_rate": 2.483675282590025e-05,
+      "loss": 0.4504,
+      "step": 685
+    },
+    {
+      "epoch": 1.5581920903954802,
+      "grad_norm": 1.1162137985229492,
+      "learning_rate": 2.4751935432074513e-05,
+      "loss": 0.4305,
+      "step": 690
+    },
+    {
+      "epoch": 1.5694915254237287,
+      "grad_norm": 1.1118879318237305,
+      "learning_rate": 2.4666574491273882e-05,
+      "loss": 0.4413,
+      "step": 695
+    },
+    {
+      "epoch": 1.5807909604519774,
+      "grad_norm": 1.2221858501434326,
+      "learning_rate": 2.4580674761290433e-05,
+      "loss": 0.4225,
+      "step": 700
+    },
+    {
+      "epoch": 1.5920903954802261,
+      "grad_norm": 1.1549981832504272,
+      "learning_rate": 2.4494241029946927e-05,
+      "loss": 0.4359,
+      "step": 705
+    },
+    {
+      "epoch": 1.6033898305084746,
+      "grad_norm": 1.1755397319793701,
+      "learning_rate": 2.4407278114829925e-05,
+      "loss": 0.4477,
+      "step": 710
+    },
+    {
+      "epoch": 1.614689265536723,
+      "grad_norm": 1.1148909330368042,
+      "learning_rate": 2.431979086302131e-05,
+      "loss": 0.4564,
+      "step": 715
+    },
+    {
+      "epoch": 1.6259887005649718,
+      "grad_norm": 0.9955953359603882,
+      "learning_rate": 2.4231784150828087e-05,
+      "loss": 0.4501,
+      "step": 720
+    },
+    {
+      "epoch": 1.6372881355932203,
+      "grad_norm": 1.1515029668807983,
+      "learning_rate": 2.4143262883510593e-05,
+      "loss": 0.4771,
+      "step": 725
+    },
+    {
+      "epoch": 1.6485875706214688,
+      "grad_norm": 1.4518494606018066,
+      "learning_rate": 2.405423199500912e-05,
+      "loss": 0.4262,
+      "step": 730
+    },
+    {
+      "epoch": 1.6598870056497175,
+      "grad_norm": 0.9435175657272339,
+      "learning_rate": 2.3964696447668873e-05,
+      "loss": 0.4424,
+      "step": 735
+    },
+    {
+      "epoch": 1.6711864406779662,
+      "grad_norm": 1.2055212259292603,
+      "learning_rate": 2.3874661231963422e-05,
+      "loss": 0.4333,
+      "step": 740
+    },
+    {
+      "epoch": 1.6824858757062147,
+      "grad_norm": 1.2880557775497437,
+      "learning_rate": 2.3784131366216504e-05,
+      "loss": 0.452,
+      "step": 745
+    },
+    {
+      "epoch": 1.6937853107344631,
+      "grad_norm": 1.093306303024292,
+      "learning_rate": 2.369311189632234e-05,
+      "loss": 0.4526,
+      "step": 750
+    },
+    {
+      "epoch": 1.7050847457627119,
+      "grad_norm": 1.1268608570098877,
+      "learning_rate": 2.3601607895464387e-05,
+      "loss": 0.4542,
+      "step": 755
+    },
+    {
+      "epoch": 1.7163841807909606,
+      "grad_norm": 1.1216182708740234,
+      "learning_rate": 2.350962446383258e-05,
+      "loss": 0.4287,
+      "step": 760
+    },
+    {
+      "epoch": 1.727683615819209,
+      "grad_norm": 1.0930184125900269,
+      "learning_rate": 2.3417166728339034e-05,
+      "loss": 0.3796,
+      "step": 765
+    },
+    {
+      "epoch": 1.7389830508474575,
+      "grad_norm": 1.2358158826828003,
+      "learning_rate": 2.3324239842332324e-05,
+      "loss": 0.4139,
+      "step": 770
+    },
+    {
+      "epoch": 1.7502824858757062,
+      "grad_norm": 1.2803504467010498,
+      "learning_rate": 2.32308489853102e-05,
+      "loss": 0.3922,
+      "step": 775
+    },
+    {
+      "epoch": 1.761581920903955,
+      "grad_norm": 1.0053050518035889,
+      "learning_rate": 2.3136999362630945e-05,
+      "loss": 0.3947,
+      "step": 780
+    },
+    {
+      "epoch": 1.7728813559322034,
+      "grad_norm": 1.2856465578079224,
+      "learning_rate": 2.304269620522322e-05,
+      "loss": 0.3861,
+      "step": 785
+    },
+    {
+      "epoch": 1.784180790960452,
+      "grad_norm": 0.9932456612586975,
+      "learning_rate": 2.294794476929449e-05,
+      "loss": 0.4505,
+      "step": 790
+    },
+    {
+      "epoch": 1.7954802259887006,
+      "grad_norm": 1.2482821941375732,
+      "learning_rate": 2.28527503360381e-05,
+      "loss": 0.3839,
+      "step": 795
+    },
+    {
+      "epoch": 1.8067796610169493,
+      "grad_norm": 0.9733690023422241,
+      "learning_rate": 2.2757118211338863e-05,
+      "loss": 0.387,
+      "step": 800
+    },
+    {
+      "epoch": 1.8180790960451978,
+      "grad_norm": 1.034924030303955,
+      "learning_rate": 2.2661053725477374e-05,
+      "loss": 0.413,
+      "step": 805
+    },
+    {
+      "epoch": 1.8293785310734463,
+      "grad_norm": 1.0252270698547363,
+      "learning_rate": 2.256456223283287e-05,
+      "loss": 0.3787,
+      "step": 810
+    },
+    {
+      "epoch": 1.840677966101695,
+      "grad_norm": 1.2237545251846313,
+      "learning_rate": 2.2467649111584816e-05,
+      "loss": 0.3687,
+      "step": 815
+    },
+    {
+      "epoch": 1.8519774011299435,
+      "grad_norm": 1.0817660093307495,
+      "learning_rate": 2.2370319763413157e-05,
+      "loss": 0.3826,
+      "step": 820
+    },
+    {
+      "epoch": 1.863276836158192,
+      "grad_norm": 1.099799394607544,
+      "learning_rate": 2.2272579613197183e-05,
+      "loss": 0.4199,
+      "step": 825
+    },
+    {
+      "epoch": 1.8745762711864407,
+      "grad_norm": 1.1659348011016846,
+      "learning_rate": 2.217443410871324e-05,
+      "loss": 0.3739,
+      "step": 830
+    },
+    {
+      "epoch": 1.8858757062146894,
+      "grad_norm": 1.0663833618164062,
+      "learning_rate": 2.207588872033101e-05,
+      "loss": 0.4162,
+      "step": 835
+    },
+    {
+      "epoch": 1.8971751412429378,
+      "grad_norm": 1.1548362970352173,
+      "learning_rate": 2.197694894070866e-05,
+      "loss": 0.387,
+      "step": 840
+    },
+    {
+      "epoch": 1.9084745762711863,
+      "grad_norm": 1.2925057411193848,
+      "learning_rate": 2.187762028448667e-05,
+      "loss": 0.3588,
+      "step": 845
+    },
+    {
+      "epoch": 1.919774011299435,
+      "grad_norm": 1.1187832355499268,
+      "learning_rate": 2.177790828798047e-05,
+      "loss": 0.3681,
+      "step": 850
+    },
+    {
+      "epoch": 1.9310734463276837,
+      "grad_norm": 1.1352661848068237,
+      "learning_rate": 2.1677818508871862e-05,
+      "loss": 0.3755,
+      "step": 855
+    },
+    {
+      "epoch": 1.9423728813559322,
+      "grad_norm": 1.1208044290542603,
+      "learning_rate": 2.1577356525899244e-05,
+      "loss": 0.3914,
+      "step": 860
+    },
+    {
+      "epoch": 1.9536723163841807,
+      "grad_norm": 1.1486340761184692,
+      "learning_rate": 2.1476527938546665e-05,
+      "loss": 0.3681,
+      "step": 865
+    },
+    {
+      "epoch": 1.9649717514124294,
+      "grad_norm": 1.209509253501892,
+      "learning_rate": 2.1375338366731743e-05,
+      "loss": 0.3958,
+      "step": 870
+    },
+    {
+      "epoch": 1.9762711864406781,
+      "grad_norm": 1.1091386079788208,
+      "learning_rate": 2.127379345049239e-05,
+      "loss": 0.3282,
+      "step": 875
+    },
+    {
+      "epoch": 1.9875706214689266,
+      "grad_norm": 1.0583096742630005,
+      "learning_rate": 2.117189884967248e-05,
+      "loss": 0.3461,
+      "step": 880
+    },
+    {
+      "epoch": 1.998870056497175,
+      "grad_norm": 1.161463737487793,
+      "learning_rate": 2.1069660243606384e-05,
+      "loss": 0.3753,
+      "step": 885
+    },
+    {
+      "epoch": 2.009039548022599,
+      "grad_norm": 1.1983983516693115,
+      "learning_rate": 2.096708333080241e-05,
+      "loss": 0.2699,
+      "step": 890
+    },
+    {
+      "epoch": 2.0203389830508476,
+      "grad_norm": 1.1153104305267334,
+      "learning_rate": 2.086417382862518e-05,
+      "loss": 0.2679,
+      "step": 895
+    },
+    {
+      "epoch": 2.031638418079096,
+      "grad_norm": 1.1299874782562256,
+      "learning_rate": 2.076093747297696e-05,
+      "loss": 0.3141,
+      "step": 900
+    },
+    {
+      "epoch": 2.0429378531073445,
+      "grad_norm": 1.0530294179916382,
+      "learning_rate": 2.0657380017977964e-05,
+      "loss": 0.2703,
+      "step": 905
+    },
+    {
+      "epoch": 2.0542372881355933,
+      "grad_norm": 1.1320830583572388,
+      "learning_rate": 2.0553507235645648e-05,
+      "loss": 0.2921,
+      "step": 910
+    },
+    {
+      "epoch": 2.065536723163842,
+      "grad_norm": 1.1087229251861572,
+      "learning_rate": 2.0449324915572942e-05,
+      "loss": 0.2601,
+      "step": 915
+    },
+    {
+      "epoch": 2.07683615819209,
+      "grad_norm": 1.1149791479110718,
+      "learning_rate": 2.0344838864605614e-05,
+      "loss": 0.2695,
+      "step": 920
+    },
+    {
+      "epoch": 2.088135593220339,
+      "grad_norm": 1.2066401243209839,
+      "learning_rate": 2.0240054906518574e-05,
+      "loss": 0.2857,
+      "step": 925
+    },
+    {
+      "epoch": 2.0994350282485876,
+      "grad_norm": 1.0312755107879639,
+      "learning_rate": 2.0134978881691292e-05,
+      "loss": 0.2736,
+      "step": 930
+    },
+    {
+      "epoch": 2.1107344632768363,
+      "grad_norm": 1.054076075553894,
+      "learning_rate": 2.0029616646782255e-05,
+      "loss": 0.2591,
+      "step": 935
+    },
+    {
+      "epoch": 2.1220338983050846,
+      "grad_norm": 1.2717868089675903,
+      "learning_rate": 1.9923974074402544e-05,
+      "loss": 0.2926,
+      "step": 940
+    },
+    {
+      "epoch": 2.1333333333333333,
+      "grad_norm": 1.154904842376709,
+      "learning_rate": 1.981805705278849e-05,
+      "loss": 0.3017,
+      "step": 945
+    },
+    {
+      "epoch": 2.144632768361582,
+      "grad_norm": 1.2696698904037476,
+      "learning_rate": 1.9711871485473523e-05,
+      "loss": 0.2602,
+      "step": 950
+    },
+    {
+      "epoch": 2.1559322033898307,
+      "grad_norm": 1.1353667974472046,
+      "learning_rate": 1.9605423290959077e-05,
+      "loss": 0.278,
+      "step": 955
+    },
+    {
+      "epoch": 2.167231638418079,
+      "grad_norm": 1.2267999649047852,
+      "learning_rate": 1.9498718402384732e-05,
+      "loss": 0.2874,
+      "step": 960
+    },
+    {
+      "epoch": 2.1785310734463277,
+      "grad_norm": 1.1276496648788452,
+      "learning_rate": 1.9391762767197516e-05,
+      "loss": 0.2615,
+      "step": 965
+    },
+    {
+      "epoch": 2.1898305084745764,
+      "grad_norm": 1.0736314058303833,
+      "learning_rate": 1.928456234682041e-05,
+      "loss": 0.2785,
+      "step": 970
+    },
+    {
+      "epoch": 2.201129943502825,
+      "grad_norm": 1.2606271505355835,
+      "learning_rate": 1.917712311632006e-05,
+      "loss": 0.2556,
+      "step": 975
+    },
+    {
+      "epoch": 2.2124293785310734,
+      "grad_norm": 1.205222249031067,
+      "learning_rate": 1.9069451064073764e-05,
+      "loss": 0.289,
+      "step": 980
+    },
+    {
+      "epoch": 2.223728813559322,
+      "grad_norm": 1.2200062274932861,
+      "learning_rate": 1.8961552191435684e-05,
+      "loss": 0.2479,
+      "step": 985
+    },
+    {
+      "epoch": 2.2350282485875708,
+      "grad_norm": 0.9834157228469849,
+      "learning_rate": 1.8853432512402348e-05,
+      "loss": 0.2476,
+      "step": 990
+    },
+    {
+      "epoch": 2.246327683615819,
+      "grad_norm": 1.0837196111679077,
+      "learning_rate": 1.874509805327744e-05,
+      "loss": 0.2739,
+      "step": 995
+    },
+    {
+      "epoch": 2.2576271186440677,
+      "grad_norm": 1.0946136713027954,
+      "learning_rate": 1.863655485233592e-05,
+      "loss": 0.2474,
+      "step": 1000
+    },
+    {
+      "epoch": 2.2689265536723164,
+      "grad_norm": 1.2106841802597046,
+      "learning_rate": 1.8527808959487468e-05,
+      "loss": 0.2697,
+      "step": 1005
+    },
+    {
+      "epoch": 2.280225988700565,
+      "grad_norm": 1.2549164295196533,
+      "learning_rate": 1.8418866435939267e-05,
+      "loss": 0.2757,
+      "step": 1010
+    },
+    {
+      "epoch": 2.2915254237288134,
+      "grad_norm": 1.1625417470932007,
+      "learning_rate": 1.8309733353858167e-05,
+      "loss": 0.257,
+      "step": 1015
+    },
+    {
+      "epoch": 2.302824858757062,
+      "grad_norm": 1.1273975372314453,
+      "learning_rate": 1.8200415796032254e-05,
+      "loss": 0.255,
+      "step": 1020
+    },
+    {
+      "epoch": 2.314124293785311,
+      "grad_norm": 1.0395770072937012,
+      "learning_rate": 1.80909198555318e-05,
+      "loss": 0.2498,
+      "step": 1025
+    },
+    {
+      "epoch": 2.3254237288135595,
+      "grad_norm": 1.2258102893829346,
+      "learning_rate": 1.798125163536965e-05,
+      "loss": 0.2448,
+      "step": 1030
+    },
+    {
+      "epoch": 2.336723163841808,
+      "grad_norm": 1.081221103668213,
+      "learning_rate": 1.787141724816106e-05,
+      "loss": 0.2508,
+      "step": 1035
+    },
+    {
+      "epoch": 2.3480225988700565,
+      "grad_norm": 1.3911811113357544,
+      "learning_rate": 1.7761422815783e-05,
+      "loss": 0.2542,
+      "step": 1040
+    },
+    {
+      "epoch": 2.359322033898305,
+      "grad_norm": 1.5137418508529663,
+      "learning_rate": 1.765127446903293e-05,
+      "loss": 0.2536,
+      "step": 1045
+    },
+    {
+      "epoch": 2.3706214689265535,
+      "grad_norm": 1.079277515411377,
+      "learning_rate": 1.754097834728708e-05,
+      "loss": 0.2452,
+      "step": 1050
+    },
+    {
+      "epoch": 2.381920903954802,
+      "grad_norm": 1.3662875890731812,
+      "learning_rate": 1.7430540598158287e-05,
+      "loss": 0.2538,
+      "step": 1055
+    },
+    {
+      "epoch": 2.393220338983051,
+      "grad_norm": 1.214477777481079,
+      "learning_rate": 1.7319967377153295e-05,
+      "loss": 0.2698,
+      "step": 1060
+    },
+    {
+      "epoch": 2.4045197740112996,
+      "grad_norm": 1.1170893907546997,
+      "learning_rate": 1.7209264847329717e-05,
+      "loss": 0.2284,
+      "step": 1065
+    },
+    {
+      "epoch": 2.415819209039548,
+      "grad_norm": 1.0510598421096802,
+      "learning_rate": 1.709843917895247e-05,
+      "loss": 0.2073,
+      "step": 1070
+    },
+    {
+      "epoch": 2.4271186440677965,
+      "grad_norm": 1.0740697383880615,
+      "learning_rate": 1.698749654914991e-05,
+      "loss": 0.2402,
+      "step": 1075
+    },
+    {
+      "epoch": 2.4384180790960452,
+      "grad_norm": 1.167296051979065,
+      "learning_rate": 1.6876443141569503e-05,
+      "loss": 0.2535,
+      "step": 1080
+    },
+    {
+      "epoch": 2.449717514124294,
+      "grad_norm": 1.1952712535858154,
+      "learning_rate": 1.6765285146033165e-05,
+      "loss": 0.2423,
+      "step": 1085
+    },
+    {
+      "epoch": 2.461016949152542,
+      "grad_norm": 0.95307457447052,
+      "learning_rate": 1.6654028758192286e-05,
+      "loss": 0.2457,
+      "step": 1090
+    },
+    {
+      "epoch": 2.472316384180791,
+      "grad_norm": 1.0912442207336426,
+      "learning_rate": 1.654268017918236e-05,
+      "loss": 0.2326,
+      "step": 1095
+    },
+    {
+      "epoch": 2.4836158192090396,
+      "grad_norm": 1.280784010887146,
+      "learning_rate": 1.643124561527739e-05,
+      "loss": 0.2414,
+      "step": 1100
+    },
+    {
+      "epoch": 2.4949152542372883,
+      "grad_norm": 1.0802363157272339,
+      "learning_rate": 1.631973127754394e-05,
+      "loss": 0.2236,
+      "step": 1105
+    },
+    {
+      "epoch": 2.5062146892655366,
+      "grad_norm": 1.074249267578125,
+      "learning_rate": 1.6208143381494955e-05,
+      "loss": 0.2317,
+      "step": 1110
+    },
+    {
+      "epoch": 2.5175141242937853,
+      "grad_norm": 1.433241844177246,
+      "learning_rate": 1.609648814674332e-05,
+      "loss": 0.2381,
+      "step": 1115
+    },
+    {
+      "epoch": 2.528813559322034,
+      "grad_norm": 1.030909776687622,
+      "learning_rate": 1.5984771796655223e-05,
+      "loss": 0.2604,
+      "step": 1120
+    },
+    {
+      "epoch": 2.5401129943502827,
+      "grad_norm": 1.213474988937378,
+      "learning_rate": 1.5873000558003227e-05,
+      "loss": 0.2453,
+      "step": 1125
+    },
+    {
+      "epoch": 2.551412429378531,
+      "grad_norm": 1.3832768201828003,
+      "learning_rate": 1.5761180660619267e-05,
+      "loss": 0.2381,
+      "step": 1130
+    },
+    {
+      "epoch": 2.5627118644067797,
+      "grad_norm": 1.210844874382019,
+      "learning_rate": 1.5649318337047365e-05,
+      "loss": 0.2279,
+      "step": 1135
+    },
+    {
+      "epoch": 2.5740112994350284,
+      "grad_norm": 1.2258585691452026,
+      "learning_rate": 1.553741982219627e-05,
+      "loss": 0.2251,
+      "step": 1140
+    },
+    {
+      "epoch": 2.5853107344632766,
+      "grad_norm": 1.11599862575531,
+      "learning_rate": 1.5425491352991947e-05,
+      "loss": 0.2224,
+      "step": 1145
+    },
+    {
+      "epoch": 2.5966101694915253,
+      "grad_norm": 1.0045859813690186,
+      "learning_rate": 1.5313539168029932e-05,
+      "loss": 0.213,
+      "step": 1150
+    },
+    {
+      "epoch": 2.607909604519774,
+      "grad_norm": 0.8923414349555969,
+      "learning_rate": 1.5201569507227609e-05,
+      "loss": 0.2049,
+      "step": 1155
+    },
+    {
+      "epoch": 2.6192090395480228,
+      "grad_norm": 1.2115402221679688,
+      "learning_rate": 1.5089588611476426e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 2.6305084745762715,
+      "grad_norm": 1.0908693075180054,
+      "learning_rate": 1.497760272229403e-05,
+      "loss": 0.25,
+      "step": 1165
+    },
+    {
+      "epoch": 2.6418079096045197,
+      "grad_norm": 1.0529178380966187,
+      "learning_rate": 1.4865618081476394e-05,
+      "loss": 0.2034,
+      "step": 1170
+    },
+    {
+      "epoch": 2.6531073446327684,
+      "grad_norm": 1.1474601030349731,
+      "learning_rate": 1.475364093074991e-05,
+      "loss": 0.224,
+      "step": 1175
+    },
+    {
+      "epoch": 2.664406779661017,
+      "grad_norm": 1.0988585948944092,
+      "learning_rate": 1.4641677511423497e-05,
+      "loss": 0.2367,
+      "step": 1180
+    },
+    {
+      "epoch": 2.6757062146892654,
+      "grad_norm": 1.174198031425476,
+      "learning_rate": 1.4529734064040704e-05,
+      "loss": 0.229,
+      "step": 1185
+    },
+    {
+      "epoch": 2.687005649717514,
+      "grad_norm": 1.0259079933166504,
+      "learning_rate": 1.4417816828031916e-05,
+      "loss": 0.2048,
+      "step": 1190
+    },
+    {
+      "epoch": 2.698305084745763,
+      "grad_norm": 1.0843926668167114,
+      "learning_rate": 1.4305932041366566e-05,
+      "loss": 0.2094,
+      "step": 1195
+    },
+    {
+      "epoch": 2.709604519774011,
+      "grad_norm": 1.0300178527832031,
+      "learning_rate": 1.4194085940205427e-05,
+      "loss": 0.2288,
+      "step": 1200
+    },
+    {
+      "epoch": 2.7209039548022598,
+      "grad_norm": 1.406386375427246,
+      "learning_rate": 1.4082284758553067e-05,
+      "loss": 0.2153,
+      "step": 1205
+    },
+    {
+      "epoch": 2.7322033898305085,
+      "grad_norm": 1.2302467823028564,
+      "learning_rate": 1.3970534727910346e-05,
+      "loss": 0.219,
+      "step": 1210
+    },
+    {
+      "epoch": 2.743502824858757,
+      "grad_norm": 1.163407564163208,
+      "learning_rate": 1.3858842076927114e-05,
+      "loss": 0.2121,
+      "step": 1215
+    },
+    {
+      "epoch": 2.754802259887006,
+      "grad_norm": 1.185469150543213,
+      "learning_rate": 1.3747213031055015e-05,
+      "loss": 0.215,
+      "step": 1220
+    },
+    {
+      "epoch": 2.766101694915254,
+      "grad_norm": 1.2026503086090088,
+      "learning_rate": 1.363565381220053e-05,
+      "loss": 0.2025,
+      "step": 1225
+    },
+    {
+      "epoch": 2.777401129943503,
+      "grad_norm": 1.0308500528335571,
+      "learning_rate": 1.352417063837816e-05,
+      "loss": 0.1976,
+      "step": 1230
+    },
+    {
+      "epoch": 2.7887005649717516,
+      "grad_norm": 1.136925220489502,
+      "learning_rate": 1.3412769723363848e-05,
+      "loss": 0.2244,
+      "step": 1235
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 1.0420323610305786,
+      "learning_rate": 1.3301457276348668e-05,
+      "loss": 0.216,
+      "step": 1240
+    },
+    {
+      "epoch": 2.8112994350282485,
+      "grad_norm": 1.2677199840545654,
+      "learning_rate": 1.3190239501592721e-05,
+      "loss": 0.2261,
+      "step": 1245
+    },
+    {
+      "epoch": 2.8225988700564972,
+      "grad_norm": 1.1373021602630615,
+      "learning_rate": 1.3079122598079307e-05,
+      "loss": 0.2101,
+      "step": 1250
+    },
+    {
+      "epoch": 2.8338983050847455,
+      "grad_norm": 1.2039943933486938,
+      "learning_rate": 1.2968112759169449e-05,
+      "loss": 0.1857,
+      "step": 1255
+    },
+    {
+      "epoch": 2.845197740112994,
+      "grad_norm": 1.2662607431411743,
+      "learning_rate": 1.285721617225668e-05,
+      "loss": 0.1921,
+      "step": 1260
+    },
+    {
+      "epoch": 2.856497175141243,
+      "grad_norm": 1.051405668258667,
+      "learning_rate": 1.2746439018422142e-05,
+      "loss": 0.1874,
+      "step": 1265
+    },
+    {
+      "epoch": 2.8677966101694916,
+      "grad_norm": 1.2025455236434937,
+      "learning_rate": 1.2635787472090115e-05,
+      "loss": 0.2216,
+      "step": 1270
+    },
+    {
+      "epoch": 2.8790960451977403,
+      "grad_norm": 1.1472687721252441,
+      "learning_rate": 1.2525267700683846e-05,
+      "loss": 0.1972,
+      "step": 1275
+    },
+    {
+      "epoch": 2.8903954802259886,
+      "grad_norm": 1.0994764566421509,
+      "learning_rate": 1.2414885864281804e-05,
+      "loss": 0.1978,
+      "step": 1280
+    },
+    {
+      "epoch": 2.9016949152542373,
+      "grad_norm": 1.150343894958496,
+      "learning_rate": 1.2304648115274303e-05,
+      "loss": 0.2111,
+      "step": 1285
+    },
+    {
+      "epoch": 2.912994350282486,
+      "grad_norm": 1.064850091934204,
+      "learning_rate": 1.2194560598020637e-05,
+      "loss": 0.2019,
+      "step": 1290
+    },
+    {
+      "epoch": 2.9242937853107343,
+      "grad_norm": 1.053019642829895,
+      "learning_rate": 1.2084629448506568e-05,
+      "loss": 0.1839,
+      "step": 1295
+    },
+    {
+      "epoch": 2.935593220338983,
+      "grad_norm": 1.14827561378479,
+      "learning_rate": 1.1974860794002337e-05,
+      "loss": 0.1863,
+      "step": 1300
+    },
+    {
+      "epoch": 2.9468926553672317,
+      "grad_norm": 1.2147302627563477,
+      "learning_rate": 1.1865260752721141e-05,
+      "loss": 0.1949,
+      "step": 1305
+    },
+    {
+      "epoch": 2.9581920903954804,
+      "grad_norm": 1.1208956241607666,
+      "learning_rate": 1.1755835433478136e-05,
+      "loss": 0.1789,
+      "step": 1310
+    },
+    {
+      "epoch": 2.969491525423729,
+      "grad_norm": 0.9414421319961548,
+      "learning_rate": 1.1646590935349917e-05,
+      "loss": 0.1913,
+      "step": 1315
+    },
+    {
+      "epoch": 2.9807909604519773,
+      "grad_norm": 1.3583221435546875,
+      "learning_rate": 1.1537533347334606e-05,
+      "loss": 0.1887,
+      "step": 1320
+    },
+    {
+      "epoch": 2.992090395480226,
+      "grad_norm": 1.1631615161895752,
+      "learning_rate": 1.1428668748012453e-05,
+      "loss": 0.1915,
+      "step": 1325
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2215,
+  "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": 1.9186562394750976e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

37_128_e5_3e-5/checkpoint-1329/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2fc91f88ecbc07e44fa9bd105383aeb3bf6c3739d7fa469fb2915f1ea1cc71e
+size 7736

37_128_e5_3e-5/checkpoint-1329/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)

37_128_e5_3e-5/checkpoint-1772/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

37_128_e5_3e-5/checkpoint-1772/adapter_config.json

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

37_128_e5_3e-5/checkpoint-1772/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a3a8e6cca969bddf8b5bdd3a59a5e060715f09f4dc48b77ca9bbd927f88d861
+size 791751704

37_128_e5_3e-5/checkpoint-1772/latest

@@ -0,0 +1 @@
+global_step1772

37_128_e5_3e-5/checkpoint-1772/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:314bbf777da15d8aa5def2f25e7fed54c3cf2bb98785655107491c89be6c5fb6
+size 15920

37_128_e5_3e-5/checkpoint-1772/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:220cef892237e9f19011362bbb8ae6d4d3ea06d39e0853fac153b1e1ef38a27c
+size 15920

37_128_e5_3e-5/checkpoint-1772/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:483c65906411d3777536a6f9a5b3224ee86ba69d161241783187f1a5c44ae31d
+size 15920

37_128_e5_3e-5/checkpoint-1772/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:91a198b2bb611561e26ec8f8e670ab5922ef5116bb7bec8a3fb6fbe13c65d057
+size 15920

37_128_e5_3e-5/checkpoint-1772/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55f9edbfb7e302bff1695f8111c0456e3613d929f083c8a85bb1907c5cd3ecd8
+size 15920

37_128_e5_3e-5/checkpoint-1772/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:940063fd3f060f156eb7fac25c463077a2e8a0c97934e2997fd1a5f3dbc69122
+size 15920

37_128_e5_3e-5/checkpoint-1772/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a595c994639e08543a511c3487f36648d8d861fe48b43df7929e6fdbe6993d9
+size 15920

37_128_e5_3e-5/checkpoint-1772/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b5efbb812d8bc2e3b3d5fc748b7b4838b47442226f0fefca844234c2a47d4d5
+size 15920

37_128_e5_3e-5/checkpoint-1772/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:64c282d58785dfd23259d64feb719ef582af412274a87cfaaaad51d1d392d5af
+size 1064

37_128_e5_3e-5/checkpoint-1772/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
+  }
+}

37_128_e5_3e-5/checkpoint-1772/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
+}

37_128_e5_3e-5/checkpoint-1772/trainer_state.json

@@ -0,0 +1,2512 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1772,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011299435028248588,
+      "grad_norm": 1.2254456281661987,
+      "learning_rate": 1.0810810810810812e-06,
+      "loss": 1.3157,
+      "step": 5
+    },
+    {
+      "epoch": 0.022598870056497175,
+      "grad_norm": 1.2222959995269775,
+      "learning_rate": 2.4324324324324325e-06,
+      "loss": 1.3831,
+      "step": 10
+    },
+    {
+      "epoch": 0.03389830508474576,
+      "grad_norm": 0.7486730813980103,
+      "learning_rate": 3.7837837837837835e-06,
+      "loss": 1.3448,
+      "step": 15
+    },
+    {
+      "epoch": 0.04519774011299435,
+      "grad_norm": 0.7303342819213867,
+      "learning_rate": 5.135135135135135e-06,
+      "loss": 1.2883,
+      "step": 20
+    },
+    {
+      "epoch": 0.05649717514124294,
+      "grad_norm": 0.6827901005744934,
+      "learning_rate": 6.486486486486487e-06,
+      "loss": 1.2782,
+      "step": 25
+    },
+    {
+      "epoch": 0.06779661016949153,
+      "grad_norm": 0.6475429534912109,
+      "learning_rate": 7.837837837837838e-06,
+      "loss": 1.2937,
+      "step": 30
+    },
+    {
+      "epoch": 0.07909604519774012,
+      "grad_norm": 0.5939517617225647,
+      "learning_rate": 9.18918918918919e-06,
+      "loss": 1.2553,
+      "step": 35
+    },
+    {
+      "epoch": 0.0903954802259887,
+      "grad_norm": 0.5064879655838013,
+      "learning_rate": 1.0540540540540541e-05,
+      "loss": 1.2067,
+      "step": 40
+    },
+    {
+      "epoch": 0.1016949152542373,
+      "grad_norm": 0.6584354639053345,
+      "learning_rate": 1.1891891891891893e-05,
+      "loss": 1.2246,
+      "step": 45
+    },
+    {
+      "epoch": 0.11299435028248588,
+      "grad_norm": 0.4975661337375641,
+      "learning_rate": 1.3243243243243242e-05,
+      "loss": 1.2558,
+      "step": 50
+    },
+    {
+      "epoch": 0.12429378531073447,
+      "grad_norm": 0.5031336545944214,
+      "learning_rate": 1.4594594594594596e-05,
+      "loss": 1.2726,
+      "step": 55
+    },
+    {
+      "epoch": 0.13559322033898305,
+      "grad_norm": 0.4471903443336487,
+      "learning_rate": 1.5945945945945947e-05,
+      "loss": 1.2089,
+      "step": 60
+    },
+    {
+      "epoch": 0.14689265536723164,
+      "grad_norm": 0.5343024134635925,
+      "learning_rate": 1.72972972972973e-05,
+      "loss": 1.2018,
+      "step": 65
+    },
+    {
+      "epoch": 0.15819209039548024,
+      "grad_norm": 0.47138091921806335,
+      "learning_rate": 1.864864864864865e-05,
+      "loss": 1.2826,
+      "step": 70
+    },
+    {
+      "epoch": 0.1694915254237288,
+      "grad_norm": 0.490500271320343,
+      "learning_rate": 1.9999999999999998e-05,
+      "loss": 1.2116,
+      "step": 75
+    },
+    {
+      "epoch": 0.1807909604519774,
+      "grad_norm": 0.5078205466270447,
+      "learning_rate": 2.135135135135135e-05,
+      "loss": 1.1873,
+      "step": 80
+    },
+    {
+      "epoch": 0.192090395480226,
+      "grad_norm": 1.9143681526184082,
+      "learning_rate": 2.2702702702702705e-05,
+      "loss": 1.2159,
+      "step": 85
+    },
+    {
+      "epoch": 0.2033898305084746,
+      "grad_norm": 0.5685150027275085,
+      "learning_rate": 2.4054054054054056e-05,
+      "loss": 1.1666,
+      "step": 90
+    },
+    {
+      "epoch": 0.21468926553672316,
+      "grad_norm": 0.5490728616714478,
+      "learning_rate": 2.5405405405405408e-05,
+      "loss": 1.1348,
+      "step": 95
+    },
+    {
+      "epoch": 0.22598870056497175,
+      "grad_norm": 0.5205124020576477,
+      "learning_rate": 2.6756756756756756e-05,
+      "loss": 1.1213,
+      "step": 100
+    },
+    {
+      "epoch": 0.23728813559322035,
+      "grad_norm": 0.4488866925239563,
+      "learning_rate": 2.8108108108108107e-05,
+      "loss": 1.1414,
+      "step": 105
+    },
+    {
+      "epoch": 0.24858757062146894,
+      "grad_norm": 0.4913260340690613,
+      "learning_rate": 2.945945945945946e-05,
+      "loss": 1.139,
+      "step": 110
+    },
+    {
+      "epoch": 0.2598870056497175,
+      "grad_norm": 0.5297622084617615,
+      "learning_rate": 2.9999849508725208e-05,
+      "loss": 1.134,
+      "step": 115
+    },
+    {
+      "epoch": 0.2711864406779661,
+      "grad_norm": 0.6094189882278442,
+      "learning_rate": 2.9998929850759165e-05,
+      "loss": 1.1023,
+      "step": 120
+    },
+    {
+      "epoch": 0.2824858757062147,
+      "grad_norm": 0.6322339177131653,
+      "learning_rate": 2.9997174192288006e-05,
+      "loss": 1.1164,
+      "step": 125
+    },
+    {
+      "epoch": 0.2937853107344633,
+      "grad_norm": 0.6201522946357727,
+      "learning_rate": 2.999458263116747e-05,
+      "loss": 1.1188,
+      "step": 130
+    },
+    {
+      "epoch": 0.3050847457627119,
+      "grad_norm": 0.5387688875198364,
+      "learning_rate": 2.9991155311844292e-05,
+      "loss": 1.0651,
+      "step": 135
+    },
+    {
+      "epoch": 0.3163841807909605,
+      "grad_norm": 0.5230715274810791,
+      "learning_rate": 2.998689242534815e-05,
+      "loss": 1.0626,
+      "step": 140
+    },
+    {
+      "epoch": 0.327683615819209,
+      "grad_norm": 0.5703102946281433,
+      "learning_rate": 2.9981794209281003e-05,
+      "loss": 1.0626,
+      "step": 145
+    },
+    {
+      "epoch": 0.3389830508474576,
+      "grad_norm": 0.5561928153038025,
+      "learning_rate": 2.997586094780388e-05,
+      "loss": 1.0494,
+      "step": 150
+    },
+    {
+      "epoch": 0.3502824858757062,
+      "grad_norm": 0.6391568183898926,
+      "learning_rate": 2.9969092971621006e-05,
+      "loss": 1.05,
+      "step": 155
+    },
+    {
+      "epoch": 0.3615819209039548,
+      "grad_norm": 0.6165516376495361,
+      "learning_rate": 2.9961490657961404e-05,
+      "loss": 1.0617,
+      "step": 160
+    },
+    {
+      "epoch": 0.3728813559322034,
+      "grad_norm": 0.6393241286277771,
+      "learning_rate": 2.9953054430557835e-05,
+      "loss": 1.0036,
+      "step": 165
+    },
+    {
+      "epoch": 0.384180790960452,
+      "grad_norm": 0.6466289758682251,
+      "learning_rate": 2.9943784759623205e-05,
+      "loss": 1.0362,
+      "step": 170
+    },
+    {
+      "epoch": 0.3954802259887006,
+      "grad_norm": 0.7020736932754517,
+      "learning_rate": 2.993368216182435e-05,
+      "loss": 1.0052,
+      "step": 175
+    },
+    {
+      "epoch": 0.4067796610169492,
+      "grad_norm": 0.630094051361084,
+      "learning_rate": 2.992274720025323e-05,
+      "loss": 1.0279,
+      "step": 180
+    },
+    {
+      "epoch": 0.4180790960451977,
+      "grad_norm": 0.629647433757782,
+      "learning_rate": 2.9910980484395555e-05,
+      "loss": 1.0309,
+      "step": 185
+    },
+    {
+      "epoch": 0.4293785310734463,
+      "grad_norm": 0.6576105952262878,
+      "learning_rate": 2.989838267009681e-05,
+      "loss": 0.9227,
+      "step": 190
+    },
+    {
+      "epoch": 0.4406779661016949,
+      "grad_norm": 0.6078037023544312,
+      "learning_rate": 2.988495445952569e-05,
+      "loss": 1.0005,
+      "step": 195
+    },
+    {
+      "epoch": 0.4519774011299435,
+      "grad_norm": 0.657957911491394,
+      "learning_rate": 2.9870696601134994e-05,
+      "loss": 0.9715,
+      "step": 200
+    },
+    {
+      "epoch": 0.4632768361581921,
+      "grad_norm": 0.6783279180526733,
+      "learning_rate": 2.9855609889619864e-05,
+      "loss": 0.927,
+      "step": 205
+    },
+    {
+      "epoch": 0.4745762711864407,
+      "grad_norm": 0.6905434727668762,
+      "learning_rate": 2.983969516587352e-05,
+      "loss": 0.9506,
+      "step": 210
+    },
+    {
+      "epoch": 0.4858757062146893,
+      "grad_norm": 0.8056045174598694,
+      "learning_rate": 2.9822953316940393e-05,
+      "loss": 0.9867,
+      "step": 215
+    },
+    {
+      "epoch": 0.4971751412429379,
+      "grad_norm": 0.7994968295097351,
+      "learning_rate": 2.9805385275966658e-05,
+      "loss": 0.9489,
+      "step": 220
+    },
+    {
+      "epoch": 0.5084745762711864,
+      "grad_norm": 0.7940033078193665,
+      "learning_rate": 2.978699202214826e-05,
+      "loss": 0.9538,
+      "step": 225
+    },
+    {
+      "epoch": 0.519774011299435,
+      "grad_norm": 0.7576586604118347,
+      "learning_rate": 2.97677745806763e-05,
+      "loss": 0.9529,
+      "step": 230
+    },
+    {
+      "epoch": 0.5310734463276836,
+      "grad_norm": 0.8559620976448059,
+      "learning_rate": 2.9747734022679913e-05,
+      "loss": 0.9188,
+      "step": 235
+    },
+    {
+      "epoch": 0.5423728813559322,
+      "grad_norm": 0.9041480422019958,
+      "learning_rate": 2.9726871465166567e-05,
+      "loss": 0.9145,
+      "step": 240
+    },
+    {
+      "epoch": 0.5536723163841808,
+      "grad_norm": 0.7606709003448486,
+      "learning_rate": 2.970518807095979e-05,
+      "loss": 0.9179,
+      "step": 245
+    },
+    {
+      "epoch": 0.5649717514124294,
+      "grad_norm": 0.904378354549408,
+      "learning_rate": 2.9682685048634384e-05,
+      "loss": 0.8682,
+      "step": 250
+    },
+    {
+      "epoch": 0.576271186440678,
+      "grad_norm": 0.8670118451118469,
+      "learning_rate": 2.9659363652449027e-05,
+      "loss": 0.9448,
+      "step": 255
+    },
+    {
+      "epoch": 0.5875706214689266,
+      "grad_norm": 0.7508774399757385,
+      "learning_rate": 2.9635225182276387e-05,
+      "loss": 0.9223,
+      "step": 260
+    },
+    {
+      "epoch": 0.5988700564971752,
+      "grad_norm": 0.6919876933097839,
+      "learning_rate": 2.961027098353067e-05,
+      "loss": 0.8971,
+      "step": 265
+    },
+    {
+      "epoch": 0.6101694915254238,
+      "grad_norm": 0.7949540615081787,
+      "learning_rate": 2.9584502447092623e-05,
+      "loss": 0.91,
+      "step": 270
+    },
+    {
+      "epoch": 0.6214689265536724,
+      "grad_norm": 0.7679325938224792,
+      "learning_rate": 2.9557921009232013e-05,
+      "loss": 0.903,
+      "step": 275
+    },
+    {
+      "epoch": 0.632768361581921,
+      "grad_norm": 0.8364932537078857,
+      "learning_rate": 2.953052815152757e-05,
+      "loss": 0.8941,
+      "step": 280
+    },
+    {
+      "epoch": 0.6440677966101694,
+      "grad_norm": 0.8773370981216431,
+      "learning_rate": 2.950232540078442e-05,
+      "loss": 0.8812,
+      "step": 285
+    },
+    {
+      "epoch": 0.655367231638418,
+      "grad_norm": 1.0121004581451416,
+      "learning_rate": 2.947331432894896e-05,
+      "loss": 0.901,
+      "step": 290
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.8700821995735168,
+      "learning_rate": 2.9443496553021268e-05,
+      "loss": 0.8385,
+      "step": 295
+    },
+    {
+      "epoch": 0.6779661016949152,
+      "grad_norm": 0.8492361903190613,
+      "learning_rate": 2.9412873734964973e-05,
+      "loss": 0.8498,
+      "step": 300
+    },
+    {
+      "epoch": 0.6892655367231638,
+      "grad_norm": 0.860744833946228,
+      "learning_rate": 2.938144758161459e-05,
+      "loss": 0.8381,
+      "step": 305
+    },
+    {
+      "epoch": 0.7005649717514124,
+      "grad_norm": 0.8146252036094666,
+      "learning_rate": 2.934921984458043e-05,
+      "loss": 0.8505,
+      "step": 310
+    },
+    {
+      "epoch": 0.711864406779661,
+      "grad_norm": 0.828873336315155,
+      "learning_rate": 2.931619232015094e-05,
+      "loss": 0.8359,
+      "step": 315
+    },
+    {
+      "epoch": 0.7231638418079096,
+      "grad_norm": 0.8928766250610352,
+      "learning_rate": 2.9282366849192596e-05,
+      "loss": 0.8697,
+      "step": 320
+    },
+    {
+      "epoch": 0.7344632768361582,
+      "grad_norm": 0.8543341755867004,
+      "learning_rate": 2.924774531704729e-05,
+      "loss": 0.7961,
+      "step": 325
+    },
+    {
+      "epoch": 0.7457627118644068,
+      "grad_norm": 0.8456112146377563,
+      "learning_rate": 2.921232965342725e-05,
+      "loss": 0.8298,
+      "step": 330
+    },
+    {
+      "epoch": 0.7570621468926554,
+      "grad_norm": 0.8231430053710938,
+      "learning_rate": 2.9176121832307487e-05,
+      "loss": 0.7866,
+      "step": 335
+    },
+    {
+      "epoch": 0.768361581920904,
+      "grad_norm": 0.8562153577804565,
+      "learning_rate": 2.9139123871815762e-05,
+      "loss": 0.7833,
+      "step": 340
+    },
+    {
+      "epoch": 0.7796610169491526,
+      "grad_norm": 0.8552905917167664,
+      "learning_rate": 2.9101337834120113e-05,
+      "loss": 0.7777,
+      "step": 345
+    },
+    {
+      "epoch": 0.7909604519774012,
+      "grad_norm": 0.9859520792961121,
+      "learning_rate": 2.9062765825313887e-05,
+      "loss": 0.8016,
+      "step": 350
+    },
+    {
+      "epoch": 0.8022598870056498,
+      "grad_norm": 0.867014467716217,
+      "learning_rate": 2.9023409995298406e-05,
+      "loss": 0.7895,
+      "step": 355
+    },
+    {
+      "epoch": 0.8135593220338984,
+      "grad_norm": 0.890496015548706,
+      "learning_rate": 2.8983272537663082e-05,
+      "loss": 0.7795,
+      "step": 360
+    },
+    {
+      "epoch": 0.8248587570621468,
+      "grad_norm": 0.9742592573165894,
+      "learning_rate": 2.8942355689563166e-05,
+      "loss": 0.7661,
+      "step": 365
+    },
+    {
+      "epoch": 0.8361581920903954,
+      "grad_norm": 0.855208694934845,
+      "learning_rate": 2.890066173159509e-05,
+      "loss": 0.7863,
+      "step": 370
+    },
+    {
+      "epoch": 0.847457627118644,
+      "grad_norm": 0.9460350871086121,
+      "learning_rate": 2.8858192987669303e-05,
+      "loss": 0.7475,
+      "step": 375
+    },
+    {
+      "epoch": 0.8587570621468926,
+      "grad_norm": 0.9050624966621399,
+      "learning_rate": 2.881495182488077e-05,
+      "loss": 0.739,
+      "step": 380
+    },
+    {
+      "epoch": 0.8700564971751412,
+      "grad_norm": 0.8351825475692749,
+      "learning_rate": 2.8770940653377047e-05,
+      "loss": 0.7322,
+      "step": 385
+    },
+    {
+      "epoch": 0.8813559322033898,
+      "grad_norm": 0.9222843050956726,
+      "learning_rate": 2.8726161926223904e-05,
+      "loss": 0.7435,
+      "step": 390
+    },
+    {
+      "epoch": 0.8926553672316384,
+      "grad_norm": 0.8897649645805359,
+      "learning_rate": 2.8680618139268643e-05,
+      "loss": 0.7229,
+      "step": 395
+    },
+    {
+      "epoch": 0.903954802259887,
+      "grad_norm": 1.0437895059585571,
+      "learning_rate": 2.8634311831000966e-05,
+      "loss": 0.7132,
+      "step": 400
+    },
+    {
+      "epoch": 0.9152542372881356,
+      "grad_norm": 0.9872046113014221,
+      "learning_rate": 2.8587245582411475e-05,
+      "loss": 0.7257,
+      "step": 405
+    },
+    {
+      "epoch": 0.9265536723163842,
+      "grad_norm": 0.9081714749336243,
+      "learning_rate": 2.8539422016847834e-05,
+      "loss": 0.6956,
+      "step": 410
+    },
+    {
+      "epoch": 0.9378531073446328,
+      "grad_norm": 0.9939802289009094,
+      "learning_rate": 2.8490843799868556e-05,
+      "loss": 0.723,
+      "step": 415
+    },
+    {
+      "epoch": 0.9491525423728814,
+      "grad_norm": 0.8449594974517822,
+      "learning_rate": 2.8441513639094396e-05,
+      "loss": 0.7294,
+      "step": 420
+    },
+    {
+      "epoch": 0.96045197740113,
+      "grad_norm": 0.9042068719863892,
+      "learning_rate": 2.8391434284057466e-05,
+      "loss": 0.6565,
+      "step": 425
+    },
+    {
+      "epoch": 0.9717514124293786,
+      "grad_norm": 0.9338545203208923,
+      "learning_rate": 2.8340608526047995e-05,
+      "loss": 0.711,
+      "step": 430
+    },
+    {
+      "epoch": 0.9830508474576272,
+      "grad_norm": 0.9345859885215759,
+      "learning_rate": 2.82890391979587e-05,
+      "loss": 0.7006,
+      "step": 435
+    },
+    {
+      "epoch": 0.9943502824858758,
+      "grad_norm": 0.9804854393005371,
+      "learning_rate": 2.8236729174126948e-05,
+      "loss": 0.7183,
+      "step": 440
+    },
+    {
+      "epoch": 1.0045197740112994,
+      "grad_norm": 0.9410772919654846,
+      "learning_rate": 2.818368137017449e-05,
+      "loss": 0.6665,
+      "step": 445
+    },
+    {
+      "epoch": 1.015819209039548,
+      "grad_norm": 1.156546950340271,
+      "learning_rate": 2.8129898742845006e-05,
+      "loss": 0.616,
+      "step": 450
+    },
+    {
+      "epoch": 1.0271186440677966,
+      "grad_norm": 1.0281895399093628,
+      "learning_rate": 2.8075384289839274e-05,
+      "loss": 0.5688,
+      "step": 455
+    },
+    {
+      "epoch": 1.038418079096045,
+      "grad_norm": 0.9605177044868469,
+      "learning_rate": 2.8020141049648072e-05,
+      "loss": 0.6208,
+      "step": 460
+    },
+    {
+      "epoch": 1.0497175141242938,
+      "grad_norm": 1.0274258852005005,
+      "learning_rate": 2.796417210138287e-05,
+      "loss": 0.578,
+      "step": 465
+    },
+    {
+      "epoch": 1.0610169491525423,
+      "grad_norm": 1.0010178089141846,
+      "learning_rate": 2.7907480564604155e-05,
+      "loss": 0.5861,
+      "step": 470
+    },
+    {
+      "epoch": 1.072316384180791,
+      "grad_norm": 1.5158096551895142,
+      "learning_rate": 2.7850069599147584e-05,
+      "loss": 0.5705,
+      "step": 475
+    },
+    {
+      "epoch": 1.0836158192090395,
+      "grad_norm": 1.0114028453826904,
+      "learning_rate": 2.7791942404947857e-05,
+      "loss": 0.5903,
+      "step": 480
+    },
+    {
+      "epoch": 1.0949152542372882,
+      "grad_norm": 0.9214546084403992,
+      "learning_rate": 2.773310222186037e-05,
+      "loss": 0.5517,
+      "step": 485
+    },
+    {
+      "epoch": 1.1062146892655367,
+      "grad_norm": 1.062912106513977,
+      "learning_rate": 2.7673552329480615e-05,
+      "loss": 0.6101,
+      "step": 490
+    },
+    {
+      "epoch": 1.1175141242937854,
+      "grad_norm": 1.0093401670455933,
+      "learning_rate": 2.76132960469614e-05,
+      "loss": 0.6117,
+      "step": 495
+    },
+    {
+      "epoch": 1.1288135593220339,
+      "grad_norm": 1.0780502557754517,
+      "learning_rate": 2.7552336732827846e-05,
+      "loss": 0.5611,
+      "step": 500
+    },
+    {
+      "epoch": 1.1401129943502826,
+      "grad_norm": 1.2409229278564453,
+      "learning_rate": 2.749067778479017e-05,
+      "loss": 0.5852,
+      "step": 505
+    },
+    {
+      "epoch": 1.151412429378531,
+      "grad_norm": 1.0648524761199951,
+      "learning_rate": 2.742832263955435e-05,
+      "loss": 0.5707,
+      "step": 510
+    },
+    {
+      "epoch": 1.1627118644067798,
+      "grad_norm": 1.1011556386947632,
+      "learning_rate": 2.7365274772630535e-05,
+      "loss": 0.5322,
+      "step": 515
+    },
+    {
+      "epoch": 1.1740112994350282,
+      "grad_norm": 1.0446882247924805,
+      "learning_rate": 2.730153769813933e-05,
+      "loss": 0.5517,
+      "step": 520
+    },
+    {
+      "epoch": 1.1853107344632767,
+      "grad_norm": 1.1494063138961792,
+      "learning_rate": 2.723711496861596e-05,
+      "loss": 0.5803,
+      "step": 525
+    },
+    {
+      "epoch": 1.1966101694915254,
+      "grad_norm": 0.9970608949661255,
+      "learning_rate": 2.7172010174812226e-05,
+      "loss": 0.611,
+      "step": 530
+    },
+    {
+      "epoch": 1.207909604519774,
+      "grad_norm": 1.0200440883636475,
+      "learning_rate": 2.710622694549638e-05,
+      "loss": 0.5742,
+      "step": 535
+    },
+    {
+      "epoch": 1.2192090395480226,
+      "grad_norm": 1.040436863899231,
+      "learning_rate": 2.7039768947250874e-05,
+      "loss": 0.5131,
+      "step": 540
+    },
+    {
+      "epoch": 1.230508474576271,
+      "grad_norm": 0.9933497309684753,
+      "learning_rate": 2.697263988426798e-05,
+      "loss": 0.5384,
+      "step": 545
+    },
+    {
+      "epoch": 1.2418079096045198,
+      "grad_norm": 1.1182832717895508,
+      "learning_rate": 2.6904843498143344e-05,
+      "loss": 0.5149,
+      "step": 550
+    },
+    {
+      "epoch": 1.2531073446327683,
+      "grad_norm": 1.002934217453003,
+      "learning_rate": 2.6836383567667423e-05,
+      "loss": 0.5396,
+      "step": 555
+    },
+    {
+      "epoch": 1.264406779661017,
+      "grad_norm": 1.090778112411499,
+      "learning_rate": 2.6767263908614884e-05,
+      "loss": 0.5593,
+      "step": 560
+    },
+    {
+      "epoch": 1.2757062146892655,
+      "grad_norm": 1.0608378648757935,
+      "learning_rate": 2.6697488373531904e-05,
+      "loss": 0.523,
+      "step": 565
+    },
+    {
+      "epoch": 1.2870056497175142,
+      "grad_norm": 1.0161354541778564,
+      "learning_rate": 2.662706085152144e-05,
+      "loss": 0.56,
+      "step": 570
+    },
+    {
+      "epoch": 1.2983050847457627,
+      "grad_norm": 1.0922820568084717,
+      "learning_rate": 2.655598526802649e-05,
+      "loss": 0.518,
+      "step": 575
+    },
+    {
+      "epoch": 1.3096045197740114,
+      "grad_norm": 0.9997158050537109,
+      "learning_rate": 2.6484265584611258e-05,
+      "loss": 0.5293,
+      "step": 580
+    },
+    {
+      "epoch": 1.3209039548022599,
+      "grad_norm": 1.1040353775024414,
+      "learning_rate": 2.6411905798740374e-05,
+      "loss": 0.5622,
+      "step": 585
+    },
+    {
+      "epoch": 1.3322033898305086,
+      "grad_norm": 0.9653554558753967,
+      "learning_rate": 2.6338909943556083e-05,
+      "loss": 0.482,
+      "step": 590
+    },
+    {
+      "epoch": 1.343502824858757,
+      "grad_norm": 1.0184459686279297,
+      "learning_rate": 2.6265282087653427e-05,
+      "loss": 0.5402,
+      "step": 595
+    },
+    {
+      "epoch": 1.3548022598870055,
+      "grad_norm": 1.1181055307388306,
+      "learning_rate": 2.619102633485352e-05,
+      "loss": 0.5269,
+      "step": 600
+    },
+    {
+      "epoch": 1.3661016949152542,
+      "grad_norm": 1.0076478719711304,
+      "learning_rate": 2.6116146823974747e-05,
+      "loss": 0.496,
+      "step": 605
+    },
+    {
+      "epoch": 1.377401129943503,
+      "grad_norm": 1.2491661310195923,
+      "learning_rate": 2.604064772860213e-05,
+      "loss": 0.5157,
+      "step": 610
+    },
+    {
+      "epoch": 1.3887005649717514,
+      "grad_norm": 1.0203555822372437,
+      "learning_rate": 2.5964533256854686e-05,
+      "loss": 0.4646,
+      "step": 615
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 1.079098105430603,
+      "learning_rate": 2.5887807651150865e-05,
+      "loss": 0.497,
+      "step": 620
+    },
+    {
+      "epoch": 1.4112994350282486,
+      "grad_norm": 1.1061233282089233,
+      "learning_rate": 2.5810475187972113e-05,
+      "loss": 0.4781,
+      "step": 625
+    },
+    {
+      "epoch": 1.422598870056497,
+      "grad_norm": 1.256387710571289,
+      "learning_rate": 2.5732540177624487e-05,
+      "loss": 0.4991,
+      "step": 630
+    },
+    {
+      "epoch": 1.4338983050847458,
+      "grad_norm": 1.09014892578125,
+      "learning_rate": 2.565400696399843e-05,
+      "loss": 0.5003,
+      "step": 635
+    },
+    {
+      "epoch": 1.4451977401129943,
+      "grad_norm": 0.9970406889915466,
+      "learning_rate": 2.557487992432665e-05,
+      "loss": 0.4976,
+      "step": 640
+    },
+    {
+      "epoch": 1.456497175141243,
+      "grad_norm": 0.9552879333496094,
+      "learning_rate": 2.5495163468940135e-05,
+      "loss": 0.4523,
+      "step": 645
+    },
+    {
+      "epoch": 1.4677966101694915,
+      "grad_norm": 1.1708117723464966,
+      "learning_rate": 2.5414862041022336e-05,
+      "loss": 0.5089,
+      "step": 650
+    },
+    {
+      "epoch": 1.4790960451977402,
+      "grad_norm": 1.2075222730636597,
+      "learning_rate": 2.533398011636153e-05,
+      "loss": 0.5079,
+      "step": 655
+    },
+    {
+      "epoch": 1.4903954802259887,
+      "grad_norm": 1.1360461711883545,
+      "learning_rate": 2.5252522203101336e-05,
+      "loss": 0.453,
+      "step": 660
+    },
+    {
+      "epoch": 1.5016949152542374,
+      "grad_norm": 1.0806200504302979,
+      "learning_rate": 2.517049284148944e-05,
+      "loss": 0.454,
+      "step": 665
+    },
+    {
+      "epoch": 1.5129943502824859,
+      "grad_norm": 1.00311279296875,
+      "learning_rate": 2.5087896603624545e-05,
+      "loss": 0.4649,
+      "step": 670
+    },
+    {
+      "epoch": 1.5242937853107343,
+      "grad_norm": 1.0430840253829956,
+      "learning_rate": 2.5004738093201534e-05,
+      "loss": 0.4552,
+      "step": 675
+    },
+    {
+      "epoch": 1.535593220338983,
+      "grad_norm": 1.133973479270935,
+      "learning_rate": 2.4921021945254872e-05,
+      "loss": 0.5055,
+      "step": 680
+    },
+    {
+      "epoch": 1.5468926553672318,
+      "grad_norm": 1.0552384853363037,
+      "learning_rate": 2.483675282590025e-05,
+      "loss": 0.4504,
+      "step": 685
+    },
+    {
+      "epoch": 1.5581920903954802,
+      "grad_norm": 1.1162137985229492,
+      "learning_rate": 2.4751935432074513e-05,
+      "loss": 0.4305,
+      "step": 690
+    },
+    {
+      "epoch": 1.5694915254237287,
+      "grad_norm": 1.1118879318237305,
+      "learning_rate": 2.4666574491273882e-05,
+      "loss": 0.4413,
+      "step": 695
+    },
+    {
+      "epoch": 1.5807909604519774,
+      "grad_norm": 1.2221858501434326,
+      "learning_rate": 2.4580674761290433e-05,
+      "loss": 0.4225,
+      "step": 700
+    },
+    {
+      "epoch": 1.5920903954802261,
+      "grad_norm": 1.1549981832504272,
+      "learning_rate": 2.4494241029946927e-05,
+      "loss": 0.4359,
+      "step": 705
+    },
+    {
+      "epoch": 1.6033898305084746,
+      "grad_norm": 1.1755397319793701,
+      "learning_rate": 2.4407278114829925e-05,
+      "loss": 0.4477,
+      "step": 710
+    },
+    {
+      "epoch": 1.614689265536723,
+      "grad_norm": 1.1148909330368042,
+      "learning_rate": 2.431979086302131e-05,
+      "loss": 0.4564,
+      "step": 715
+    },
+    {
+      "epoch": 1.6259887005649718,
+      "grad_norm": 0.9955953359603882,
+      "learning_rate": 2.4231784150828087e-05,
+      "loss": 0.4501,
+      "step": 720
+    },
+    {
+      "epoch": 1.6372881355932203,
+      "grad_norm": 1.1515029668807983,
+      "learning_rate": 2.4143262883510593e-05,
+      "loss": 0.4771,
+      "step": 725
+    },
+    {
+      "epoch": 1.6485875706214688,
+      "grad_norm": 1.4518494606018066,
+      "learning_rate": 2.405423199500912e-05,
+      "loss": 0.4262,
+      "step": 730
+    },
+    {
+      "epoch": 1.6598870056497175,
+      "grad_norm": 0.9435175657272339,
+      "learning_rate": 2.3964696447668873e-05,
+      "loss": 0.4424,
+      "step": 735
+    },
+    {
+      "epoch": 1.6711864406779662,
+      "grad_norm": 1.2055212259292603,
+      "learning_rate": 2.3874661231963422e-05,
+      "loss": 0.4333,
+      "step": 740
+    },
+    {
+      "epoch": 1.6824858757062147,
+      "grad_norm": 1.2880557775497437,
+      "learning_rate": 2.3784131366216504e-05,
+      "loss": 0.452,
+      "step": 745
+    },
+    {
+      "epoch": 1.6937853107344631,
+      "grad_norm": 1.093306303024292,
+      "learning_rate": 2.369311189632234e-05,
+      "loss": 0.4526,
+      "step": 750
+    },
+    {
+      "epoch": 1.7050847457627119,
+      "grad_norm": 1.1268608570098877,
+      "learning_rate": 2.3601607895464387e-05,
+      "loss": 0.4542,
+      "step": 755
+    },
+    {
+      "epoch": 1.7163841807909606,
+      "grad_norm": 1.1216182708740234,
+      "learning_rate": 2.350962446383258e-05,
+      "loss": 0.4287,
+      "step": 760
+    },
+    {
+      "epoch": 1.727683615819209,
+      "grad_norm": 1.0930184125900269,
+      "learning_rate": 2.3417166728339034e-05,
+      "loss": 0.3796,
+      "step": 765
+    },
+    {
+      "epoch": 1.7389830508474575,
+      "grad_norm": 1.2358158826828003,
+      "learning_rate": 2.3324239842332324e-05,
+      "loss": 0.4139,
+      "step": 770
+    },
+    {
+      "epoch": 1.7502824858757062,
+      "grad_norm": 1.2803504467010498,
+      "learning_rate": 2.32308489853102e-05,
+      "loss": 0.3922,
+      "step": 775
+    },
+    {
+      "epoch": 1.761581920903955,
+      "grad_norm": 1.0053050518035889,
+      "learning_rate": 2.3136999362630945e-05,
+      "loss": 0.3947,
+      "step": 780
+    },
+    {
+      "epoch": 1.7728813559322034,
+      "grad_norm": 1.2856465578079224,
+      "learning_rate": 2.304269620522322e-05,
+      "loss": 0.3861,
+      "step": 785
+    },
+    {
+      "epoch": 1.784180790960452,
+      "grad_norm": 0.9932456612586975,
+      "learning_rate": 2.294794476929449e-05,
+      "loss": 0.4505,
+      "step": 790
+    },
+    {
+      "epoch": 1.7954802259887006,
+      "grad_norm": 1.2482821941375732,
+      "learning_rate": 2.28527503360381e-05,
+      "loss": 0.3839,
+      "step": 795
+    },
+    {
+      "epoch": 1.8067796610169493,
+      "grad_norm": 0.9733690023422241,
+      "learning_rate": 2.2757118211338863e-05,
+      "loss": 0.387,
+      "step": 800
+    },
+    {
+      "epoch": 1.8180790960451978,
+      "grad_norm": 1.034924030303955,
+      "learning_rate": 2.2661053725477374e-05,
+      "loss": 0.413,
+      "step": 805
+    },
+    {
+      "epoch": 1.8293785310734463,
+      "grad_norm": 1.0252270698547363,
+      "learning_rate": 2.256456223283287e-05,
+      "loss": 0.3787,
+      "step": 810
+    },
+    {
+      "epoch": 1.840677966101695,
+      "grad_norm": 1.2237545251846313,
+      "learning_rate": 2.2467649111584816e-05,
+      "loss": 0.3687,
+      "step": 815
+    },
+    {
+      "epoch": 1.8519774011299435,
+      "grad_norm": 1.0817660093307495,
+      "learning_rate": 2.2370319763413157e-05,
+      "loss": 0.3826,
+      "step": 820
+    },
+    {
+      "epoch": 1.863276836158192,
+      "grad_norm": 1.099799394607544,
+      "learning_rate": 2.2272579613197183e-05,
+      "loss": 0.4199,
+      "step": 825
+    },
+    {
+      "epoch": 1.8745762711864407,
+      "grad_norm": 1.1659348011016846,
+      "learning_rate": 2.217443410871324e-05,
+      "loss": 0.3739,
+      "step": 830
+    },
+    {
+      "epoch": 1.8858757062146894,
+      "grad_norm": 1.0663833618164062,
+      "learning_rate": 2.207588872033101e-05,
+      "loss": 0.4162,
+      "step": 835
+    },
+    {
+      "epoch": 1.8971751412429378,
+      "grad_norm": 1.1548362970352173,
+      "learning_rate": 2.197694894070866e-05,
+      "loss": 0.387,
+      "step": 840
+    },
+    {
+      "epoch": 1.9084745762711863,
+      "grad_norm": 1.2925057411193848,
+      "learning_rate": 2.187762028448667e-05,
+      "loss": 0.3588,
+      "step": 845
+    },
+    {
+      "epoch": 1.919774011299435,
+      "grad_norm": 1.1187832355499268,
+      "learning_rate": 2.177790828798047e-05,
+      "loss": 0.3681,
+      "step": 850
+    },
+    {
+      "epoch": 1.9310734463276837,
+      "grad_norm": 1.1352661848068237,
+      "learning_rate": 2.1677818508871862e-05,
+      "loss": 0.3755,
+      "step": 855
+    },
+    {
+      "epoch": 1.9423728813559322,
+      "grad_norm": 1.1208044290542603,
+      "learning_rate": 2.1577356525899244e-05,
+      "loss": 0.3914,
+      "step": 860
+    },
+    {
+      "epoch": 1.9536723163841807,
+      "grad_norm": 1.1486340761184692,
+      "learning_rate": 2.1476527938546665e-05,
+      "loss": 0.3681,
+      "step": 865
+    },
+    {
+      "epoch": 1.9649717514124294,
+      "grad_norm": 1.209509253501892,
+      "learning_rate": 2.1375338366731743e-05,
+      "loss": 0.3958,
+      "step": 870
+    },
+    {
+      "epoch": 1.9762711864406781,
+      "grad_norm": 1.1091386079788208,
+      "learning_rate": 2.127379345049239e-05,
+      "loss": 0.3282,
+      "step": 875
+    },
+    {
+      "epoch": 1.9875706214689266,
+      "grad_norm": 1.0583096742630005,
+      "learning_rate": 2.117189884967248e-05,
+      "loss": 0.3461,
+      "step": 880
+    },
+    {
+      "epoch": 1.998870056497175,
+      "grad_norm": 1.161463737487793,
+      "learning_rate": 2.1069660243606384e-05,
+      "loss": 0.3753,
+      "step": 885
+    },
+    {
+      "epoch": 2.009039548022599,
+      "grad_norm": 1.1983983516693115,
+      "learning_rate": 2.096708333080241e-05,
+      "loss": 0.2699,
+      "step": 890
+    },
+    {
+      "epoch": 2.0203389830508476,
+      "grad_norm": 1.1153104305267334,
+      "learning_rate": 2.086417382862518e-05,
+      "loss": 0.2679,
+      "step": 895
+    },
+    {
+      "epoch": 2.031638418079096,
+      "grad_norm": 1.1299874782562256,
+      "learning_rate": 2.076093747297696e-05,
+      "loss": 0.3141,
+      "step": 900
+    },
+    {
+      "epoch": 2.0429378531073445,
+      "grad_norm": 1.0530294179916382,
+      "learning_rate": 2.0657380017977964e-05,
+      "loss": 0.2703,
+      "step": 905
+    },
+    {
+      "epoch": 2.0542372881355933,
+      "grad_norm": 1.1320830583572388,
+      "learning_rate": 2.0553507235645648e-05,
+      "loss": 0.2921,
+      "step": 910
+    },
+    {
+      "epoch": 2.065536723163842,
+      "grad_norm": 1.1087229251861572,
+      "learning_rate": 2.0449324915572942e-05,
+      "loss": 0.2601,
+      "step": 915
+    },
+    {
+      "epoch": 2.07683615819209,
+      "grad_norm": 1.1149791479110718,
+      "learning_rate": 2.0344838864605614e-05,
+      "loss": 0.2695,
+      "step": 920
+    },
+    {
+      "epoch": 2.088135593220339,
+      "grad_norm": 1.2066401243209839,
+      "learning_rate": 2.0240054906518574e-05,
+      "loss": 0.2857,
+      "step": 925
+    },
+    {
+      "epoch": 2.0994350282485876,
+      "grad_norm": 1.0312755107879639,
+      "learning_rate": 2.0134978881691292e-05,
+      "loss": 0.2736,
+      "step": 930
+    },
+    {
+      "epoch": 2.1107344632768363,
+      "grad_norm": 1.054076075553894,
+      "learning_rate": 2.0029616646782255e-05,
+      "loss": 0.2591,
+      "step": 935
+    },
+    {
+      "epoch": 2.1220338983050846,
+      "grad_norm": 1.2717868089675903,
+      "learning_rate": 1.9923974074402544e-05,
+      "loss": 0.2926,
+      "step": 940
+    },
+    {
+      "epoch": 2.1333333333333333,
+      "grad_norm": 1.154904842376709,
+      "learning_rate": 1.981805705278849e-05,
+      "loss": 0.3017,
+      "step": 945
+    },
+    {
+      "epoch": 2.144632768361582,
+      "grad_norm": 1.2696698904037476,
+      "learning_rate": 1.9711871485473523e-05,
+      "loss": 0.2602,
+      "step": 950
+    },
+    {
+      "epoch": 2.1559322033898307,
+      "grad_norm": 1.1353667974472046,
+      "learning_rate": 1.9605423290959077e-05,
+      "loss": 0.278,
+      "step": 955
+    },
+    {
+      "epoch": 2.167231638418079,
+      "grad_norm": 1.2267999649047852,
+      "learning_rate": 1.9498718402384732e-05,
+      "loss": 0.2874,
+      "step": 960
+    },
+    {
+      "epoch": 2.1785310734463277,
+      "grad_norm": 1.1276496648788452,
+      "learning_rate": 1.9391762767197516e-05,
+      "loss": 0.2615,
+      "step": 965
+    },
+    {
+      "epoch": 2.1898305084745764,
+      "grad_norm": 1.0736314058303833,
+      "learning_rate": 1.928456234682041e-05,
+      "loss": 0.2785,
+      "step": 970
+    },
+    {
+      "epoch": 2.201129943502825,
+      "grad_norm": 1.2606271505355835,
+      "learning_rate": 1.917712311632006e-05,
+      "loss": 0.2556,
+      "step": 975
+    },
+    {
+      "epoch": 2.2124293785310734,
+      "grad_norm": 1.205222249031067,
+      "learning_rate": 1.9069451064073764e-05,
+      "loss": 0.289,
+      "step": 980
+    },
+    {
+      "epoch": 2.223728813559322,
+      "grad_norm": 1.2200062274932861,
+      "learning_rate": 1.8961552191435684e-05,
+      "loss": 0.2479,
+      "step": 985
+    },
+    {
+      "epoch": 2.2350282485875708,
+      "grad_norm": 0.9834157228469849,
+      "learning_rate": 1.8853432512402348e-05,
+      "loss": 0.2476,
+      "step": 990
+    },
+    {
+      "epoch": 2.246327683615819,
+      "grad_norm": 1.0837196111679077,
+      "learning_rate": 1.874509805327744e-05,
+      "loss": 0.2739,
+      "step": 995
+    },
+    {
+      "epoch": 2.2576271186440677,
+      "grad_norm": 1.0946136713027954,
+      "learning_rate": 1.863655485233592e-05,
+      "loss": 0.2474,
+      "step": 1000
+    },
+    {
+      "epoch": 2.2689265536723164,
+      "grad_norm": 1.2106841802597046,
+      "learning_rate": 1.8527808959487468e-05,
+      "loss": 0.2697,
+      "step": 1005
+    },
+    {
+      "epoch": 2.280225988700565,
+      "grad_norm": 1.2549164295196533,
+      "learning_rate": 1.8418866435939267e-05,
+      "loss": 0.2757,
+      "step": 1010
+    },
+    {
+      "epoch": 2.2915254237288134,
+      "grad_norm": 1.1625417470932007,
+      "learning_rate": 1.8309733353858167e-05,
+      "loss": 0.257,
+      "step": 1015
+    },
+    {
+      "epoch": 2.302824858757062,
+      "grad_norm": 1.1273975372314453,
+      "learning_rate": 1.8200415796032254e-05,
+      "loss": 0.255,
+      "step": 1020
+    },
+    {
+      "epoch": 2.314124293785311,
+      "grad_norm": 1.0395770072937012,
+      "learning_rate": 1.80909198555318e-05,
+      "loss": 0.2498,
+      "step": 1025
+    },
+    {
+      "epoch": 2.3254237288135595,
+      "grad_norm": 1.2258102893829346,
+      "learning_rate": 1.798125163536965e-05,
+      "loss": 0.2448,
+      "step": 1030
+    },
+    {
+      "epoch": 2.336723163841808,
+      "grad_norm": 1.081221103668213,
+      "learning_rate": 1.787141724816106e-05,
+      "loss": 0.2508,
+      "step": 1035
+    },
+    {
+      "epoch": 2.3480225988700565,
+      "grad_norm": 1.3911811113357544,
+      "learning_rate": 1.7761422815783e-05,
+      "loss": 0.2542,
+      "step": 1040
+    },
+    {
+      "epoch": 2.359322033898305,
+      "grad_norm": 1.5137418508529663,
+      "learning_rate": 1.765127446903293e-05,
+      "loss": 0.2536,
+      "step": 1045
+    },
+    {
+      "epoch": 2.3706214689265535,
+      "grad_norm": 1.079277515411377,
+      "learning_rate": 1.754097834728708e-05,
+      "loss": 0.2452,
+      "step": 1050
+    },
+    {
+      "epoch": 2.381920903954802,
+      "grad_norm": 1.3662875890731812,
+      "learning_rate": 1.7430540598158287e-05,
+      "loss": 0.2538,
+      "step": 1055
+    },
+    {
+      "epoch": 2.393220338983051,
+      "grad_norm": 1.214477777481079,
+      "learning_rate": 1.7319967377153295e-05,
+      "loss": 0.2698,
+      "step": 1060
+    },
+    {
+      "epoch": 2.4045197740112996,
+      "grad_norm": 1.1170893907546997,
+      "learning_rate": 1.7209264847329717e-05,
+      "loss": 0.2284,
+      "step": 1065
+    },
+    {
+      "epoch": 2.415819209039548,
+      "grad_norm": 1.0510598421096802,
+      "learning_rate": 1.709843917895247e-05,
+      "loss": 0.2073,
+      "step": 1070
+    },
+    {
+      "epoch": 2.4271186440677965,
+      "grad_norm": 1.0740697383880615,
+      "learning_rate": 1.698749654914991e-05,
+      "loss": 0.2402,
+      "step": 1075
+    },
+    {
+      "epoch": 2.4384180790960452,
+      "grad_norm": 1.167296051979065,
+      "learning_rate": 1.6876443141569503e-05,
+      "loss": 0.2535,
+      "step": 1080
+    },
+    {
+      "epoch": 2.449717514124294,
+      "grad_norm": 1.1952712535858154,
+      "learning_rate": 1.6765285146033165e-05,
+      "loss": 0.2423,
+      "step": 1085
+    },
+    {
+      "epoch": 2.461016949152542,
+      "grad_norm": 0.95307457447052,
+      "learning_rate": 1.6654028758192286e-05,
+      "loss": 0.2457,
+      "step": 1090
+    },
+    {
+      "epoch": 2.472316384180791,
+      "grad_norm": 1.0912442207336426,
+      "learning_rate": 1.654268017918236e-05,
+      "loss": 0.2326,
+      "step": 1095
+    },
+    {
+      "epoch": 2.4836158192090396,
+      "grad_norm": 1.280784010887146,
+      "learning_rate": 1.643124561527739e-05,
+      "loss": 0.2414,
+      "step": 1100
+    },
+    {
+      "epoch": 2.4949152542372883,
+      "grad_norm": 1.0802363157272339,
+      "learning_rate": 1.631973127754394e-05,
+      "loss": 0.2236,
+      "step": 1105
+    },
+    {
+      "epoch": 2.5062146892655366,
+      "grad_norm": 1.074249267578125,
+      "learning_rate": 1.6208143381494955e-05,
+      "loss": 0.2317,
+      "step": 1110
+    },
+    {
+      "epoch": 2.5175141242937853,
+      "grad_norm": 1.433241844177246,
+      "learning_rate": 1.609648814674332e-05,
+      "loss": 0.2381,
+      "step": 1115
+    },
+    {
+      "epoch": 2.528813559322034,
+      "grad_norm": 1.030909776687622,
+      "learning_rate": 1.5984771796655223e-05,
+      "loss": 0.2604,
+      "step": 1120
+    },
+    {
+      "epoch": 2.5401129943502827,
+      "grad_norm": 1.213474988937378,
+      "learning_rate": 1.5873000558003227e-05,
+      "loss": 0.2453,
+      "step": 1125
+    },
+    {
+      "epoch": 2.551412429378531,
+      "grad_norm": 1.3832768201828003,
+      "learning_rate": 1.5761180660619267e-05,
+      "loss": 0.2381,
+      "step": 1130
+    },
+    {
+      "epoch": 2.5627118644067797,
+      "grad_norm": 1.210844874382019,
+      "learning_rate": 1.5649318337047365e-05,
+      "loss": 0.2279,
+      "step": 1135
+    },
+    {
+      "epoch": 2.5740112994350284,
+      "grad_norm": 1.2258585691452026,
+      "learning_rate": 1.553741982219627e-05,
+      "loss": 0.2251,
+      "step": 1140
+    },
+    {
+      "epoch": 2.5853107344632766,
+      "grad_norm": 1.11599862575531,
+      "learning_rate": 1.5425491352991947e-05,
+      "loss": 0.2224,
+      "step": 1145
+    },
+    {
+      "epoch": 2.5966101694915253,
+      "grad_norm": 1.0045859813690186,
+      "learning_rate": 1.5313539168029932e-05,
+      "loss": 0.213,
+      "step": 1150
+    },
+    {
+      "epoch": 2.607909604519774,
+      "grad_norm": 0.8923414349555969,
+      "learning_rate": 1.5201569507227609e-05,
+      "loss": 0.2049,
+      "step": 1155
+    },
+    {
+      "epoch": 2.6192090395480228,
+      "grad_norm": 1.2115402221679688,
+      "learning_rate": 1.5089588611476426e-05,
+      "loss": 0.2118,
+      "step": 1160
+    },
+    {
+      "epoch": 2.6305084745762715,
+      "grad_norm": 1.0908693075180054,
+      "learning_rate": 1.497760272229403e-05,
+      "loss": 0.25,
+      "step": 1165
+    },
+    {
+      "epoch": 2.6418079096045197,
+      "grad_norm": 1.0529178380966187,
+      "learning_rate": 1.4865618081476394e-05,
+      "loss": 0.2034,
+      "step": 1170
+    },
+    {
+      "epoch": 2.6531073446327684,
+      "grad_norm": 1.1474601030349731,
+      "learning_rate": 1.475364093074991e-05,
+      "loss": 0.224,
+      "step": 1175
+    },
+    {
+      "epoch": 2.664406779661017,
+      "grad_norm": 1.0988585948944092,
+      "learning_rate": 1.4641677511423497e-05,
+      "loss": 0.2367,
+      "step": 1180
+    },
+    {
+      "epoch": 2.6757062146892654,
+      "grad_norm": 1.174198031425476,
+      "learning_rate": 1.4529734064040704e-05,
+      "loss": 0.229,
+      "step": 1185
+    },
+    {
+      "epoch": 2.687005649717514,
+      "grad_norm": 1.0259079933166504,
+      "learning_rate": 1.4417816828031916e-05,
+      "loss": 0.2048,
+      "step": 1190
+    },
+    {
+      "epoch": 2.698305084745763,
+      "grad_norm": 1.0843926668167114,
+      "learning_rate": 1.4305932041366566e-05,
+      "loss": 0.2094,
+      "step": 1195
+    },
+    {
+      "epoch": 2.709604519774011,
+      "grad_norm": 1.0300178527832031,
+      "learning_rate": 1.4194085940205427e-05,
+      "loss": 0.2288,
+      "step": 1200
+    },
+    {
+      "epoch": 2.7209039548022598,
+      "grad_norm": 1.406386375427246,
+      "learning_rate": 1.4082284758553067e-05,
+      "loss": 0.2153,
+      "step": 1205
+    },
+    {
+      "epoch": 2.7322033898305085,
+      "grad_norm": 1.2302467823028564,
+      "learning_rate": 1.3970534727910346e-05,
+      "loss": 0.219,
+      "step": 1210
+    },
+    {
+      "epoch": 2.743502824858757,
+      "grad_norm": 1.163407564163208,
+      "learning_rate": 1.3858842076927114e-05,
+      "loss": 0.2121,
+      "step": 1215
+    },
+    {
+      "epoch": 2.754802259887006,
+      "grad_norm": 1.185469150543213,
+      "learning_rate": 1.3747213031055015e-05,
+      "loss": 0.215,
+      "step": 1220
+    },
+    {
+      "epoch": 2.766101694915254,
+      "grad_norm": 1.2026503086090088,
+      "learning_rate": 1.363565381220053e-05,
+      "loss": 0.2025,
+      "step": 1225
+    },
+    {
+      "epoch": 2.777401129943503,
+      "grad_norm": 1.0308500528335571,
+      "learning_rate": 1.352417063837816e-05,
+      "loss": 0.1976,
+      "step": 1230
+    },
+    {
+      "epoch": 2.7887005649717516,
+      "grad_norm": 1.136925220489502,
+      "learning_rate": 1.3412769723363848e-05,
+      "loss": 0.2244,
+      "step": 1235
+    },
+    {
+      "epoch": 2.8,
+      "grad_norm": 1.0420323610305786,
+      "learning_rate": 1.3301457276348668e-05,
+      "loss": 0.216,
+      "step": 1240
+    },
+    {
+      "epoch": 2.8112994350282485,
+      "grad_norm": 1.2677199840545654,
+      "learning_rate": 1.3190239501592721e-05,
+      "loss": 0.2261,
+      "step": 1245
+    },
+    {
+      "epoch": 2.8225988700564972,
+      "grad_norm": 1.1373021602630615,
+      "learning_rate": 1.3079122598079307e-05,
+      "loss": 0.2101,
+      "step": 1250
+    },
+    {
+      "epoch": 2.8338983050847455,
+      "grad_norm": 1.2039943933486938,
+      "learning_rate": 1.2968112759169449e-05,
+      "loss": 0.1857,
+      "step": 1255
+    },
+    {
+      "epoch": 2.845197740112994,
+      "grad_norm": 1.2662607431411743,
+      "learning_rate": 1.285721617225668e-05,
+      "loss": 0.1921,
+      "step": 1260
+    },
+    {
+      "epoch": 2.856497175141243,
+      "grad_norm": 1.051405668258667,
+      "learning_rate": 1.2746439018422142e-05,
+      "loss": 0.1874,
+      "step": 1265
+    },
+    {
+      "epoch": 2.8677966101694916,
+      "grad_norm": 1.2025455236434937,
+      "learning_rate": 1.2635787472090115e-05,
+      "loss": 0.2216,
+      "step": 1270
+    },
+    {
+      "epoch": 2.8790960451977403,
+      "grad_norm": 1.1472687721252441,
+      "learning_rate": 1.2525267700683846e-05,
+      "loss": 0.1972,
+      "step": 1275
+    },
+    {
+      "epoch": 2.8903954802259886,
+      "grad_norm": 1.0994764566421509,
+      "learning_rate": 1.2414885864281804e-05,
+      "loss": 0.1978,
+      "step": 1280
+    },
+    {
+      "epoch": 2.9016949152542373,
+      "grad_norm": 1.150343894958496,
+      "learning_rate": 1.2304648115274303e-05,
+      "loss": 0.2111,
+      "step": 1285
+    },
+    {
+      "epoch": 2.912994350282486,
+      "grad_norm": 1.064850091934204,
+      "learning_rate": 1.2194560598020637e-05,
+      "loss": 0.2019,
+      "step": 1290
+    },
+    {
+      "epoch": 2.9242937853107343,
+      "grad_norm": 1.053019642829895,
+      "learning_rate": 1.2084629448506568e-05,
+      "loss": 0.1839,
+      "step": 1295
+    },
+    {
+      "epoch": 2.935593220338983,
+      "grad_norm": 1.14827561378479,
+      "learning_rate": 1.1974860794002337e-05,
+      "loss": 0.1863,
+      "step": 1300
+    },
+    {
+      "epoch": 2.9468926553672317,
+      "grad_norm": 1.2147302627563477,
+      "learning_rate": 1.1865260752721141e-05,
+      "loss": 0.1949,
+      "step": 1305
+    },
+    {
+      "epoch": 2.9581920903954804,
+      "grad_norm": 1.1208956241607666,
+      "learning_rate": 1.1755835433478136e-05,
+      "loss": 0.1789,
+      "step": 1310
+    },
+    {
+      "epoch": 2.969491525423729,
+      "grad_norm": 0.9414421319961548,
+      "learning_rate": 1.1646590935349917e-05,
+      "loss": 0.1913,
+      "step": 1315
+    },
+    {
+      "epoch": 2.9807909604519773,
+      "grad_norm": 1.3583221435546875,
+      "learning_rate": 1.1537533347334606e-05,
+      "loss": 0.1887,
+      "step": 1320
+    },
+    {
+      "epoch": 2.992090395480226,
+      "grad_norm": 1.1631615161895752,
+      "learning_rate": 1.1428668748012453e-05,
+      "loss": 0.1915,
+      "step": 1325
+    },
+    {
+      "epoch": 3.00225988700565,
+      "grad_norm": 0.91972815990448,
+      "learning_rate": 1.1320003205207011e-05,
+      "loss": 0.1953,
+      "step": 1330
+    },
+    {
+      "epoch": 3.013559322033898,
+      "grad_norm": 1.2306082248687744,
+      "learning_rate": 1.1211542775646973e-05,
+      "loss": 0.1453,
+      "step": 1335
+    },
+    {
+      "epoch": 3.024858757062147,
+      "grad_norm": 0.9653990268707275,
+      "learning_rate": 1.1103293504628554e-05,
+      "loss": 0.1736,
+      "step": 1340
+    },
+    {
+      "epoch": 3.0361581920903955,
+      "grad_norm": 1.0750000476837158,
+      "learning_rate": 1.0995261425678567e-05,
+      "loss": 0.1481,
+      "step": 1345
+    },
+    {
+      "epoch": 3.047457627118644,
+      "grad_norm": 0.9752047657966614,
+      "learning_rate": 1.0887452560218087e-05,
+      "loss": 0.1234,
+      "step": 1350
+    },
+    {
+      "epoch": 3.0587570621468925,
+      "grad_norm": 0.9885937571525574,
+      "learning_rate": 1.0779872917226888e-05,
+      "loss": 0.1517,
+      "step": 1355
+    },
+    {
+      "epoch": 3.070056497175141,
+      "grad_norm": 1.0810526609420776,
+      "learning_rate": 1.0672528492908489e-05,
+      "loss": 0.1395,
+      "step": 1360
+    },
+    {
+      "epoch": 3.08135593220339,
+      "grad_norm": 1.0768548250198364,
+      "learning_rate": 1.056542527035593e-05,
+      "loss": 0.1245,
+      "step": 1365
+    },
+    {
+      "epoch": 3.0926553672316386,
+      "grad_norm": 1.1355918645858765,
+      "learning_rate": 1.0458569219218324e-05,
+      "loss": 0.1428,
+      "step": 1370
+    },
+    {
+      "epoch": 3.103954802259887,
+      "grad_norm": 1.1265407800674438,
+      "learning_rate": 1.0351966295368097e-05,
+      "loss": 0.1286,
+      "step": 1375
+    },
+    {
+      "epoch": 3.1152542372881356,
+      "grad_norm": 1.0180362462997437,
+      "learning_rate": 1.0245622440569028e-05,
+      "loss": 0.1263,
+      "step": 1380
+    },
+    {
+      "epoch": 3.1265536723163843,
+      "grad_norm": 1.0521749258041382,
+      "learning_rate": 1.0139543582145081e-05,
+      "loss": 0.1352,
+      "step": 1385
+    },
+    {
+      "epoch": 3.137853107344633,
+      "grad_norm": 1.0448542833328247,
+      "learning_rate": 1.0033735632650032e-05,
+      "loss": 0.126,
+      "step": 1390
+    },
+    {
+      "epoch": 3.1491525423728812,
+      "grad_norm": 0.9419913291931152,
+      "learning_rate": 9.9282044895379e-06,
+      "loss": 0.1268,
+      "step": 1395
+    },
+    {
+      "epoch": 3.16045197740113,
+      "grad_norm": 0.9798169136047363,
+      "learning_rate": 9.822956034834256e-06,
+      "loss": 0.1409,
+      "step": 1400
+    },
+    {
+      "epoch": 3.1717514124293786,
+      "grad_norm": 1.1786912679672241,
+      "learning_rate": 9.717996134808373e-06,
+      "loss": 0.134,
+      "step": 1405
+    },
+    {
+      "epoch": 3.1830508474576273,
+      "grad_norm": 1.0761826038360596,
+      "learning_rate": 9.613330639646263e-06,
+      "loss": 0.1275,
+      "step": 1410
+    },
+    {
+      "epoch": 3.1943502824858756,
+      "grad_norm": 1.2219995260238647,
+      "learning_rate": 9.508965383124567e-06,
+      "loss": 0.136,
+      "step": 1415
+    },
+    {
+      "epoch": 3.2056497175141243,
+      "grad_norm": 0.9917460083961487,
+      "learning_rate": 9.40490618228544e-06,
+      "loss": 0.1236,
+      "step": 1420
+    },
+    {
+      "epoch": 3.216949152542373,
+      "grad_norm": 1.1220533847808838,
+      "learning_rate": 9.301158837112308e-06,
+      "loss": 0.1397,
+      "step": 1425
+    },
+    {
+      "epoch": 3.2282485875706213,
+      "grad_norm": 0.9562025666236877,
+      "learning_rate": 9.197729130206582e-06,
+      "loss": 0.1147,
+      "step": 1430
+    },
+    {
+      "epoch": 3.23954802259887,
+      "grad_norm": 1.0364519357681274,
+      "learning_rate": 9.09462282646536e-06,
+      "loss": 0.1302,
+      "step": 1435
+    },
+    {
+      "epoch": 3.2508474576271187,
+      "grad_norm": 1.0610555410385132,
+      "learning_rate": 8.991845672760119e-06,
+      "loss": 0.1357,
+      "step": 1440
+    },
+    {
+      "epoch": 3.2621468926553674,
+      "grad_norm": 0.9001578688621521,
+      "learning_rate": 8.889403397616364e-06,
+      "loss": 0.1244,
+      "step": 1445
+    },
+    {
+      "epoch": 3.2734463276836157,
+      "grad_norm": 1.051056981086731,
+      "learning_rate": 8.787301710894386e-06,
+      "loss": 0.1493,
+      "step": 1450
+    },
+    {
+      "epoch": 3.2847457627118644,
+      "grad_norm": 1.0633341073989868,
+      "learning_rate": 8.68554630347097e-06,
+      "loss": 0.136,
+      "step": 1455
+    },
+    {
+      "epoch": 3.296045197740113,
+      "grad_norm": 0.9047091007232666,
+      "learning_rate": 8.584142846922199e-06,
+      "loss": 0.1278,
+      "step": 1460
+    },
+    {
+      "epoch": 3.3073446327683618,
+      "grad_norm": 1.1870874166488647,
+      "learning_rate": 8.483096993207367e-06,
+      "loss": 0.1577,
+      "step": 1465
+    },
+    {
+      "epoch": 3.31864406779661,
+      "grad_norm": 0.974368691444397,
+      "learning_rate": 8.382414374353936e-06,
+      "loss": 0.1193,
+      "step": 1470
+    },
+    {
+      "epoch": 3.3299435028248587,
+      "grad_norm": 0.9335852861404419,
+      "learning_rate": 8.282100602143613e-06,
+      "loss": 0.1305,
+      "step": 1475
+    },
+    {
+      "epoch": 3.3412429378531074,
+      "grad_norm": 1.0982964038848877,
+      "learning_rate": 8.182161267799576e-06,
+      "loss": 0.1542,
+      "step": 1480
+    },
+    {
+      "epoch": 3.3525423728813557,
+      "grad_norm": 0.9854334592819214,
+      "learning_rate": 8.082601941674834e-06,
+      "loss": 0.1339,
+      "step": 1485
+    },
+    {
+      "epoch": 3.3638418079096044,
+      "grad_norm": 1.1188992261886597,
+      "learning_rate": 7.98342817294175e-06,
+      "loss": 0.1304,
+      "step": 1490
+    },
+    {
+      "epoch": 3.375141242937853,
+      "grad_norm": 1.0093649625778198,
+      "learning_rate": 7.884645489282726e-06,
+      "loss": 0.1417,
+      "step": 1495
+    },
+    {
+      "epoch": 3.386440677966102,
+      "grad_norm": 1.1569143533706665,
+      "learning_rate": 7.786259396582139e-06,
+      "loss": 0.1394,
+      "step": 1500
+    },
+    {
+      "epoch": 3.3977401129943505,
+      "grad_norm": 0.9898454546928406,
+      "learning_rate": 7.688275378619437e-06,
+      "loss": 0.1414,
+      "step": 1505
+    },
+    {
+      "epoch": 3.409039548022599,
+      "grad_norm": 1.0313230752944946,
+      "learning_rate": 7.5906988967634725e-06,
+      "loss": 0.1323,
+      "step": 1510
+    },
+    {
+      "epoch": 3.4203389830508475,
+      "grad_norm": 0.9393590092658997,
+      "learning_rate": 7.493535389668144e-06,
+      "loss": 0.1376,
+      "step": 1515
+    },
+    {
+      "epoch": 3.431638418079096,
+      "grad_norm": 1.0178027153015137,
+      "learning_rate": 7.3967902729692195e-06,
+      "loss": 0.128,
+      "step": 1520
+    },
+    {
+      "epoch": 3.4429378531073445,
+      "grad_norm": 0.9743438959121704,
+      "learning_rate": 7.300468938982493e-06,
+      "loss": 0.1339,
+      "step": 1525
+    },
+    {
+      "epoch": 3.454237288135593,
+      "grad_norm": 0.8919374346733093,
+      "learning_rate": 7.204576756403246e-06,
+      "loss": 0.1135,
+      "step": 1530
+    },
+    {
+      "epoch": 3.465536723163842,
+      "grad_norm": 1.075640320777893,
+      "learning_rate": 7.109119070006999e-06,
+      "loss": 0.1303,
+      "step": 1535
+    },
+    {
+      "epoch": 3.4768361581920906,
+      "grad_norm": 1.0930207967758179,
+      "learning_rate": 7.014101200351612e-06,
+      "loss": 0.1336,
+      "step": 1540
+    },
+    {
+      "epoch": 3.488135593220339,
+      "grad_norm": 0.9221169352531433,
+      "learning_rate": 6.919528443480715e-06,
+      "loss": 0.123,
+      "step": 1545
+    },
+    {
+      "epoch": 3.4994350282485875,
+      "grad_norm": 1.0594903230667114,
+      "learning_rate": 6.8254060706285435e-06,
+      "loss": 0.1238,
+      "step": 1550
+    },
+    {
+      "epoch": 3.5107344632768362,
+      "grad_norm": 1.0262291431427002,
+      "learning_rate": 6.7317393279261334e-06,
+      "loss": 0.1111,
+      "step": 1555
+    },
+    {
+      "epoch": 3.522033898305085,
+      "grad_norm": 1.030395269393921,
+      "learning_rate": 6.638533436108885e-06,
+      "loss": 0.1273,
+      "step": 1560
+    },
+    {
+      "epoch": 3.533333333333333,
+      "grad_norm": 1.1622343063354492,
+      "learning_rate": 6.54579359022561e-06,
+      "loss": 0.151,
+      "step": 1565
+    },
+    {
+      "epoch": 3.544632768361582,
+      "grad_norm": 0.9246652722358704,
+      "learning_rate": 6.4535249593489595e-06,
+      "loss": 0.1195,
+      "step": 1570
+    },
+    {
+      "epoch": 3.5559322033898306,
+      "grad_norm": 1.036953330039978,
+      "learning_rate": 6.3617326862872956e-06,
+      "loss": 0.1031,
+      "step": 1575
+    },
+    {
+      "epoch": 3.567231638418079,
+      "grad_norm": 0.9925161004066467,
+      "learning_rate": 6.270421887298079e-06,
+      "loss": 0.1188,
+      "step": 1580
+    },
+    {
+      "epoch": 3.5785310734463276,
+      "grad_norm": 1.0026724338531494,
+      "learning_rate": 6.17959765180268e-06,
+      "loss": 0.1198,
+      "step": 1585
+    },
+    {
+      "epoch": 3.5898305084745763,
+      "grad_norm": 1.069403052330017,
+      "learning_rate": 6.0892650421027005e-06,
+      "loss": 0.1313,
+      "step": 1590
+    },
+    {
+      "epoch": 3.601129943502825,
+      "grad_norm": 1.0291143655776978,
+      "learning_rate": 5.999429093097844e-06,
+      "loss": 0.1168,
+      "step": 1595
+    },
+    {
+      "epoch": 3.6124293785310737,
+      "grad_norm": 1.038604974746704,
+      "learning_rate": 5.910094812005254e-06,
+      "loss": 0.119,
+      "step": 1600
+    },
+    {
+      "epoch": 3.623728813559322,
+      "grad_norm": 0.943396806716919,
+      "learning_rate": 5.8212671780804485e-06,
+      "loss": 0.1237,
+      "step": 1605
+    },
+    {
+      "epoch": 3.6350282485875707,
+      "grad_norm": 0.9925442337989807,
+      "learning_rate": 5.732951142339759e-06,
+      "loss": 0.1234,
+      "step": 1610
+    },
+    {
+      "epoch": 3.6463276836158194,
+      "grad_norm": 0.9433180689811707,
+      "learning_rate": 5.645151627284415e-06,
+      "loss": 0.1287,
+      "step": 1615
+    },
+    {
+      "epoch": 3.6576271186440676,
+      "grad_norm": 1.0339361429214478,
+      "learning_rate": 5.557873526626141e-06,
+      "loss": 0.1249,
+      "step": 1620
+    },
+    {
+      "epoch": 3.6689265536723163,
+      "grad_norm": 0.8291221261024475,
+      "learning_rate": 5.471121705014415e-06,
+      "loss": 0.1192,
+      "step": 1625
+    },
+    {
+      "epoch": 3.680225988700565,
+      "grad_norm": 0.9069194197654724,
+      "learning_rate": 5.384900997765316e-06,
+      "loss": 0.114,
+      "step": 1630
+    },
+    {
+      "epoch": 3.6915254237288133,
+      "grad_norm": 0.8561278581619263,
+      "learning_rate": 5.299216210592026e-06,
+      "loss": 0.1182,
+      "step": 1635
+    },
+    {
+      "epoch": 3.702824858757062,
+      "grad_norm": 1.0266520977020264,
+      "learning_rate": 5.214072119336951e-06,
+      "loss": 0.1078,
+      "step": 1640
+    },
+    {
+      "epoch": 3.7141242937853107,
+      "grad_norm": 0.87870192527771,
+      "learning_rate": 5.129473469705553e-06,
+      "loss": 0.1076,
+      "step": 1645
+    },
+    {
+      "epoch": 3.7254237288135594,
+      "grad_norm": 0.9351235628128052,
+      "learning_rate": 5.045424977001833e-06,
+      "loss": 0.1163,
+      "step": 1650
+    },
+    {
+      "epoch": 3.736723163841808,
+      "grad_norm": 0.8863831758499146,
+      "learning_rate": 4.961931325865489e-06,
+      "loss": 0.1034,
+      "step": 1655
+    },
+    {
+      "epoch": 3.7480225988700564,
+      "grad_norm": 1.0970642566680908,
+      "learning_rate": 4.878997170010838e-06,
+      "loss": 0.1111,
+      "step": 1660
+    },
+    {
+      "epoch": 3.759322033898305,
+      "grad_norm": 1.004388451576233,
+      "learning_rate": 4.7966271319674135e-06,
+      "loss": 0.1067,
+      "step": 1665
+    },
+    {
+      "epoch": 3.770621468926554,
+      "grad_norm": 0.9965023398399353,
+      "learning_rate": 4.714825802822321e-06,
+      "loss": 0.1201,
+      "step": 1670
+    },
+    {
+      "epoch": 3.781920903954802,
+      "grad_norm": 0.799984335899353,
+      "learning_rate": 4.633597741964335e-06,
+      "loss": 0.1122,
+      "step": 1675
+    },
+    {
+      "epoch": 3.7932203389830508,
+      "grad_norm": 0.9405682682991028,
+      "learning_rate": 4.552947476829788e-06,
+      "loss": 0.1118,
+      "step": 1680
+    },
+    {
+      "epoch": 3.8045197740112995,
+      "grad_norm": 1.1903618574142456,
+      "learning_rate": 4.472879502650216e-06,
+      "loss": 0.1103,
+      "step": 1685
+    },
+    {
+      "epoch": 3.815819209039548,
+      "grad_norm": 0.8348179459571838,
+      "learning_rate": 4.393398282201788e-06,
+      "loss": 0.1175,
+      "step": 1690
+    },
+    {
+      "epoch": 3.8271186440677964,
+      "grad_norm": 0.9724533557891846,
+      "learning_rate": 4.314508245556596e-06,
+      "loss": 0.1068,
+      "step": 1695
+    },
+    {
+      "epoch": 3.838418079096045,
+      "grad_norm": 0.8165696263313293,
+      "learning_rate": 4.2362137898357125e-06,
+      "loss": 0.1139,
+      "step": 1700
+    },
+    {
+      "epoch": 3.849717514124294,
+      "grad_norm": 0.9819647073745728,
+      "learning_rate": 4.158519278964104e-06,
+      "loss": 0.107,
+      "step": 1705
+    },
+    {
+      "epoch": 3.8610169491525426,
+      "grad_norm": 0.9945287108421326,
+      "learning_rate": 4.081429043427413e-06,
+      "loss": 0.1107,
+      "step": 1710
+    },
+    {
+      "epoch": 3.872316384180791,
+      "grad_norm": 1.185699224472046,
+      "learning_rate": 4.00494738003058e-06,
+      "loss": 0.0984,
+      "step": 1715
+    },
+    {
+      "epoch": 3.8836158192090395,
+      "grad_norm": 0.8769948482513428,
+      "learning_rate": 3.929078551658344e-06,
+      "loss": 0.0965,
+      "step": 1720
+    },
+    {
+      "epoch": 3.8949152542372882,
+      "grad_norm": 0.8651741147041321,
+      "learning_rate": 3.853826787037652e-06,
+      "loss": 0.1134,
+      "step": 1725
+    },
+    {
+      "epoch": 3.9062146892655365,
+      "grad_norm": 0.9241393804550171,
+      "learning_rate": 3.779196280501962e-06,
+      "loss": 0.0964,
+      "step": 1730
+    },
+    {
+      "epoch": 3.917514124293785,
+      "grad_norm": 0.8641093373298645,
+      "learning_rate": 3.705191191757453e-06,
+      "loss": 0.1074,
+      "step": 1735
+    },
+    {
+      "epoch": 3.928813559322034,
+      "grad_norm": 0.8302857279777527,
+      "learning_rate": 3.631815645651168e-06,
+      "loss": 0.1035,
+      "step": 1740
+    },
+    {
+      "epoch": 3.9401129943502826,
+      "grad_norm": 0.9152923822402954,
+      "learning_rate": 3.5590737319411303e-06,
+      "loss": 0.1026,
+      "step": 1745
+    },
+    {
+      "epoch": 3.9514124293785313,
+      "grad_norm": 0.8876755237579346,
+      "learning_rate": 3.4869695050683742e-06,
+      "loss": 0.1122,
+      "step": 1750
+    },
+    {
+      "epoch": 3.9627118644067796,
+      "grad_norm": 0.9697118997573853,
+      "learning_rate": 3.415506983930952e-06,
+      "loss": 0.1121,
+      "step": 1755
+    },
+    {
+      "epoch": 3.9740112994350283,
+      "grad_norm": 0.998383104801178,
+      "learning_rate": 3.344690151659954e-06,
+      "loss": 0.1102,
+      "step": 1760
+    },
+    {
+      "epoch": 3.985310734463277,
+      "grad_norm": 0.9575672745704651,
+      "learning_rate": 3.2745229553974856e-06,
+      "loss": 0.0971,
+      "step": 1765
+    },
+    {
+      "epoch": 3.9966101694915253,
+      "grad_norm": 0.8986596465110779,
+      "learning_rate": 3.2050093060766695e-06,
+      "loss": 0.1117,
+      "step": 1770
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 2215,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 2000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.558445157934432e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

37_128_e5_3e-5/checkpoint-1772/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c2fc91f88ecbc07e44fa9bd105383aeb3bf6c3739d7fa469fb2915f1ea1cc71e
+size 7736

37_128_e5_3e-5/checkpoint-1772/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)

37_128_e5_3e-5/checkpoint-2215/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

37_128_e5_3e-5/checkpoint-2215/adapter_config.json

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

37_128_e5_3e-5/checkpoint-2215/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dd4c8aaaedd4cdd961fe4f7d9e985db59633f95fda60f91e558c3f09e38bb811
+size 791751704

37_128_e5_3e-5/checkpoint-2215/latest

@@ -0,0 +1 @@
+global_step2215

37_128_e5_3e-5/checkpoint-2215/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e530996bcc34dfd48bd46e49b1775c368325503d2c810e96e426fcac8df0ea5
+size 15920

37_128_e5_3e-5/checkpoint-2215/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc690fbcd10659815021efcccf1d2bab8128fad8917d58a83b8fae4a5a6c70f7
+size 15920

37_128_e5_3e-5/checkpoint-2215/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8cbca5bbd8a372bc967dcd94148ef11de0625d2c8cf776b2738fea89f4432818
+size 15920