Upload folder using huggingface_hub

49_128_e5_3e-5/checkpoint-1116/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

49_128_e5_3e-5/checkpoint-1116/adapter_config.json

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

49_128_e5_3e-5/checkpoint-1116/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca93b0614566db9baa59fc8993f125675800f0c3d236ee81a7899b2ce208a0d4
+size 791751704

49_128_e5_3e-5/checkpoint-1116/latest

@@ -0,0 +1 @@
+global_step1116

49_128_e5_3e-5/checkpoint-1116/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30b590cfd05ff6749e750bd00dc550722b425103c929b10cf9b6067ab419949d
+size 15920

49_128_e5_3e-5/checkpoint-1116/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1395f3e18a4db8eee65dbf030c660a415ed11ffc846b07370999e5613ed803d3
+size 15920

49_128_e5_3e-5/checkpoint-1116/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d91052454525486c42eebfab712e96c789ad540d93499d38e2f13b7dc0be701
+size 15920

49_128_e5_3e-5/checkpoint-1116/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:358b81b13f856d3ca834140bc0c300cfda87c97c0ddaeed002b114ff03887ae4
+size 15920

49_128_e5_3e-5/checkpoint-1116/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5175a20233d360c54a49364417aa0926915c0b70442b576f95117f7b39bddf2
+size 15920

49_128_e5_3e-5/checkpoint-1116/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f0ddf129009f953cd53267d328b842ebe104b06b2dea5a28b1e05bd5542498e
+size 15920

49_128_e5_3e-5/checkpoint-1116/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1524ade282739ce68d80412ac5bbe42ac86c977a4bcbb9d0fe6500395f6fb2bd
+size 15920

49_128_e5_3e-5/checkpoint-1116/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec4ede89b1a702308690a1d1cbfba89469ae3f53614e4011689edf146b6256bb
+size 15920

49_128_e5_3e-5/checkpoint-1116/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8ec96a43b749e20c25a87e0f8ed048c223539051a9b103d4387d3bea3489d12f
+size 1064

49_128_e5_3e-5/checkpoint-1116/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
+  }
+}

49_128_e5_3e-5/checkpoint-1116/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
+}

49_128_e5_3e-5/checkpoint-1116/trainer_state.json

@@ -0,0 +1,1595 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1116,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.013458950201884253,
+      "grad_norm": 1.00712251663208,
+      "learning_rate": 1.2903225806451614e-06,
+      "loss": 1.264,
+      "step": 5
+    },
+    {
+      "epoch": 0.026917900403768506,
+      "grad_norm": 0.9430832266807556,
+      "learning_rate": 2.9032258064516128e-06,
+      "loss": 1.2378,
+      "step": 10
+    },
+    {
+      "epoch": 0.040376850605652756,
+      "grad_norm": 0.6684543490409851,
+      "learning_rate": 4.516129032258065e-06,
+      "loss": 1.239,
+      "step": 15
+    },
+    {
+      "epoch": 0.05383580080753701,
+      "grad_norm": 0.7191868424415588,
+      "learning_rate": 6.129032258064517e-06,
+      "loss": 1.2218,
+      "step": 20
+    },
+    {
+      "epoch": 0.06729475100942127,
+      "grad_norm": 0.6867708563804626,
+      "learning_rate": 7.741935483870968e-06,
+      "loss": 1.2301,
+      "step": 25
+    },
+    {
+      "epoch": 0.08075370121130551,
+      "grad_norm": 0.6379647254943848,
+      "learning_rate": 9.35483870967742e-06,
+      "loss": 1.2194,
+      "step": 30
+    },
+    {
+      "epoch": 0.09421265141318977,
+      "grad_norm": 0.5798360109329224,
+      "learning_rate": 1.0967741935483872e-05,
+      "loss": 1.2002,
+      "step": 35
+    },
+    {
+      "epoch": 0.10767160161507403,
+      "grad_norm": 0.5719746351242065,
+      "learning_rate": 1.2580645161290324e-05,
+      "loss": 1.1982,
+      "step": 40
+    },
+    {
+      "epoch": 0.12113055181695828,
+      "grad_norm": 0.5450358390808105,
+      "learning_rate": 1.4193548387096776e-05,
+      "loss": 1.1986,
+      "step": 45
+    },
+    {
+      "epoch": 0.13458950201884254,
+      "grad_norm": 0.5413178205490112,
+      "learning_rate": 1.5806451612903226e-05,
+      "loss": 1.1955,
+      "step": 50
+    },
+    {
+      "epoch": 0.1480484522207268,
+      "grad_norm": 0.5677768588066101,
+      "learning_rate": 1.741935483870968e-05,
+      "loss": 1.1496,
+      "step": 55
+    },
+    {
+      "epoch": 0.16150740242261102,
+      "grad_norm": 0.5705103278160095,
+      "learning_rate": 1.903225806451613e-05,
+      "loss": 1.1377,
+      "step": 60
+    },
+    {
+      "epoch": 0.17496635262449528,
+      "grad_norm": 0.6191664934158325,
+      "learning_rate": 2.0645161290322582e-05,
+      "loss": 1.1242,
+      "step": 65
+    },
+    {
+      "epoch": 0.18842530282637954,
+      "grad_norm": 0.41417160630226135,
+      "learning_rate": 2.2258064516129034e-05,
+      "loss": 1.0923,
+      "step": 70
+    },
+    {
+      "epoch": 0.2018842530282638,
+      "grad_norm": 0.5877333283424377,
+      "learning_rate": 2.3870967741935483e-05,
+      "loss": 1.1575,
+      "step": 75
+    },
+    {
+      "epoch": 0.21534320323014805,
+      "grad_norm": 0.5006933808326721,
+      "learning_rate": 2.548387096774194e-05,
+      "loss": 1.0873,
+      "step": 80
+    },
+    {
+      "epoch": 0.2288021534320323,
+      "grad_norm": 0.5454208254814148,
+      "learning_rate": 2.7096774193548387e-05,
+      "loss": 1.073,
+      "step": 85
+    },
+    {
+      "epoch": 0.24226110363391656,
+      "grad_norm": 0.4708370268344879,
+      "learning_rate": 2.870967741935484e-05,
+      "loss": 1.1246,
+      "step": 90
+    },
+    {
+      "epoch": 0.2557200538358008,
+      "grad_norm": 0.7446239590644836,
+      "learning_rate": 2.9999976292388853e-05,
+      "loss": 1.0893,
+      "step": 95
+    },
+    {
+      "epoch": 0.2691790040376851,
+      "grad_norm": 0.5205251574516296,
+      "learning_rate": 2.999914653386729e-05,
+      "loss": 1.0233,
+      "step": 100
+    },
+    {
+      "epoch": 0.28263795423956933,
+      "grad_norm": 0.5542618632316589,
+      "learning_rate": 2.9997131469727287e-05,
+      "loss": 1.0534,
+      "step": 105
+    },
+    {
+      "epoch": 0.2960969044414536,
+      "grad_norm": 0.5413870811462402,
+      "learning_rate": 2.9993931259209015e-05,
+      "loss": 1.0217,
+      "step": 110
+    },
+    {
+      "epoch": 0.30955585464333785,
+      "grad_norm": 0.5619310736656189,
+      "learning_rate": 2.998954615520871e-05,
+      "loss": 1.0357,
+      "step": 115
+    },
+    {
+      "epoch": 0.32301480484522205,
+      "grad_norm": 0.5382521152496338,
+      "learning_rate": 2.998397650425863e-05,
+      "loss": 1.0228,
+      "step": 120
+    },
+    {
+      "epoch": 0.3364737550471063,
+      "grad_norm": 0.6172555088996887,
+      "learning_rate": 2.997722274649974e-05,
+      "loss": 0.9711,
+      "step": 125
+    },
+    {
+      "epoch": 0.34993270524899056,
+      "grad_norm": 0.6441250443458557,
+      "learning_rate": 2.9969285415646874e-05,
+      "loss": 0.9977,
+      "step": 130
+    },
+    {
+      "epoch": 0.3633916554508748,
+      "grad_norm": 0.5927826166152954,
+      "learning_rate": 2.9960165138946572e-05,
+      "loss": 0.9778,
+      "step": 135
+    },
+    {
+      "epoch": 0.3768506056527591,
+      "grad_norm": 0.6550005078315735,
+      "learning_rate": 2.9949862637127523e-05,
+      "loss": 0.9335,
+      "step": 140
+    },
+    {
+      "epoch": 0.39030955585464333,
+      "grad_norm": 0.5886467099189758,
+      "learning_rate": 2.9938378724343604e-05,
+      "loss": 0.9681,
+      "step": 145
+    },
+    {
+      "epoch": 0.4037685060565276,
+      "grad_norm": 0.7246305346488953,
+      "learning_rate": 2.992571430810954e-05,
+      "loss": 0.9701,
+      "step": 150
+    },
+    {
+      "epoch": 0.41722745625841184,
+      "grad_norm": 0.7467737197875977,
+      "learning_rate": 2.9911870389229192e-05,
+      "loss": 0.9141,
+      "step": 155
+    },
+    {
+      "epoch": 0.4306864064602961,
+      "grad_norm": 0.7680078744888306,
+      "learning_rate": 2.9896848061716456e-05,
+      "loss": 0.8685,
+      "step": 160
+    },
+    {
+      "epoch": 0.44414535666218036,
+      "grad_norm": 0.7801365256309509,
+      "learning_rate": 2.9880648512708834e-05,
+      "loss": 0.9138,
+      "step": 165
+    },
+    {
+      "epoch": 0.4576043068640646,
+      "grad_norm": 0.6462291479110718,
+      "learning_rate": 2.9863273022373585e-05,
+      "loss": 0.9223,
+      "step": 170
+    },
+    {
+      "epoch": 0.47106325706594887,
+      "grad_norm": 0.6398594975471497,
+      "learning_rate": 2.9844722963806592e-05,
+      "loss": 0.9021,
+      "step": 175
+    },
+    {
+      "epoch": 0.4845222072678331,
+      "grad_norm": 0.683367133140564,
+      "learning_rate": 2.9824999802923846e-05,
+      "loss": 0.8599,
+      "step": 180
+    },
+    {
+      "epoch": 0.4979811574697174,
+      "grad_norm": 0.7672361731529236,
+      "learning_rate": 2.9804105098345575e-05,
+      "loss": 0.9282,
+      "step": 185
+    },
+    {
+      "epoch": 0.5114401076716016,
+      "grad_norm": 0.6266528367996216,
+      "learning_rate": 2.9782040501273126e-05,
+      "loss": 0.8664,
+      "step": 190
+    },
+    {
+      "epoch": 0.5248990578734859,
+      "grad_norm": 0.7748366594314575,
+      "learning_rate": 2.975880775535843e-05,
+      "loss": 0.8588,
+      "step": 195
+    },
+    {
+      "epoch": 0.5383580080753702,
+      "grad_norm": 0.892792820930481,
+      "learning_rate": 2.9734408696566224e-05,
+      "loss": 0.8517,
+      "step": 200
+    },
+    {
+      "epoch": 0.5518169582772544,
+      "grad_norm": 0.8755563497543335,
+      "learning_rate": 2.9708845253028988e-05,
+      "loss": 0.9136,
+      "step": 205
+    },
+    {
+      "epoch": 0.5652759084791387,
+      "grad_norm": 0.7859370708465576,
+      "learning_rate": 2.9682119444894545e-05,
+      "loss": 0.7955,
+      "step": 210
+    },
+    {
+      "epoch": 0.5787348586810229,
+      "grad_norm": 0.7947808504104614,
+      "learning_rate": 2.9654233384166435e-05,
+      "loss": 0.8695,
+      "step": 215
+    },
+    {
+      "epoch": 0.5921938088829072,
+      "grad_norm": 0.7226110696792603,
+      "learning_rate": 2.9625189274537004e-05,
+      "loss": 0.7672,
+      "step": 220
+    },
+    {
+      "epoch": 0.6056527590847914,
+      "grad_norm": 1.0677162408828735,
+      "learning_rate": 2.9594989411213266e-05,
+      "loss": 0.816,
+      "step": 225
+    },
+    {
+      "epoch": 0.6191117092866757,
+      "grad_norm": 0.8201079964637756,
+      "learning_rate": 2.9563636180735524e-05,
+      "loss": 0.7877,
+      "step": 230
+    },
+    {
+      "epoch": 0.6325706594885598,
+      "grad_norm": 0.961162805557251,
+      "learning_rate": 2.9531132060788763e-05,
+      "loss": 0.7566,
+      "step": 235
+    },
+    {
+      "epoch": 0.6460296096904441,
+      "grad_norm": 0.9038267731666565,
+      "learning_rate": 2.9497479620006873e-05,
+      "loss": 0.8061,
+      "step": 240
+    },
+    {
+      "epoch": 0.6594885598923284,
+      "grad_norm": 0.9068340063095093,
+      "learning_rate": 2.9462681517769638e-05,
+      "loss": 0.787,
+      "step": 245
+    },
+    {
+      "epoch": 0.6729475100942126,
+      "grad_norm": 0.8361156582832336,
+      "learning_rate": 2.94267405039926e-05,
+      "loss": 0.7835,
+      "step": 250
+    },
+    {
+      "epoch": 0.6864064602960969,
+      "grad_norm": 0.8174508213996887,
+      "learning_rate": 2.938965941890972e-05,
+      "loss": 0.7641,
+      "step": 255
+    },
+    {
+      "epoch": 0.6998654104979811,
+      "grad_norm": 0.8101845383644104,
+      "learning_rate": 2.9351441192848972e-05,
+      "loss": 0.76,
+      "step": 260
+    },
+    {
+      "epoch": 0.7133243606998654,
+      "grad_norm": 0.8047847747802734,
+      "learning_rate": 2.9312088846000733e-05,
+      "loss": 0.7404,
+      "step": 265
+    },
+    {
+      "epoch": 0.7267833109017496,
+      "grad_norm": 0.8787263035774231,
+      "learning_rate": 2.927160548817913e-05,
+      "loss": 0.7335,
+      "step": 270
+    },
+    {
+      "epoch": 0.7402422611036339,
+      "grad_norm": 0.9566526412963867,
+      "learning_rate": 2.9229994318576295e-05,
+      "loss": 0.7546,
+      "step": 275
+    },
+    {
+      "epoch": 0.7537012113055181,
+      "grad_norm": 0.9340652227401733,
+      "learning_rate": 2.9187258625509518e-05,
+      "loss": 0.6814,
+      "step": 280
+    },
+    {
+      "epoch": 0.7671601615074024,
+      "grad_norm": 1.0249980688095093,
+      "learning_rate": 2.9143401786161445e-05,
+      "loss": 0.7116,
+      "step": 285
+    },
+    {
+      "epoch": 0.7806191117092867,
+      "grad_norm": 0.8947973847389221,
+      "learning_rate": 2.9098427266313138e-05,
+      "loss": 0.6845,
+      "step": 290
+    },
+    {
+      "epoch": 0.7940780619111709,
+      "grad_norm": 0.9918004274368286,
+      "learning_rate": 2.9052338620070215e-05,
+      "loss": 0.7287,
+      "step": 295
+    },
+    {
+      "epoch": 0.8075370121130552,
+      "grad_norm": 0.9002223610877991,
+      "learning_rate": 2.9005139489582002e-05,
+      "loss": 0.7129,
+      "step": 300
+    },
+    {
+      "epoch": 0.8209959623149394,
+      "grad_norm": 0.8825509548187256,
+      "learning_rate": 2.8956833604753688e-05,
+      "loss": 0.6633,
+      "step": 305
+    },
+    {
+      "epoch": 0.8344549125168237,
+      "grad_norm": 0.8505289554595947,
+      "learning_rate": 2.8907424782951587e-05,
+      "loss": 0.6826,
+      "step": 310
+    },
+    {
+      "epoch": 0.847913862718708,
+      "grad_norm": 1.0375418663024902,
+      "learning_rate": 2.8856916928701467e-05,
+      "loss": 0.7003,
+      "step": 315
+    },
+    {
+      "epoch": 0.8613728129205922,
+      "grad_norm": 1.0597079992294312,
+      "learning_rate": 2.8805314033379985e-05,
+      "loss": 0.7151,
+      "step": 320
+    },
+    {
+      "epoch": 0.8748317631224765,
+      "grad_norm": 1.1800124645233154,
+      "learning_rate": 2.8752620174899284e-05,
+      "loss": 0.6594,
+      "step": 325
+    },
+    {
+      "epoch": 0.8882907133243607,
+      "grad_norm": 0.9276929497718811,
+      "learning_rate": 2.8698839517384726e-05,
+      "loss": 0.651,
+      "step": 330
+    },
+    {
+      "epoch": 0.901749663526245,
+      "grad_norm": 1.094962239265442,
+      "learning_rate": 2.8643976310845828e-05,
+      "loss": 0.6412,
+      "step": 335
+    },
+    {
+      "epoch": 0.9152086137281292,
+      "grad_norm": 0.9263754487037659,
+      "learning_rate": 2.8588034890840408e-05,
+      "loss": 0.6269,
+      "step": 340
+    },
+    {
+      "epoch": 0.9286675639300135,
+      "grad_norm": 0.9816908240318298,
+      "learning_rate": 2.8531019678131952e-05,
+      "loss": 0.6577,
+      "step": 345
+    },
+    {
+      "epoch": 0.9421265141318977,
+      "grad_norm": 1.0188283920288086,
+      "learning_rate": 2.847293517834029e-05,
+      "loss": 0.66,
+      "step": 350
+    },
+    {
+      "epoch": 0.955585464333782,
+      "grad_norm": 1.0503355264663696,
+      "learning_rate": 2.8413785981585518e-05,
+      "loss": 0.621,
+      "step": 355
+    },
+    {
+      "epoch": 0.9690444145356663,
+      "grad_norm": 0.9092134237289429,
+      "learning_rate": 2.8353576762125274e-05,
+      "loss": 0.6343,
+      "step": 360
+    },
+    {
+      "epoch": 0.9825033647375505,
+      "grad_norm": 0.9565970301628113,
+      "learning_rate": 2.8292312277985354e-05,
+      "loss": 0.6538,
+      "step": 365
+    },
+    {
+      "epoch": 0.9959623149394348,
+      "grad_norm": 1.014952301979065,
+      "learning_rate": 2.82299973705837e-05,
+      "loss": 0.6081,
+      "step": 370
+    },
+    {
+      "epoch": 1.0080753701211305,
+      "grad_norm": 1.1951587200164795,
+      "learning_rate": 2.816663696434783e-05,
+      "loss": 0.5937,
+      "step": 375
+    },
+    {
+      "epoch": 1.0215343203230147,
+      "grad_norm": 0.9546276926994324,
+      "learning_rate": 2.810223606632567e-05,
+      "loss": 0.5175,
+      "step": 380
+    },
+    {
+      "epoch": 1.034993270524899,
+      "grad_norm": 0.9498586654663086,
+      "learning_rate": 2.803679976578987e-05,
+      "loss": 0.534,
+      "step": 385
+    },
+    {
+      "epoch": 1.0484522207267832,
+      "grad_norm": 1.189976453781128,
+      "learning_rate": 2.797033323383563e-05,
+      "loss": 0.5639,
+      "step": 390
+    },
+    {
+      "epoch": 1.0619111709286675,
+      "grad_norm": 1.038994550704956,
+      "learning_rate": 2.7902841722972062e-05,
+      "loss": 0.5405,
+      "step": 395
+    },
+    {
+      "epoch": 1.0753701211305517,
+      "grad_norm": 1.0657168626785278,
+      "learning_rate": 2.7834330566707086e-05,
+      "loss": 0.5407,
+      "step": 400
+    },
+    {
+      "epoch": 1.088829071332436,
+      "grad_norm": 1.0063115358352661,
+      "learning_rate": 2.7764805179126e-05,
+      "loss": 0.524,
+      "step": 405
+    },
+    {
+      "epoch": 1.1022880215343203,
+      "grad_norm": 0.9580488801002502,
+      "learning_rate": 2.7694271054463565e-05,
+      "loss": 0.5357,
+      "step": 410
+    },
+    {
+      "epoch": 1.1157469717362045,
+      "grad_norm": 1.0191570520401,
+      "learning_rate": 2.762273376666989e-05,
+      "loss": 0.4963,
+      "step": 415
+    },
+    {
+      "epoch": 1.1292059219380888,
+      "grad_norm": 1.0942199230194092,
+      "learning_rate": 2.755019896896989e-05,
+      "loss": 0.4825,
+      "step": 420
+    },
+    {
+      "epoch": 1.142664872139973,
+      "grad_norm": 0.9650993347167969,
+      "learning_rate": 2.7476672393416603e-05,
+      "loss": 0.4862,
+      "step": 425
+    },
+    {
+      "epoch": 1.1561238223418573,
+      "grad_norm": 1.0294479131698608,
+      "learning_rate": 2.7402159850438173e-05,
+      "loss": 0.4972,
+      "step": 430
+    },
+    {
+      "epoch": 1.1695827725437415,
+      "grad_norm": 1.1111294031143188,
+      "learning_rate": 2.7326667228378677e-05,
+      "loss": 0.4757,
+      "step": 435
+    },
+    {
+      "epoch": 1.1830417227456258,
+      "grad_norm": 1.099021077156067,
+      "learning_rate": 2.7250200493032836e-05,
+      "loss": 0.4916,
+      "step": 440
+    },
+    {
+      "epoch": 1.19650067294751,
+      "grad_norm": 1.0875200033187866,
+      "learning_rate": 2.7172765687174536e-05,
+      "loss": 0.5004,
+      "step": 445
+    },
+    {
+      "epoch": 1.2099596231493943,
+      "grad_norm": 1.0877183675765991,
+      "learning_rate": 2.7094368930079312e-05,
+      "loss": 0.4856,
+      "step": 450
+    },
+    {
+      "epoch": 1.2234185733512786,
+      "grad_norm": 1.0833150148391724,
+      "learning_rate": 2.701501641704077e-05,
+      "loss": 0.5081,
+      "step": 455
+    },
+    {
+      "epoch": 1.2368775235531628,
+      "grad_norm": 1.0706162452697754,
+      "learning_rate": 2.693471441888101e-05,
+      "loss": 0.4808,
+      "step": 460
+    },
+    {
+      "epoch": 1.250336473755047,
+      "grad_norm": 1.0003634691238403,
+      "learning_rate": 2.685346928145508e-05,
+      "loss": 0.5242,
+      "step": 465
+    },
+    {
+      "epoch": 1.2637954239569313,
+      "grad_norm": 1.081905722618103,
+      "learning_rate": 2.6771287425149465e-05,
+      "loss": 0.4724,
+      "step": 470
+    },
+    {
+      "epoch": 1.2772543741588156,
+      "grad_norm": 1.0233476161956787,
+      "learning_rate": 2.6688175344374748e-05,
+      "loss": 0.4461,
+      "step": 475
+    },
+    {
+      "epoch": 1.2907133243606999,
+      "grad_norm": 0.9034926295280457,
+      "learning_rate": 2.660413960705239e-05,
+      "loss": 0.5191,
+      "step": 480
+    },
+    {
+      "epoch": 1.304172274562584,
+      "grad_norm": 0.9046325087547302,
+      "learning_rate": 2.651918685409568e-05,
+      "loss": 0.4811,
+      "step": 485
+    },
+    {
+      "epoch": 1.3176312247644684,
+      "grad_norm": 0.9982317686080933,
+      "learning_rate": 2.6433323798884955e-05,
+      "loss": 0.4671,
+      "step": 490
+    },
+    {
+      "epoch": 1.3310901749663526,
+      "grad_norm": 1.2173210382461548,
+      "learning_rate": 2.6346557226737064e-05,
+      "loss": 0.4575,
+      "step": 495
+    },
+    {
+      "epoch": 1.3445491251682369,
+      "grad_norm": 1.0846431255340576,
+      "learning_rate": 2.6258893994369176e-05,
+      "loss": 0.436,
+      "step": 500
+    },
+    {
+      "epoch": 1.3580080753701211,
+      "grad_norm": 1.0849974155426025,
+      "learning_rate": 2.6170341029356906e-05,
+      "loss": 0.4323,
+      "step": 505
+    },
+    {
+      "epoch": 1.3714670255720054,
+      "grad_norm": 1.0231881141662598,
+      "learning_rate": 2.6080905329586883e-05,
+      "loss": 0.4259,
+      "step": 510
+    },
+    {
+      "epoch": 1.3849259757738897,
+      "grad_norm": 1.1137603521347046,
+      "learning_rate": 2.599059396270374e-05,
+      "loss": 0.4539,
+      "step": 515
+    },
+    {
+      "epoch": 1.398384925975774,
+      "grad_norm": 1.066973090171814,
+      "learning_rate": 2.5899414065551578e-05,
+      "loss": 0.4475,
+      "step": 520
+    },
+    {
+      "epoch": 1.4118438761776582,
+      "grad_norm": 1.1666553020477295,
+      "learning_rate": 2.5807372843610002e-05,
+      "loss": 0.436,
+      "step": 525
+    },
+    {
+      "epoch": 1.4253028263795424,
+      "grad_norm": 1.1762378215789795,
+      "learning_rate": 2.5714477570424692e-05,
+      "loss": 0.4234,
+      "step": 530
+    },
+    {
+      "epoch": 1.4387617765814267,
+      "grad_norm": 1.053694486618042,
+      "learning_rate": 2.5620735587032622e-05,
+      "loss": 0.4242,
+      "step": 535
+    },
+    {
+      "epoch": 1.452220726783311,
+      "grad_norm": 1.1577174663543701,
+      "learning_rate": 2.5526154301381926e-05,
+      "loss": 0.408,
+      "step": 540
+    },
+    {
+      "epoch": 1.4656796769851952,
+      "grad_norm": 1.1527884006500244,
+      "learning_rate": 2.5430741187746485e-05,
+      "loss": 0.4154,
+      "step": 545
+    },
+    {
+      "epoch": 1.4791386271870794,
+      "grad_norm": 1.005540132522583,
+      "learning_rate": 2.53345037861353e-05,
+      "loss": 0.4474,
+      "step": 550
+    },
+    {
+      "epoch": 1.4925975773889637,
+      "grad_norm": 1.0806723833084106,
+      "learning_rate": 2.52374497016966e-05,
+      "loss": 0.4015,
+      "step": 555
+    },
+    {
+      "epoch": 1.506056527590848,
+      "grad_norm": 1.0257374048233032,
+      "learning_rate": 2.513958660411688e-05,
+      "loss": 0.4348,
+      "step": 560
+    },
+    {
+      "epoch": 1.5195154777927322,
+      "grad_norm": 0.9938416481018066,
+      "learning_rate": 2.5040922227014787e-05,
+      "loss": 0.4049,
+      "step": 565
+    },
+    {
+      "epoch": 1.5329744279946165,
+      "grad_norm": 1.0034414529800415,
+      "learning_rate": 2.4941464367329993e-05,
+      "loss": 0.4436,
+      "step": 570
+    },
+    {
+      "epoch": 1.5464333781965007,
+      "grad_norm": 1.0372949838638306,
+      "learning_rate": 2.4841220884707014e-05,
+      "loss": 0.3635,
+      "step": 575
+    },
+    {
+      "epoch": 1.559892328398385,
+      "grad_norm": 1.1650141477584839,
+      "learning_rate": 2.4740199700874124e-05,
+      "loss": 0.3933,
+      "step": 580
+    },
+    {
+      "epoch": 1.5733512786002692,
+      "grad_norm": 1.1232426166534424,
+      "learning_rate": 2.4638408799017333e-05,
+      "loss": 0.3762,
+      "step": 585
+    },
+    {
+      "epoch": 1.5868102288021535,
+      "grad_norm": 1.0809695720672607,
+      "learning_rate": 2.4535856223149525e-05,
+      "loss": 0.364,
+      "step": 590
+    },
+    {
+      "epoch": 1.6002691790040378,
+      "grad_norm": 1.0743389129638672,
+      "learning_rate": 2.4432550077474753e-05,
+      "loss": 0.4135,
+      "step": 595
+    },
+    {
+      "epoch": 1.613728129205922,
+      "grad_norm": 1.1608853340148926,
+      "learning_rate": 2.4328498525747862e-05,
+      "loss": 0.373,
+      "step": 600
+    },
+    {
+      "epoch": 1.6271870794078063,
+      "grad_norm": 1.0588221549987793,
+      "learning_rate": 2.4223709790629275e-05,
+      "loss": 0.3898,
+      "step": 605
+    },
+    {
+      "epoch": 1.6406460296096905,
+      "grad_norm": 1.08194100856781,
+      "learning_rate": 2.4118192153035273e-05,
+      "loss": 0.3759,
+      "step": 610
+    },
+    {
+      "epoch": 1.6541049798115748,
+      "grad_norm": 0.9936762452125549,
+      "learning_rate": 2.4011953951483532e-05,
+      "loss": 0.4186,
+      "step": 615
+    },
+    {
+      "epoch": 1.667563930013459,
+      "grad_norm": 1.1843287944793701,
+      "learning_rate": 2.3905003581434227e-05,
+      "loss": 0.3562,
+      "step": 620
+    },
+    {
+      "epoch": 1.6810228802153433,
+      "grad_norm": 1.525776743888855,
+      "learning_rate": 2.379734949462653e-05,
+      "loss": 0.3998,
+      "step": 625
+    },
+    {
+      "epoch": 1.6944818304172276,
+      "grad_norm": 1.056809425354004,
+      "learning_rate": 2.3689000198410762e-05,
+      "loss": 0.4076,
+      "step": 630
+    },
+    {
+      "epoch": 1.7079407806191118,
+      "grad_norm": 1.3260629177093506,
+      "learning_rate": 2.3579964255076064e-05,
+      "loss": 0.3742,
+      "step": 635
+    },
+    {
+      "epoch": 1.721399730820996,
+      "grad_norm": 1.0809186697006226,
+      "learning_rate": 2.3470250281173787e-05,
+      "loss": 0.4176,
+      "step": 640
+    },
+    {
+      "epoch": 1.7348586810228803,
+      "grad_norm": 0.9884938597679138,
+      "learning_rate": 2.3359866946836546e-05,
+      "loss": 0.3475,
+      "step": 645
+    },
+    {
+      "epoch": 1.7483176312247646,
+      "grad_norm": 1.0996809005737305,
+      "learning_rate": 2.3248822975093094e-05,
+      "loss": 0.349,
+      "step": 650
+    },
+    {
+      "epoch": 1.7617765814266488,
+      "grad_norm": 1.3934807777404785,
+      "learning_rate": 2.3137127141178962e-05,
+      "loss": 0.317,
+      "step": 655
+    },
+    {
+      "epoch": 1.775235531628533,
+      "grad_norm": 1.1072278022766113,
+      "learning_rate": 2.3024788271842997e-05,
+      "loss": 0.3442,
+      "step": 660
+    },
+    {
+      "epoch": 1.7886944818304173,
+      "grad_norm": 1.1622191667556763,
+      "learning_rate": 2.2911815244649855e-05,
+      "loss": 0.3707,
+      "step": 665
+    },
+    {
+      "epoch": 1.8021534320323016,
+      "grad_norm": 1.1260006427764893,
+      "learning_rate": 2.2798216987278417e-05,
+      "loss": 0.3506,
+      "step": 670
+    },
+    {
+      "epoch": 1.8156123822341859,
+      "grad_norm": 0.9603449702262878,
+      "learning_rate": 2.2684002476816317e-05,
+      "loss": 0.3696,
+      "step": 675
+    },
+    {
+      "epoch": 1.8290713324360701,
+      "grad_norm": 1.0738201141357422,
+      "learning_rate": 2.2569180739050486e-05,
+      "loss": 0.3615,
+      "step": 680
+    },
+    {
+      "epoch": 1.8425302826379544,
+      "grad_norm": 1.082394003868103,
+      "learning_rate": 2.2453760847753937e-05,
+      "loss": 0.3259,
+      "step": 685
+    },
+    {
+      "epoch": 1.8559892328398386,
+      "grad_norm": 1.0013381242752075,
+      "learning_rate": 2.2337751923968673e-05,
+      "loss": 0.3602,
+      "step": 690
+    },
+    {
+      "epoch": 1.8694481830417229,
+      "grad_norm": 0.9863987565040588,
+      "learning_rate": 2.2221163135284926e-05,
+      "loss": 0.3096,
+      "step": 695
+    },
+    {
+      "epoch": 1.8829071332436071,
+      "grad_norm": 1.0083246231079102,
+      "learning_rate": 2.2104003695116662e-05,
+      "loss": 0.3408,
+      "step": 700
+    },
+    {
+      "epoch": 1.8963660834454914,
+      "grad_norm": 1.1064270734786987,
+      "learning_rate": 2.1986282861973525e-05,
+      "loss": 0.3569,
+      "step": 705
+    },
+    {
+      "epoch": 1.9098250336473757,
+      "grad_norm": 1.062014102935791,
+      "learning_rate": 2.1868009938729145e-05,
+      "loss": 0.3376,
+      "step": 710
+    },
+    {
+      "epoch": 1.92328398384926,
+      "grad_norm": 1.1426458358764648,
+      "learning_rate": 2.174919427188602e-05,
+      "loss": 0.3375,
+      "step": 715
+    },
+    {
+      "epoch": 1.9367429340511442,
+      "grad_norm": 1.0394279956817627,
+      "learning_rate": 2.1629845250836882e-05,
+      "loss": 0.2991,
+      "step": 720
+    },
+    {
+      "epoch": 1.9502018842530284,
+      "grad_norm": 1.1961948871612549,
+      "learning_rate": 2.150997230712271e-05,
+      "loss": 0.3048,
+      "step": 725
+    },
+    {
+      "epoch": 1.9636608344549125,
+      "grad_norm": 1.011620283126831,
+      "learning_rate": 2.138958491368741e-05,
+      "loss": 0.3241,
+      "step": 730
+    },
+    {
+      "epoch": 1.9771197846567967,
+      "grad_norm": 1.053614616394043,
+      "learning_rate": 2.1268692584129197e-05,
+      "loss": 0.3141,
+      "step": 735
+    },
+    {
+      "epoch": 1.990578734858681,
+      "grad_norm": 1.0024995803833008,
+      "learning_rate": 2.1147304871948812e-05,
+      "loss": 0.3211,
+      "step": 740
+    },
+    {
+      "epoch": 2.0026917900403767,
+      "grad_norm": 1.0640769004821777,
+      "learning_rate": 2.1025431369794546e-05,
+      "loss": 0.2784,
+      "step": 745
+    },
+    {
+      "epoch": 2.016150740242261,
+      "grad_norm": 1.1959905624389648,
+      "learning_rate": 2.090308170870416e-05,
+      "loss": 0.2488,
+      "step": 750
+    },
+    {
+      "epoch": 2.029609690444145,
+      "grad_norm": 1.0885748863220215,
+      "learning_rate": 2.0780265557343833e-05,
+      "loss": 0.2466,
+      "step": 755
+    },
+    {
+      "epoch": 2.0430686406460294,
+      "grad_norm": 1.041433334350586,
+      "learning_rate": 2.065699262124407e-05,
+      "loss": 0.2308,
+      "step": 760
+    },
+    {
+      "epoch": 2.0565275908479137,
+      "grad_norm": 1.1424509286880493,
+      "learning_rate": 2.0533272642032734e-05,
+      "loss": 0.2545,
+      "step": 765
+    },
+    {
+      "epoch": 2.069986541049798,
+      "grad_norm": 1.0592752695083618,
+      "learning_rate": 2.0409115396665202e-05,
+      "loss": 0.2373,
+      "step": 770
+    },
+    {
+      "epoch": 2.083445491251682,
+      "grad_norm": 1.3094651699066162,
+      "learning_rate": 2.0284530696651772e-05,
+      "loss": 0.2629,
+      "step": 775
+    },
+    {
+      "epoch": 2.0969044414535665,
+      "grad_norm": 1.171372890472412,
+      "learning_rate": 2.0159528387282273e-05,
+      "loss": 0.2368,
+      "step": 780
+    },
+    {
+      "epoch": 2.1103633916554507,
+      "grad_norm": 1.0821523666381836,
+      "learning_rate": 2.0034118346848057e-05,
+      "loss": 0.2267,
+      "step": 785
+    },
+    {
+      "epoch": 2.123822341857335,
+      "grad_norm": 1.0307716131210327,
+      "learning_rate": 1.9908310485861393e-05,
+      "loss": 0.2223,
+      "step": 790
+    },
+    {
+      "epoch": 2.1372812920592192,
+      "grad_norm": 1.2713267803192139,
+      "learning_rate": 1.9782114746272252e-05,
+      "loss": 0.2493,
+      "step": 795
+    },
+    {
+      "epoch": 2.1507402422611035,
+      "grad_norm": 1.1776148080825806,
+      "learning_rate": 1.965554110068265e-05,
+      "loss": 0.2308,
+      "step": 800
+    },
+    {
+      "epoch": 2.1641991924629878,
+      "grad_norm": 1.030351996421814,
+      "learning_rate": 1.95285995515586e-05,
+      "loss": 0.2718,
+      "step": 805
+    },
+    {
+      "epoch": 2.177658142664872,
+      "grad_norm": 0.9716264009475708,
+      "learning_rate": 1.9401300130439635e-05,
+      "loss": 0.2126,
+      "step": 810
+    },
+    {
+      "epoch": 2.1911170928667563,
+      "grad_norm": 1.125478982925415,
+      "learning_rate": 1.9273652897146063e-05,
+      "loss": 0.2335,
+      "step": 815
+    },
+    {
+      "epoch": 2.2045760430686405,
+      "grad_norm": 1.0152146816253662,
+      "learning_rate": 1.9145667938984034e-05,
+      "loss": 0.2171,
+      "step": 820
+    },
+    {
+      "epoch": 2.218034993270525,
+      "grad_norm": 1.0698730945587158,
+      "learning_rate": 1.9017355369948327e-05,
+      "loss": 0.2297,
+      "step": 825
+    },
+    {
+      "epoch": 2.231493943472409,
+      "grad_norm": 1.0559672117233276,
+      "learning_rate": 1.8888725329923147e-05,
+      "loss": 0.2439,
+      "step": 830
+    },
+    {
+      "epoch": 2.2449528936742933,
+      "grad_norm": 0.9136818051338196,
+      "learning_rate": 1.8759787983880807e-05,
+      "loss": 0.214,
+      "step": 835
+    },
+    {
+      "epoch": 2.2584118438761775,
+      "grad_norm": 1.2098388671875,
+      "learning_rate": 1.8630553521078432e-05,
+      "loss": 0.2305,
+      "step": 840
+    },
+    {
+      "epoch": 2.271870794078062,
+      "grad_norm": 1.0378409624099731,
+      "learning_rate": 1.850103215425276e-05,
+      "loss": 0.2223,
+      "step": 845
+    },
+    {
+      "epoch": 2.285329744279946,
+      "grad_norm": 1.0797078609466553,
+      "learning_rate": 1.8371234118813082e-05,
+      "loss": 0.221,
+      "step": 850
+    },
+    {
+      "epoch": 2.2987886944818303,
+      "grad_norm": 1.0039594173431396,
+      "learning_rate": 1.8241169672032392e-05,
+      "loss": 0.2353,
+      "step": 855
+    },
+    {
+      "epoch": 2.3122476446837146,
+      "grad_norm": 1.3303050994873047,
+      "learning_rate": 1.81108490922368e-05,
+      "loss": 0.2333,
+      "step": 860
+    },
+    {
+      "epoch": 2.325706594885599,
+      "grad_norm": 1.0132439136505127,
+      "learning_rate": 1.7980282677993308e-05,
+      "loss": 0.2138,
+      "step": 865
+    },
+    {
+      "epoch": 2.339165545087483,
+      "grad_norm": 1.0430055856704712,
+      "learning_rate": 1.7849480747295944e-05,
+      "loss": 0.2258,
+      "step": 870
+    },
+    {
+      "epoch": 2.3526244952893673,
+      "grad_norm": 1.0881597995758057,
+      "learning_rate": 1.7718453636750392e-05,
+      "loss": 0.2061,
+      "step": 875
+    },
+    {
+      "epoch": 2.3660834454912516,
+      "grad_norm": 1.0634713172912598,
+      "learning_rate": 1.758721170075715e-05,
+      "loss": 0.2306,
+      "step": 880
+    },
+    {
+      "epoch": 2.379542395693136,
+      "grad_norm": 0.9828234910964966,
+      "learning_rate": 1.7455765310693258e-05,
+      "loss": 0.228,
+      "step": 885
+    },
+    {
+      "epoch": 2.39300134589502,
+      "grad_norm": 1.1163634061813354,
+      "learning_rate": 1.7324124854092738e-05,
+      "loss": 0.2335,
+      "step": 890
+    },
+    {
+      "epoch": 2.4064602960969044,
+      "grad_norm": 1.0143085718154907,
+      "learning_rate": 1.7192300733825664e-05,
+      "loss": 0.2164,
+      "step": 895
+    },
+    {
+      "epoch": 2.4199192462987886,
+      "grad_norm": 1.1971455812454224,
+      "learning_rate": 1.7060303367276123e-05,
+      "loss": 0.2146,
+      "step": 900
+    },
+    {
+      "epoch": 2.433378196500673,
+      "grad_norm": 1.0438719987869263,
+      "learning_rate": 1.692814318551897e-05,
+      "loss": 0.212,
+      "step": 905
+    },
+    {
+      "epoch": 2.446837146702557,
+      "grad_norm": 0.9495139718055725,
+      "learning_rate": 1.6795830632495512e-05,
+      "loss": 0.2124,
+      "step": 910
+    },
+    {
+      "epoch": 2.4602960969044414,
+      "grad_norm": 1.0062249898910522,
+      "learning_rate": 1.6663376164188153e-05,
+      "loss": 0.2048,
+      "step": 915
+    },
+    {
+      "epoch": 2.4737550471063257,
+      "grad_norm": 1.0685555934906006,
+      "learning_rate": 1.6530790247794163e-05,
+      "loss": 0.1982,
+      "step": 920
+    },
+    {
+      "epoch": 2.48721399730821,
+      "grad_norm": 1.1590561866760254,
+      "learning_rate": 1.6398083360898474e-05,
+      "loss": 0.2194,
+      "step": 925
+    },
+    {
+      "epoch": 2.500672947510094,
+      "grad_norm": 0.9140127301216125,
+      "learning_rate": 1.626526599064567e-05,
+      "loss": 0.185,
+      "step": 930
+    },
+    {
+      "epoch": 2.5141318977119784,
+      "grad_norm": 1.0982829332351685,
+      "learning_rate": 1.613234863291131e-05,
+      "loss": 0.2175,
+      "step": 935
+    },
+    {
+      "epoch": 2.5275908479138627,
+      "grad_norm": 0.9458792209625244,
+      "learning_rate": 1.599934179147242e-05,
+      "loss": 0.1872,
+      "step": 940
+    },
+    {
+      "epoch": 2.541049798115747,
+      "grad_norm": 1.1163727045059204,
+      "learning_rate": 1.5866255977177477e-05,
+      "loss": 0.1964,
+      "step": 945
+    },
+    {
+      "epoch": 2.554508748317631,
+      "grad_norm": 1.2947419881820679,
+      "learning_rate": 1.5733101707115765e-05,
+      "loss": 0.1967,
+      "step": 950
+    },
+    {
+      "epoch": 2.5679676985195155,
+      "grad_norm": 1.181272268295288,
+      "learning_rate": 1.5599889503786296e-05,
+      "loss": 0.1856,
+      "step": 955
+    },
+    {
+      "epoch": 2.5814266487213997,
+      "grad_norm": 1.0785599946975708,
+      "learning_rate": 1.5466629894266233e-05,
+      "loss": 0.1932,
+      "step": 960
+    },
+    {
+      "epoch": 2.594885598923284,
+      "grad_norm": 2.4438836574554443,
+      "learning_rate": 1.5333333409379017e-05,
+      "loss": 0.1884,
+      "step": 965
+    },
+    {
+      "epoch": 2.608344549125168,
+      "grad_norm": 0.9497862458229065,
+      "learning_rate": 1.5200010582862162e-05,
+      "loss": 0.1836,
+      "step": 970
+    },
+    {
+      "epoch": 2.6218034993270525,
+      "grad_norm": 1.2000949382781982,
+      "learning_rate": 1.5066671950534817e-05,
+      "loss": 0.2057,
+      "step": 975
+    },
+    {
+      "epoch": 2.6352624495289367,
+      "grad_norm": 1.074912190437317,
+      "learning_rate": 1.493332804946519e-05,
+      "loss": 0.1973,
+      "step": 980
+    },
+    {
+      "epoch": 2.648721399730821,
+      "grad_norm": 1.1403034925460815,
+      "learning_rate": 1.4799989417137838e-05,
+      "loss": 0.1692,
+      "step": 985
+    },
+    {
+      "epoch": 2.6621803499327052,
+      "grad_norm": 1.1136847734451294,
+      "learning_rate": 1.4666666590620982e-05,
+      "loss": 0.1971,
+      "step": 990
+    },
+    {
+      "epoch": 2.6756393001345895,
+      "grad_norm": 0.9733228087425232,
+      "learning_rate": 1.4533370105733768e-05,
+      "loss": 0.2146,
+      "step": 995
+    },
+    {
+      "epoch": 2.6890982503364738,
+      "grad_norm": 1.0524123907089233,
+      "learning_rate": 1.4400110496213708e-05,
+      "loss": 0.1813,
+      "step": 1000
+    },
+    {
+      "epoch": 2.702557200538358,
+      "grad_norm": 0.796301007270813,
+      "learning_rate": 1.4266898292884239e-05,
+      "loss": 0.1834,
+      "step": 1005
+    },
+    {
+      "epoch": 2.7160161507402423,
+      "grad_norm": 1.0906836986541748,
+      "learning_rate": 1.4133744022822527e-05,
+      "loss": 0.1811,
+      "step": 1010
+    },
+    {
+      "epoch": 2.7294751009421265,
+      "grad_norm": 1.0168821811676025,
+      "learning_rate": 1.400065820852758e-05,
+      "loss": 0.1633,
+      "step": 1015
+    },
+    {
+      "epoch": 2.742934051144011,
+      "grad_norm": 1.0989868640899658,
+      "learning_rate": 1.3867651367088692e-05,
+      "loss": 0.1655,
+      "step": 1020
+    },
+    {
+      "epoch": 2.756393001345895,
+      "grad_norm": 1.2408825159072876,
+      "learning_rate": 1.373473400935433e-05,
+      "loss": 0.1898,
+      "step": 1025
+    },
+    {
+      "epoch": 2.7698519515477793,
+      "grad_norm": 0.952360212802887,
+      "learning_rate": 1.360191663910153e-05,
+      "loss": 0.1727,
+      "step": 1030
+    },
+    {
+      "epoch": 2.7833109017496636,
+      "grad_norm": 1.1368646621704102,
+      "learning_rate": 1.3469209752205834e-05,
+      "loss": 0.1884,
+      "step": 1035
+    },
+    {
+      "epoch": 2.796769851951548,
+      "grad_norm": 0.9177131652832031,
+      "learning_rate": 1.3336623835811851e-05,
+      "loss": 0.1697,
+      "step": 1040
+    },
+    {
+      "epoch": 2.810228802153432,
+      "grad_norm": 1.218162178993225,
+      "learning_rate": 1.3204169367504494e-05,
+      "loss": 0.1875,
+      "step": 1045
+    },
+    {
+      "epoch": 2.8236877523553163,
+      "grad_norm": 1.0008909702301025,
+      "learning_rate": 1.3071856814481031e-05,
+      "loss": 0.1757,
+      "step": 1050
+    },
+    {
+      "epoch": 2.8371467025572006,
+      "grad_norm": 0.9754431843757629,
+      "learning_rate": 1.2939696632723877e-05,
+      "loss": 0.167,
+      "step": 1055
+    },
+    {
+      "epoch": 2.850605652759085,
+      "grad_norm": 1.071413516998291,
+      "learning_rate": 1.2807699266174337e-05,
+      "loss": 0.1689,
+      "step": 1060
+    },
+    {
+      "epoch": 2.864064602960969,
+      "grad_norm": 0.8607226610183716,
+      "learning_rate": 1.2675875145907268e-05,
+      "loss": 0.1641,
+      "step": 1065
+    },
+    {
+      "epoch": 2.8775235531628534,
+      "grad_norm": 1.0274795293807983,
+      "learning_rate": 1.2544234689306743e-05,
+      "loss": 0.1752,
+      "step": 1070
+    },
+    {
+      "epoch": 2.8909825033647376,
+      "grad_norm": 1.068657636642456,
+      "learning_rate": 1.2412788299242854e-05,
+      "loss": 0.181,
+      "step": 1075
+    },
+    {
+      "epoch": 2.904441453566622,
+      "grad_norm": 1.0901589393615723,
+      "learning_rate": 1.228154636324961e-05,
+      "loss": 0.1718,
+      "step": 1080
+    },
+    {
+      "epoch": 2.917900403768506,
+      "grad_norm": 0.9238830804824829,
+      "learning_rate": 1.215051925270406e-05,
+      "loss": 0.1532,
+      "step": 1085
+    },
+    {
+      "epoch": 2.9313593539703904,
+      "grad_norm": 1.0537570714950562,
+      "learning_rate": 1.2019717322006694e-05,
+      "loss": 0.185,
+      "step": 1090
+    },
+    {
+      "epoch": 2.9448183041722746,
+      "grad_norm": 0.9827740788459778,
+      "learning_rate": 1.1889150907763202e-05,
+      "loss": 0.1363,
+      "step": 1095
+    },
+    {
+      "epoch": 2.958277254374159,
+      "grad_norm": 1.0735872983932495,
+      "learning_rate": 1.1758830327967614e-05,
+      "loss": 0.1479,
+      "step": 1100
+    },
+    {
+      "epoch": 2.971736204576043,
+      "grad_norm": 0.90640789270401,
+      "learning_rate": 1.1628765881186924e-05,
+      "loss": 0.1612,
+      "step": 1105
+    },
+    {
+      "epoch": 2.9851951547779274,
+      "grad_norm": 0.9942119717597961,
+      "learning_rate": 1.1498967845747244e-05,
+      "loss": 0.1405,
+      "step": 1110
+    },
+    {
+      "epoch": 2.9986541049798117,
+      "grad_norm": 1.0052928924560547,
+      "learning_rate": 1.136944647892157e-05,
+      "loss": 0.1385,
+      "step": 1115
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1860,
+  "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": 1.587507359343706e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

49_128_e5_3e-5/checkpoint-1116/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:172c033ee011733dcab4f83b130b3ce88d4ead4e1c616867187f00f92a816ae0
+size 7736

49_128_e5_3e-5/checkpoint-1116/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)

49_128_e5_3e-5/checkpoint-1488/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

49_128_e5_3e-5/checkpoint-1488/adapter_config.json

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

49_128_e5_3e-5/checkpoint-1488/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f40693ca7780a30581ed802784d45e6ed1d2f9bcbf9f205ed1cb1ff33b84985c
+size 791751704

49_128_e5_3e-5/checkpoint-1488/latest

@@ -0,0 +1 @@
+global_step1488

49_128_e5_3e-5/checkpoint-1488/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:96403cb07f3aa4f7dbd9a1795f30f7fc47a44e9923e3491a57908b3a3774277e
+size 15920

49_128_e5_3e-5/checkpoint-1488/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ddb7eccf57a5d831b710c9e5d61d35af61a4672bc0764b040ed1cdb0474dd980
+size 15920

49_128_e5_3e-5/checkpoint-1488/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:270267ceb835faa9bdc1d066a9a54ff2e74cc06ba72db9d5e0650d6ecbdaa978
+size 15920

49_128_e5_3e-5/checkpoint-1488/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20cb38a87850105415059cb9aa181da14f38b7585a5af74ba82ee9f8a45ae8e2
+size 15920

49_128_e5_3e-5/checkpoint-1488/rng_state_4.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d64fff785e2e8f58b7a6c4274afaf25f85342ea821cbc5d1c5f1e62ecc6a35b
+size 15920

49_128_e5_3e-5/checkpoint-1488/rng_state_5.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:41ad1920fbc23f61f9e9d64feb97ee35cce53977768225451cc26baa9ca832ee
+size 15920

49_128_e5_3e-5/checkpoint-1488/rng_state_6.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:366509eec805aabd0b8f426b5166443053615ea9ee6c5ffde50f955b07947eb5
+size 15920

49_128_e5_3e-5/checkpoint-1488/rng_state_7.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5132e5d479d948c18f94b0800b0e8b9a3eb73aaa8474be893be7e08d520b433f
+size 15920

49_128_e5_3e-5/checkpoint-1488/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72a17d736eabbfbb8e996575434177d00f935b47be7f54ea391f10bc269c3b2b
+size 1064

49_128_e5_3e-5/checkpoint-1488/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
+  }
+}

49_128_e5_3e-5/checkpoint-1488/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
+}

49_128_e5_3e-5/checkpoint-1488/trainer_state.json

@@ -0,0 +1,2113 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 4.0,
+  "eval_steps": 500,
+  "global_step": 1488,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.013458950201884253,
+      "grad_norm": 1.00712251663208,
+      "learning_rate": 1.2903225806451614e-06,
+      "loss": 1.264,
+      "step": 5
+    },
+    {
+      "epoch": 0.026917900403768506,
+      "grad_norm": 0.9430832266807556,
+      "learning_rate": 2.9032258064516128e-06,
+      "loss": 1.2378,
+      "step": 10
+    },
+    {
+      "epoch": 0.040376850605652756,
+      "grad_norm": 0.6684543490409851,
+      "learning_rate": 4.516129032258065e-06,
+      "loss": 1.239,
+      "step": 15
+    },
+    {
+      "epoch": 0.05383580080753701,
+      "grad_norm": 0.7191868424415588,
+      "learning_rate": 6.129032258064517e-06,
+      "loss": 1.2218,
+      "step": 20
+    },
+    {
+      "epoch": 0.06729475100942127,
+      "grad_norm": 0.6867708563804626,
+      "learning_rate": 7.741935483870968e-06,
+      "loss": 1.2301,
+      "step": 25
+    },
+    {
+      "epoch": 0.08075370121130551,
+      "grad_norm": 0.6379647254943848,
+      "learning_rate": 9.35483870967742e-06,
+      "loss": 1.2194,
+      "step": 30
+    },
+    {
+      "epoch": 0.09421265141318977,
+      "grad_norm": 0.5798360109329224,
+      "learning_rate": 1.0967741935483872e-05,
+      "loss": 1.2002,
+      "step": 35
+    },
+    {
+      "epoch": 0.10767160161507403,
+      "grad_norm": 0.5719746351242065,
+      "learning_rate": 1.2580645161290324e-05,
+      "loss": 1.1982,
+      "step": 40
+    },
+    {
+      "epoch": 0.12113055181695828,
+      "grad_norm": 0.5450358390808105,
+      "learning_rate": 1.4193548387096776e-05,
+      "loss": 1.1986,
+      "step": 45
+    },
+    {
+      "epoch": 0.13458950201884254,
+      "grad_norm": 0.5413178205490112,
+      "learning_rate": 1.5806451612903226e-05,
+      "loss": 1.1955,
+      "step": 50
+    },
+    {
+      "epoch": 0.1480484522207268,
+      "grad_norm": 0.5677768588066101,
+      "learning_rate": 1.741935483870968e-05,
+      "loss": 1.1496,
+      "step": 55
+    },
+    {
+      "epoch": 0.16150740242261102,
+      "grad_norm": 0.5705103278160095,
+      "learning_rate": 1.903225806451613e-05,
+      "loss": 1.1377,
+      "step": 60
+    },
+    {
+      "epoch": 0.17496635262449528,
+      "grad_norm": 0.6191664934158325,
+      "learning_rate": 2.0645161290322582e-05,
+      "loss": 1.1242,
+      "step": 65
+    },
+    {
+      "epoch": 0.18842530282637954,
+      "grad_norm": 0.41417160630226135,
+      "learning_rate": 2.2258064516129034e-05,
+      "loss": 1.0923,
+      "step": 70
+    },
+    {
+      "epoch": 0.2018842530282638,
+      "grad_norm": 0.5877333283424377,
+      "learning_rate": 2.3870967741935483e-05,
+      "loss": 1.1575,
+      "step": 75
+    },
+    {
+      "epoch": 0.21534320323014805,
+      "grad_norm": 0.5006933808326721,
+      "learning_rate": 2.548387096774194e-05,
+      "loss": 1.0873,
+      "step": 80
+    },
+    {
+      "epoch": 0.2288021534320323,
+      "grad_norm": 0.5454208254814148,
+      "learning_rate": 2.7096774193548387e-05,
+      "loss": 1.073,
+      "step": 85
+    },
+    {
+      "epoch": 0.24226110363391656,
+      "grad_norm": 0.4708370268344879,
+      "learning_rate": 2.870967741935484e-05,
+      "loss": 1.1246,
+      "step": 90
+    },
+    {
+      "epoch": 0.2557200538358008,
+      "grad_norm": 0.7446239590644836,
+      "learning_rate": 2.9999976292388853e-05,
+      "loss": 1.0893,
+      "step": 95
+    },
+    {
+      "epoch": 0.2691790040376851,
+      "grad_norm": 0.5205251574516296,
+      "learning_rate": 2.999914653386729e-05,
+      "loss": 1.0233,
+      "step": 100
+    },
+    {
+      "epoch": 0.28263795423956933,
+      "grad_norm": 0.5542618632316589,
+      "learning_rate": 2.9997131469727287e-05,
+      "loss": 1.0534,
+      "step": 105
+    },
+    {
+      "epoch": 0.2960969044414536,
+      "grad_norm": 0.5413870811462402,
+      "learning_rate": 2.9993931259209015e-05,
+      "loss": 1.0217,
+      "step": 110
+    },
+    {
+      "epoch": 0.30955585464333785,
+      "grad_norm": 0.5619310736656189,
+      "learning_rate": 2.998954615520871e-05,
+      "loss": 1.0357,
+      "step": 115
+    },
+    {
+      "epoch": 0.32301480484522205,
+      "grad_norm": 0.5382521152496338,
+      "learning_rate": 2.998397650425863e-05,
+      "loss": 1.0228,
+      "step": 120
+    },
+    {
+      "epoch": 0.3364737550471063,
+      "grad_norm": 0.6172555088996887,
+      "learning_rate": 2.997722274649974e-05,
+      "loss": 0.9711,
+      "step": 125
+    },
+    {
+      "epoch": 0.34993270524899056,
+      "grad_norm": 0.6441250443458557,
+      "learning_rate": 2.9969285415646874e-05,
+      "loss": 0.9977,
+      "step": 130
+    },
+    {
+      "epoch": 0.3633916554508748,
+      "grad_norm": 0.5927826166152954,
+      "learning_rate": 2.9960165138946572e-05,
+      "loss": 0.9778,
+      "step": 135
+    },
+    {
+      "epoch": 0.3768506056527591,
+      "grad_norm": 0.6550005078315735,
+      "learning_rate": 2.9949862637127523e-05,
+      "loss": 0.9335,
+      "step": 140
+    },
+    {
+      "epoch": 0.39030955585464333,
+      "grad_norm": 0.5886467099189758,
+      "learning_rate": 2.9938378724343604e-05,
+      "loss": 0.9681,
+      "step": 145
+    },
+    {
+      "epoch": 0.4037685060565276,
+      "grad_norm": 0.7246305346488953,
+      "learning_rate": 2.992571430810954e-05,
+      "loss": 0.9701,
+      "step": 150
+    },
+    {
+      "epoch": 0.41722745625841184,
+      "grad_norm": 0.7467737197875977,
+      "learning_rate": 2.9911870389229192e-05,
+      "loss": 0.9141,
+      "step": 155
+    },
+    {
+      "epoch": 0.4306864064602961,
+      "grad_norm": 0.7680078744888306,
+      "learning_rate": 2.9896848061716456e-05,
+      "loss": 0.8685,
+      "step": 160
+    },
+    {
+      "epoch": 0.44414535666218036,
+      "grad_norm": 0.7801365256309509,
+      "learning_rate": 2.9880648512708834e-05,
+      "loss": 0.9138,
+      "step": 165
+    },
+    {
+      "epoch": 0.4576043068640646,
+      "grad_norm": 0.6462291479110718,
+      "learning_rate": 2.9863273022373585e-05,
+      "loss": 0.9223,
+      "step": 170
+    },
+    {
+      "epoch": 0.47106325706594887,
+      "grad_norm": 0.6398594975471497,
+      "learning_rate": 2.9844722963806592e-05,
+      "loss": 0.9021,
+      "step": 175
+    },
+    {
+      "epoch": 0.4845222072678331,
+      "grad_norm": 0.683367133140564,
+      "learning_rate": 2.9824999802923846e-05,
+      "loss": 0.8599,
+      "step": 180
+    },
+    {
+      "epoch": 0.4979811574697174,
+      "grad_norm": 0.7672361731529236,
+      "learning_rate": 2.9804105098345575e-05,
+      "loss": 0.9282,
+      "step": 185
+    },
+    {
+      "epoch": 0.5114401076716016,
+      "grad_norm": 0.6266528367996216,
+      "learning_rate": 2.9782040501273126e-05,
+      "loss": 0.8664,
+      "step": 190
+    },
+    {
+      "epoch": 0.5248990578734859,
+      "grad_norm": 0.7748366594314575,
+      "learning_rate": 2.975880775535843e-05,
+      "loss": 0.8588,
+      "step": 195
+    },
+    {
+      "epoch": 0.5383580080753702,
+      "grad_norm": 0.892792820930481,
+      "learning_rate": 2.9734408696566224e-05,
+      "loss": 0.8517,
+      "step": 200
+    },
+    {
+      "epoch": 0.5518169582772544,
+      "grad_norm": 0.8755563497543335,
+      "learning_rate": 2.9708845253028988e-05,
+      "loss": 0.9136,
+      "step": 205
+    },
+    {
+      "epoch": 0.5652759084791387,
+      "grad_norm": 0.7859370708465576,
+      "learning_rate": 2.9682119444894545e-05,
+      "loss": 0.7955,
+      "step": 210
+    },
+    {
+      "epoch": 0.5787348586810229,
+      "grad_norm": 0.7947808504104614,
+      "learning_rate": 2.9654233384166435e-05,
+      "loss": 0.8695,
+      "step": 215
+    },
+    {
+      "epoch": 0.5921938088829072,
+      "grad_norm": 0.7226110696792603,
+      "learning_rate": 2.9625189274537004e-05,
+      "loss": 0.7672,
+      "step": 220
+    },
+    {
+      "epoch": 0.6056527590847914,
+      "grad_norm": 1.0677162408828735,
+      "learning_rate": 2.9594989411213266e-05,
+      "loss": 0.816,
+      "step": 225
+    },
+    {
+      "epoch": 0.6191117092866757,
+      "grad_norm": 0.8201079964637756,
+      "learning_rate": 2.9563636180735524e-05,
+      "loss": 0.7877,
+      "step": 230
+    },
+    {
+      "epoch": 0.6325706594885598,
+      "grad_norm": 0.961162805557251,
+      "learning_rate": 2.9531132060788763e-05,
+      "loss": 0.7566,
+      "step": 235
+    },
+    {
+      "epoch": 0.6460296096904441,
+      "grad_norm": 0.9038267731666565,
+      "learning_rate": 2.9497479620006873e-05,
+      "loss": 0.8061,
+      "step": 240
+    },
+    {
+      "epoch": 0.6594885598923284,
+      "grad_norm": 0.9068340063095093,
+      "learning_rate": 2.9462681517769638e-05,
+      "loss": 0.787,
+      "step": 245
+    },
+    {
+      "epoch": 0.6729475100942126,
+      "grad_norm": 0.8361156582832336,
+      "learning_rate": 2.94267405039926e-05,
+      "loss": 0.7835,
+      "step": 250
+    },
+    {
+      "epoch": 0.6864064602960969,
+      "grad_norm": 0.8174508213996887,
+      "learning_rate": 2.938965941890972e-05,
+      "loss": 0.7641,
+      "step": 255
+    },
+    {
+      "epoch": 0.6998654104979811,
+      "grad_norm": 0.8101845383644104,
+      "learning_rate": 2.9351441192848972e-05,
+      "loss": 0.76,
+      "step": 260
+    },
+    {
+      "epoch": 0.7133243606998654,
+      "grad_norm": 0.8047847747802734,
+      "learning_rate": 2.9312088846000733e-05,
+      "loss": 0.7404,
+      "step": 265
+    },
+    {
+      "epoch": 0.7267833109017496,
+      "grad_norm": 0.8787263035774231,
+      "learning_rate": 2.927160548817913e-05,
+      "loss": 0.7335,
+      "step": 270
+    },
+    {
+      "epoch": 0.7402422611036339,
+      "grad_norm": 0.9566526412963867,
+      "learning_rate": 2.9229994318576295e-05,
+      "loss": 0.7546,
+      "step": 275
+    },
+    {
+      "epoch": 0.7537012113055181,
+      "grad_norm": 0.9340652227401733,
+      "learning_rate": 2.9187258625509518e-05,
+      "loss": 0.6814,
+      "step": 280
+    },
+    {
+      "epoch": 0.7671601615074024,
+      "grad_norm": 1.0249980688095093,
+      "learning_rate": 2.9143401786161445e-05,
+      "loss": 0.7116,
+      "step": 285
+    },
+    {
+      "epoch": 0.7806191117092867,
+      "grad_norm": 0.8947973847389221,
+      "learning_rate": 2.9098427266313138e-05,
+      "loss": 0.6845,
+      "step": 290
+    },
+    {
+      "epoch": 0.7940780619111709,
+      "grad_norm": 0.9918004274368286,
+      "learning_rate": 2.9052338620070215e-05,
+      "loss": 0.7287,
+      "step": 295
+    },
+    {
+      "epoch": 0.8075370121130552,
+      "grad_norm": 0.9002223610877991,
+      "learning_rate": 2.9005139489582002e-05,
+      "loss": 0.7129,
+      "step": 300
+    },
+    {
+      "epoch": 0.8209959623149394,
+      "grad_norm": 0.8825509548187256,
+      "learning_rate": 2.8956833604753688e-05,
+      "loss": 0.6633,
+      "step": 305
+    },
+    {
+      "epoch": 0.8344549125168237,
+      "grad_norm": 0.8505289554595947,
+      "learning_rate": 2.8907424782951587e-05,
+      "loss": 0.6826,
+      "step": 310
+    },
+    {
+      "epoch": 0.847913862718708,
+      "grad_norm": 1.0375418663024902,
+      "learning_rate": 2.8856916928701467e-05,
+      "loss": 0.7003,
+      "step": 315
+    },
+    {
+      "epoch": 0.8613728129205922,
+      "grad_norm": 1.0597079992294312,
+      "learning_rate": 2.8805314033379985e-05,
+      "loss": 0.7151,
+      "step": 320
+    },
+    {
+      "epoch": 0.8748317631224765,
+      "grad_norm": 1.1800124645233154,
+      "learning_rate": 2.8752620174899284e-05,
+      "loss": 0.6594,
+      "step": 325
+    },
+    {
+      "epoch": 0.8882907133243607,
+      "grad_norm": 0.9276929497718811,
+      "learning_rate": 2.8698839517384726e-05,
+      "loss": 0.651,
+      "step": 330
+    },
+    {
+      "epoch": 0.901749663526245,
+      "grad_norm": 1.094962239265442,
+      "learning_rate": 2.8643976310845828e-05,
+      "loss": 0.6412,
+      "step": 335
+    },
+    {
+      "epoch": 0.9152086137281292,
+      "grad_norm": 0.9263754487037659,
+      "learning_rate": 2.8588034890840408e-05,
+      "loss": 0.6269,
+      "step": 340
+    },
+    {
+      "epoch": 0.9286675639300135,
+      "grad_norm": 0.9816908240318298,
+      "learning_rate": 2.8531019678131952e-05,
+      "loss": 0.6577,
+      "step": 345
+    },
+    {
+      "epoch": 0.9421265141318977,
+      "grad_norm": 1.0188283920288086,
+      "learning_rate": 2.847293517834029e-05,
+      "loss": 0.66,
+      "step": 350
+    },
+    {
+      "epoch": 0.955585464333782,
+      "grad_norm": 1.0503355264663696,
+      "learning_rate": 2.8413785981585518e-05,
+      "loss": 0.621,
+      "step": 355
+    },
+    {
+      "epoch": 0.9690444145356663,
+      "grad_norm": 0.9092134237289429,
+      "learning_rate": 2.8353576762125274e-05,
+      "loss": 0.6343,
+      "step": 360
+    },
+    {
+      "epoch": 0.9825033647375505,
+      "grad_norm": 0.9565970301628113,
+      "learning_rate": 2.8292312277985354e-05,
+      "loss": 0.6538,
+      "step": 365
+    },
+    {
+      "epoch": 0.9959623149394348,
+      "grad_norm": 1.014952301979065,
+      "learning_rate": 2.82299973705837e-05,
+      "loss": 0.6081,
+      "step": 370
+    },
+    {
+      "epoch": 1.0080753701211305,
+      "grad_norm": 1.1951587200164795,
+      "learning_rate": 2.816663696434783e-05,
+      "loss": 0.5937,
+      "step": 375
+    },
+    {
+      "epoch": 1.0215343203230147,
+      "grad_norm": 0.9546276926994324,
+      "learning_rate": 2.810223606632567e-05,
+      "loss": 0.5175,
+      "step": 380
+    },
+    {
+      "epoch": 1.034993270524899,
+      "grad_norm": 0.9498586654663086,
+      "learning_rate": 2.803679976578987e-05,
+      "loss": 0.534,
+      "step": 385
+    },
+    {
+      "epoch": 1.0484522207267832,
+      "grad_norm": 1.189976453781128,
+      "learning_rate": 2.797033323383563e-05,
+      "loss": 0.5639,
+      "step": 390
+    },
+    {
+      "epoch": 1.0619111709286675,
+      "grad_norm": 1.038994550704956,
+      "learning_rate": 2.7902841722972062e-05,
+      "loss": 0.5405,
+      "step": 395
+    },
+    {
+      "epoch": 1.0753701211305517,
+      "grad_norm": 1.0657168626785278,
+      "learning_rate": 2.7834330566707086e-05,
+      "loss": 0.5407,
+      "step": 400
+    },
+    {
+      "epoch": 1.088829071332436,
+      "grad_norm": 1.0063115358352661,
+      "learning_rate": 2.7764805179126e-05,
+      "loss": 0.524,
+      "step": 405
+    },
+    {
+      "epoch": 1.1022880215343203,
+      "grad_norm": 0.9580488801002502,
+      "learning_rate": 2.7694271054463565e-05,
+      "loss": 0.5357,
+      "step": 410
+    },
+    {
+      "epoch": 1.1157469717362045,
+      "grad_norm": 1.0191570520401,
+      "learning_rate": 2.762273376666989e-05,
+      "loss": 0.4963,
+      "step": 415
+    },
+    {
+      "epoch": 1.1292059219380888,
+      "grad_norm": 1.0942199230194092,
+      "learning_rate": 2.755019896896989e-05,
+      "loss": 0.4825,
+      "step": 420
+    },
+    {
+      "epoch": 1.142664872139973,
+      "grad_norm": 0.9650993347167969,
+      "learning_rate": 2.7476672393416603e-05,
+      "loss": 0.4862,
+      "step": 425
+    },
+    {
+      "epoch": 1.1561238223418573,
+      "grad_norm": 1.0294479131698608,
+      "learning_rate": 2.7402159850438173e-05,
+      "loss": 0.4972,
+      "step": 430
+    },
+    {
+      "epoch": 1.1695827725437415,
+      "grad_norm": 1.1111294031143188,
+      "learning_rate": 2.7326667228378677e-05,
+      "loss": 0.4757,
+      "step": 435
+    },
+    {
+      "epoch": 1.1830417227456258,
+      "grad_norm": 1.099021077156067,
+      "learning_rate": 2.7250200493032836e-05,
+      "loss": 0.4916,
+      "step": 440
+    },
+    {
+      "epoch": 1.19650067294751,
+      "grad_norm": 1.0875200033187866,
+      "learning_rate": 2.7172765687174536e-05,
+      "loss": 0.5004,
+      "step": 445
+    },
+    {
+      "epoch": 1.2099596231493943,
+      "grad_norm": 1.0877183675765991,
+      "learning_rate": 2.7094368930079312e-05,
+      "loss": 0.4856,
+      "step": 450
+    },
+    {
+      "epoch": 1.2234185733512786,
+      "grad_norm": 1.0833150148391724,
+      "learning_rate": 2.701501641704077e-05,
+      "loss": 0.5081,
+      "step": 455
+    },
+    {
+      "epoch": 1.2368775235531628,
+      "grad_norm": 1.0706162452697754,
+      "learning_rate": 2.693471441888101e-05,
+      "loss": 0.4808,
+      "step": 460
+    },
+    {
+      "epoch": 1.250336473755047,
+      "grad_norm": 1.0003634691238403,
+      "learning_rate": 2.685346928145508e-05,
+      "loss": 0.5242,
+      "step": 465
+    },
+    {
+      "epoch": 1.2637954239569313,
+      "grad_norm": 1.081905722618103,
+      "learning_rate": 2.6771287425149465e-05,
+      "loss": 0.4724,
+      "step": 470
+    },
+    {
+      "epoch": 1.2772543741588156,
+      "grad_norm": 1.0233476161956787,
+      "learning_rate": 2.6688175344374748e-05,
+      "loss": 0.4461,
+      "step": 475
+    },
+    {
+      "epoch": 1.2907133243606999,
+      "grad_norm": 0.9034926295280457,
+      "learning_rate": 2.660413960705239e-05,
+      "loss": 0.5191,
+      "step": 480
+    },
+    {
+      "epoch": 1.304172274562584,
+      "grad_norm": 0.9046325087547302,
+      "learning_rate": 2.651918685409568e-05,
+      "loss": 0.4811,
+      "step": 485
+    },
+    {
+      "epoch": 1.3176312247644684,
+      "grad_norm": 0.9982317686080933,
+      "learning_rate": 2.6433323798884955e-05,
+      "loss": 0.4671,
+      "step": 490
+    },
+    {
+      "epoch": 1.3310901749663526,
+      "grad_norm": 1.2173210382461548,
+      "learning_rate": 2.6346557226737064e-05,
+      "loss": 0.4575,
+      "step": 495
+    },
+    {
+      "epoch": 1.3445491251682369,
+      "grad_norm": 1.0846431255340576,
+      "learning_rate": 2.6258893994369176e-05,
+      "loss": 0.436,
+      "step": 500
+    },
+    {
+      "epoch": 1.3580080753701211,
+      "grad_norm": 1.0849974155426025,
+      "learning_rate": 2.6170341029356906e-05,
+      "loss": 0.4323,
+      "step": 505
+    },
+    {
+      "epoch": 1.3714670255720054,
+      "grad_norm": 1.0231881141662598,
+      "learning_rate": 2.6080905329586883e-05,
+      "loss": 0.4259,
+      "step": 510
+    },
+    {
+      "epoch": 1.3849259757738897,
+      "grad_norm": 1.1137603521347046,
+      "learning_rate": 2.599059396270374e-05,
+      "loss": 0.4539,
+      "step": 515
+    },
+    {
+      "epoch": 1.398384925975774,
+      "grad_norm": 1.066973090171814,
+      "learning_rate": 2.5899414065551578e-05,
+      "loss": 0.4475,
+      "step": 520
+    },
+    {
+      "epoch": 1.4118438761776582,
+      "grad_norm": 1.1666553020477295,
+      "learning_rate": 2.5807372843610002e-05,
+      "loss": 0.436,
+      "step": 525
+    },
+    {
+      "epoch": 1.4253028263795424,
+      "grad_norm": 1.1762378215789795,
+      "learning_rate": 2.5714477570424692e-05,
+      "loss": 0.4234,
+      "step": 530
+    },
+    {
+      "epoch": 1.4387617765814267,
+      "grad_norm": 1.053694486618042,
+      "learning_rate": 2.5620735587032622e-05,
+      "loss": 0.4242,
+      "step": 535
+    },
+    {
+      "epoch": 1.452220726783311,
+      "grad_norm": 1.1577174663543701,
+      "learning_rate": 2.5526154301381926e-05,
+      "loss": 0.408,
+      "step": 540
+    },
+    {
+      "epoch": 1.4656796769851952,
+      "grad_norm": 1.1527884006500244,
+      "learning_rate": 2.5430741187746485e-05,
+      "loss": 0.4154,
+      "step": 545
+    },
+    {
+      "epoch": 1.4791386271870794,
+      "grad_norm": 1.005540132522583,
+      "learning_rate": 2.53345037861353e-05,
+      "loss": 0.4474,
+      "step": 550
+    },
+    {
+      "epoch": 1.4925975773889637,
+      "grad_norm": 1.0806723833084106,
+      "learning_rate": 2.52374497016966e-05,
+      "loss": 0.4015,
+      "step": 555
+    },
+    {
+      "epoch": 1.506056527590848,
+      "grad_norm": 1.0257374048233032,
+      "learning_rate": 2.513958660411688e-05,
+      "loss": 0.4348,
+      "step": 560
+    },
+    {
+      "epoch": 1.5195154777927322,
+      "grad_norm": 0.9938416481018066,
+      "learning_rate": 2.5040922227014787e-05,
+      "loss": 0.4049,
+      "step": 565
+    },
+    {
+      "epoch": 1.5329744279946165,
+      "grad_norm": 1.0034414529800415,
+      "learning_rate": 2.4941464367329993e-05,
+      "loss": 0.4436,
+      "step": 570
+    },
+    {
+      "epoch": 1.5464333781965007,
+      "grad_norm": 1.0372949838638306,
+      "learning_rate": 2.4841220884707014e-05,
+      "loss": 0.3635,
+      "step": 575
+    },
+    {
+      "epoch": 1.559892328398385,
+      "grad_norm": 1.1650141477584839,
+      "learning_rate": 2.4740199700874124e-05,
+      "loss": 0.3933,
+      "step": 580
+    },
+    {
+      "epoch": 1.5733512786002692,
+      "grad_norm": 1.1232426166534424,
+      "learning_rate": 2.4638408799017333e-05,
+      "loss": 0.3762,
+      "step": 585
+    },
+    {
+      "epoch": 1.5868102288021535,
+      "grad_norm": 1.0809695720672607,
+      "learning_rate": 2.4535856223149525e-05,
+      "loss": 0.364,
+      "step": 590
+    },
+    {
+      "epoch": 1.6002691790040378,
+      "grad_norm": 1.0743389129638672,
+      "learning_rate": 2.4432550077474753e-05,
+      "loss": 0.4135,
+      "step": 595
+    },
+    {
+      "epoch": 1.613728129205922,
+      "grad_norm": 1.1608853340148926,
+      "learning_rate": 2.4328498525747862e-05,
+      "loss": 0.373,
+      "step": 600
+    },
+    {
+      "epoch": 1.6271870794078063,
+      "grad_norm": 1.0588221549987793,
+      "learning_rate": 2.4223709790629275e-05,
+      "loss": 0.3898,
+      "step": 605
+    },
+    {
+      "epoch": 1.6406460296096905,
+      "grad_norm": 1.08194100856781,
+      "learning_rate": 2.4118192153035273e-05,
+      "loss": 0.3759,
+      "step": 610
+    },
+    {
+      "epoch": 1.6541049798115748,
+      "grad_norm": 0.9936762452125549,
+      "learning_rate": 2.4011953951483532e-05,
+      "loss": 0.4186,
+      "step": 615
+    },
+    {
+      "epoch": 1.667563930013459,
+      "grad_norm": 1.1843287944793701,
+      "learning_rate": 2.3905003581434227e-05,
+      "loss": 0.3562,
+      "step": 620
+    },
+    {
+      "epoch": 1.6810228802153433,
+      "grad_norm": 1.525776743888855,
+      "learning_rate": 2.379734949462653e-05,
+      "loss": 0.3998,
+      "step": 625
+    },
+    {
+      "epoch": 1.6944818304172276,
+      "grad_norm": 1.056809425354004,
+      "learning_rate": 2.3689000198410762e-05,
+      "loss": 0.4076,
+      "step": 630
+    },
+    {
+      "epoch": 1.7079407806191118,
+      "grad_norm": 1.3260629177093506,
+      "learning_rate": 2.3579964255076064e-05,
+      "loss": 0.3742,
+      "step": 635
+    },
+    {
+      "epoch": 1.721399730820996,
+      "grad_norm": 1.0809186697006226,
+      "learning_rate": 2.3470250281173787e-05,
+      "loss": 0.4176,
+      "step": 640
+    },
+    {
+      "epoch": 1.7348586810228803,
+      "grad_norm": 0.9884938597679138,
+      "learning_rate": 2.3359866946836546e-05,
+      "loss": 0.3475,
+      "step": 645
+    },
+    {
+      "epoch": 1.7483176312247646,
+      "grad_norm": 1.0996809005737305,
+      "learning_rate": 2.3248822975093094e-05,
+      "loss": 0.349,
+      "step": 650
+    },
+    {
+      "epoch": 1.7617765814266488,
+      "grad_norm": 1.3934807777404785,
+      "learning_rate": 2.3137127141178962e-05,
+      "loss": 0.317,
+      "step": 655
+    },
+    {
+      "epoch": 1.775235531628533,
+      "grad_norm": 1.1072278022766113,
+      "learning_rate": 2.3024788271842997e-05,
+      "loss": 0.3442,
+      "step": 660
+    },
+    {
+      "epoch": 1.7886944818304173,
+      "grad_norm": 1.1622191667556763,
+      "learning_rate": 2.2911815244649855e-05,
+      "loss": 0.3707,
+      "step": 665
+    },
+    {
+      "epoch": 1.8021534320323016,
+      "grad_norm": 1.1260006427764893,
+      "learning_rate": 2.2798216987278417e-05,
+      "loss": 0.3506,
+      "step": 670
+    },
+    {
+      "epoch": 1.8156123822341859,
+      "grad_norm": 0.9603449702262878,
+      "learning_rate": 2.2684002476816317e-05,
+      "loss": 0.3696,
+      "step": 675
+    },
+    {
+      "epoch": 1.8290713324360701,
+      "grad_norm": 1.0738201141357422,
+      "learning_rate": 2.2569180739050486e-05,
+      "loss": 0.3615,
+      "step": 680
+    },
+    {
+      "epoch": 1.8425302826379544,
+      "grad_norm": 1.082394003868103,
+      "learning_rate": 2.2453760847753937e-05,
+      "loss": 0.3259,
+      "step": 685
+    },
+    {
+      "epoch": 1.8559892328398386,
+      "grad_norm": 1.0013381242752075,
+      "learning_rate": 2.2337751923968673e-05,
+      "loss": 0.3602,
+      "step": 690
+    },
+    {
+      "epoch": 1.8694481830417229,
+      "grad_norm": 0.9863987565040588,
+      "learning_rate": 2.2221163135284926e-05,
+      "loss": 0.3096,
+      "step": 695
+    },
+    {
+      "epoch": 1.8829071332436071,
+      "grad_norm": 1.0083246231079102,
+      "learning_rate": 2.2104003695116662e-05,
+      "loss": 0.3408,
+      "step": 700
+    },
+    {
+      "epoch": 1.8963660834454914,
+      "grad_norm": 1.1064270734786987,
+      "learning_rate": 2.1986282861973525e-05,
+      "loss": 0.3569,
+      "step": 705
+    },
+    {
+      "epoch": 1.9098250336473757,
+      "grad_norm": 1.062014102935791,
+      "learning_rate": 2.1868009938729145e-05,
+      "loss": 0.3376,
+      "step": 710
+    },
+    {
+      "epoch": 1.92328398384926,
+      "grad_norm": 1.1426458358764648,
+      "learning_rate": 2.174919427188602e-05,
+      "loss": 0.3375,
+      "step": 715
+    },
+    {
+      "epoch": 1.9367429340511442,
+      "grad_norm": 1.0394279956817627,
+      "learning_rate": 2.1629845250836882e-05,
+      "loss": 0.2991,
+      "step": 720
+    },
+    {
+      "epoch": 1.9502018842530284,
+      "grad_norm": 1.1961948871612549,
+      "learning_rate": 2.150997230712271e-05,
+      "loss": 0.3048,
+      "step": 725
+    },
+    {
+      "epoch": 1.9636608344549125,
+      "grad_norm": 1.011620283126831,
+      "learning_rate": 2.138958491368741e-05,
+      "loss": 0.3241,
+      "step": 730
+    },
+    {
+      "epoch": 1.9771197846567967,
+      "grad_norm": 1.053614616394043,
+      "learning_rate": 2.1268692584129197e-05,
+      "loss": 0.3141,
+      "step": 735
+    },
+    {
+      "epoch": 1.990578734858681,
+      "grad_norm": 1.0024995803833008,
+      "learning_rate": 2.1147304871948812e-05,
+      "loss": 0.3211,
+      "step": 740
+    },
+    {
+      "epoch": 2.0026917900403767,
+      "grad_norm": 1.0640769004821777,
+      "learning_rate": 2.1025431369794546e-05,
+      "loss": 0.2784,
+      "step": 745
+    },
+    {
+      "epoch": 2.016150740242261,
+      "grad_norm": 1.1959905624389648,
+      "learning_rate": 2.090308170870416e-05,
+      "loss": 0.2488,
+      "step": 750
+    },
+    {
+      "epoch": 2.029609690444145,
+      "grad_norm": 1.0885748863220215,
+      "learning_rate": 2.0780265557343833e-05,
+      "loss": 0.2466,
+      "step": 755
+    },
+    {
+      "epoch": 2.0430686406460294,
+      "grad_norm": 1.041433334350586,
+      "learning_rate": 2.065699262124407e-05,
+      "loss": 0.2308,
+      "step": 760
+    },
+    {
+      "epoch": 2.0565275908479137,
+      "grad_norm": 1.1424509286880493,
+      "learning_rate": 2.0533272642032734e-05,
+      "loss": 0.2545,
+      "step": 765
+    },
+    {
+      "epoch": 2.069986541049798,
+      "grad_norm": 1.0592752695083618,
+      "learning_rate": 2.0409115396665202e-05,
+      "loss": 0.2373,
+      "step": 770
+    },
+    {
+      "epoch": 2.083445491251682,
+      "grad_norm": 1.3094651699066162,
+      "learning_rate": 2.0284530696651772e-05,
+      "loss": 0.2629,
+      "step": 775
+    },
+    {
+      "epoch": 2.0969044414535665,
+      "grad_norm": 1.171372890472412,
+      "learning_rate": 2.0159528387282273e-05,
+      "loss": 0.2368,
+      "step": 780
+    },
+    {
+      "epoch": 2.1103633916554507,
+      "grad_norm": 1.0821523666381836,
+      "learning_rate": 2.0034118346848057e-05,
+      "loss": 0.2267,
+      "step": 785
+    },
+    {
+      "epoch": 2.123822341857335,
+      "grad_norm": 1.0307716131210327,
+      "learning_rate": 1.9908310485861393e-05,
+      "loss": 0.2223,
+      "step": 790
+    },
+    {
+      "epoch": 2.1372812920592192,
+      "grad_norm": 1.2713267803192139,
+      "learning_rate": 1.9782114746272252e-05,
+      "loss": 0.2493,
+      "step": 795
+    },
+    {
+      "epoch": 2.1507402422611035,
+      "grad_norm": 1.1776148080825806,
+      "learning_rate": 1.965554110068265e-05,
+      "loss": 0.2308,
+      "step": 800
+    },
+    {
+      "epoch": 2.1641991924629878,
+      "grad_norm": 1.030351996421814,
+      "learning_rate": 1.95285995515586e-05,
+      "loss": 0.2718,
+      "step": 805
+    },
+    {
+      "epoch": 2.177658142664872,
+      "grad_norm": 0.9716264009475708,
+      "learning_rate": 1.9401300130439635e-05,
+      "loss": 0.2126,
+      "step": 810
+    },
+    {
+      "epoch": 2.1911170928667563,
+      "grad_norm": 1.125478982925415,
+      "learning_rate": 1.9273652897146063e-05,
+      "loss": 0.2335,
+      "step": 815
+    },
+    {
+      "epoch": 2.2045760430686405,
+      "grad_norm": 1.0152146816253662,
+      "learning_rate": 1.9145667938984034e-05,
+      "loss": 0.2171,
+      "step": 820
+    },
+    {
+      "epoch": 2.218034993270525,
+      "grad_norm": 1.0698730945587158,
+      "learning_rate": 1.9017355369948327e-05,
+      "loss": 0.2297,
+      "step": 825
+    },
+    {
+      "epoch": 2.231493943472409,
+      "grad_norm": 1.0559672117233276,
+      "learning_rate": 1.8888725329923147e-05,
+      "loss": 0.2439,
+      "step": 830
+    },
+    {
+      "epoch": 2.2449528936742933,
+      "grad_norm": 0.9136818051338196,
+      "learning_rate": 1.8759787983880807e-05,
+      "loss": 0.214,
+      "step": 835
+    },
+    {
+      "epoch": 2.2584118438761775,
+      "grad_norm": 1.2098388671875,
+      "learning_rate": 1.8630553521078432e-05,
+      "loss": 0.2305,
+      "step": 840
+    },
+    {
+      "epoch": 2.271870794078062,
+      "grad_norm": 1.0378409624099731,
+      "learning_rate": 1.850103215425276e-05,
+      "loss": 0.2223,
+      "step": 845
+    },
+    {
+      "epoch": 2.285329744279946,
+      "grad_norm": 1.0797078609466553,
+      "learning_rate": 1.8371234118813082e-05,
+      "loss": 0.221,
+      "step": 850
+    },
+    {
+      "epoch": 2.2987886944818303,
+      "grad_norm": 1.0039594173431396,
+      "learning_rate": 1.8241169672032392e-05,
+      "loss": 0.2353,
+      "step": 855
+    },
+    {
+      "epoch": 2.3122476446837146,
+      "grad_norm": 1.3303050994873047,
+      "learning_rate": 1.81108490922368e-05,
+      "loss": 0.2333,
+      "step": 860
+    },
+    {
+      "epoch": 2.325706594885599,
+      "grad_norm": 1.0132439136505127,
+      "learning_rate": 1.7980282677993308e-05,
+      "loss": 0.2138,
+      "step": 865
+    },
+    {
+      "epoch": 2.339165545087483,
+      "grad_norm": 1.0430055856704712,
+      "learning_rate": 1.7849480747295944e-05,
+      "loss": 0.2258,
+      "step": 870
+    },
+    {
+      "epoch": 2.3526244952893673,
+      "grad_norm": 1.0881597995758057,
+      "learning_rate": 1.7718453636750392e-05,
+      "loss": 0.2061,
+      "step": 875
+    },
+    {
+      "epoch": 2.3660834454912516,
+      "grad_norm": 1.0634713172912598,
+      "learning_rate": 1.758721170075715e-05,
+      "loss": 0.2306,
+      "step": 880
+    },
+    {
+      "epoch": 2.379542395693136,
+      "grad_norm": 0.9828234910964966,
+      "learning_rate": 1.7455765310693258e-05,
+      "loss": 0.228,
+      "step": 885
+    },
+    {
+      "epoch": 2.39300134589502,
+      "grad_norm": 1.1163634061813354,
+      "learning_rate": 1.7324124854092738e-05,
+      "loss": 0.2335,
+      "step": 890
+    },
+    {
+      "epoch": 2.4064602960969044,
+      "grad_norm": 1.0143085718154907,
+      "learning_rate": 1.7192300733825664e-05,
+      "loss": 0.2164,
+      "step": 895
+    },
+    {
+      "epoch": 2.4199192462987886,
+      "grad_norm": 1.1971455812454224,
+      "learning_rate": 1.7060303367276123e-05,
+      "loss": 0.2146,
+      "step": 900
+    },
+    {
+      "epoch": 2.433378196500673,
+      "grad_norm": 1.0438719987869263,
+      "learning_rate": 1.692814318551897e-05,
+      "loss": 0.212,
+      "step": 905
+    },
+    {
+      "epoch": 2.446837146702557,
+      "grad_norm": 0.9495139718055725,
+      "learning_rate": 1.6795830632495512e-05,
+      "loss": 0.2124,
+      "step": 910
+    },
+    {
+      "epoch": 2.4602960969044414,
+      "grad_norm": 1.0062249898910522,
+      "learning_rate": 1.6663376164188153e-05,
+      "loss": 0.2048,
+      "step": 915
+    },
+    {
+      "epoch": 2.4737550471063257,
+      "grad_norm": 1.0685555934906006,
+      "learning_rate": 1.6530790247794163e-05,
+      "loss": 0.1982,
+      "step": 920
+    },
+    {
+      "epoch": 2.48721399730821,
+      "grad_norm": 1.1590561866760254,
+      "learning_rate": 1.6398083360898474e-05,
+      "loss": 0.2194,
+      "step": 925
+    },
+    {
+      "epoch": 2.500672947510094,
+      "grad_norm": 0.9140127301216125,
+      "learning_rate": 1.626526599064567e-05,
+      "loss": 0.185,
+      "step": 930
+    },
+    {
+      "epoch": 2.5141318977119784,
+      "grad_norm": 1.0982829332351685,
+      "learning_rate": 1.613234863291131e-05,
+      "loss": 0.2175,
+      "step": 935
+    },
+    {
+      "epoch": 2.5275908479138627,
+      "grad_norm": 0.9458792209625244,
+      "learning_rate": 1.599934179147242e-05,
+      "loss": 0.1872,
+      "step": 940
+    },
+    {
+      "epoch": 2.541049798115747,
+      "grad_norm": 1.1163727045059204,
+      "learning_rate": 1.5866255977177477e-05,
+      "loss": 0.1964,
+      "step": 945
+    },
+    {
+      "epoch": 2.554508748317631,
+      "grad_norm": 1.2947419881820679,
+      "learning_rate": 1.5733101707115765e-05,
+      "loss": 0.1967,
+      "step": 950
+    },
+    {
+      "epoch": 2.5679676985195155,
+      "grad_norm": 1.181272268295288,
+      "learning_rate": 1.5599889503786296e-05,
+      "loss": 0.1856,
+      "step": 955
+    },
+    {
+      "epoch": 2.5814266487213997,
+      "grad_norm": 1.0785599946975708,
+      "learning_rate": 1.5466629894266233e-05,
+      "loss": 0.1932,
+      "step": 960
+    },
+    {
+      "epoch": 2.594885598923284,
+      "grad_norm": 2.4438836574554443,
+      "learning_rate": 1.5333333409379017e-05,
+      "loss": 0.1884,
+      "step": 965
+    },
+    {
+      "epoch": 2.608344549125168,
+      "grad_norm": 0.9497862458229065,
+      "learning_rate": 1.5200010582862162e-05,
+      "loss": 0.1836,
+      "step": 970
+    },
+    {
+      "epoch": 2.6218034993270525,
+      "grad_norm": 1.2000949382781982,
+      "learning_rate": 1.5066671950534817e-05,
+      "loss": 0.2057,
+      "step": 975
+    },
+    {
+      "epoch": 2.6352624495289367,
+      "grad_norm": 1.074912190437317,
+      "learning_rate": 1.493332804946519e-05,
+      "loss": 0.1973,
+      "step": 980
+    },
+    {
+      "epoch": 2.648721399730821,
+      "grad_norm": 1.1403034925460815,
+      "learning_rate": 1.4799989417137838e-05,
+      "loss": 0.1692,
+      "step": 985
+    },
+    {
+      "epoch": 2.6621803499327052,
+      "grad_norm": 1.1136847734451294,
+      "learning_rate": 1.4666666590620982e-05,
+      "loss": 0.1971,
+      "step": 990
+    },
+    {
+      "epoch": 2.6756393001345895,
+      "grad_norm": 0.9733228087425232,
+      "learning_rate": 1.4533370105733768e-05,
+      "loss": 0.2146,
+      "step": 995
+    },
+    {
+      "epoch": 2.6890982503364738,
+      "grad_norm": 1.0524123907089233,
+      "learning_rate": 1.4400110496213708e-05,
+      "loss": 0.1813,
+      "step": 1000
+    },
+    {
+      "epoch": 2.702557200538358,
+      "grad_norm": 0.796301007270813,
+      "learning_rate": 1.4266898292884239e-05,
+      "loss": 0.1834,
+      "step": 1005
+    },
+    {
+      "epoch": 2.7160161507402423,
+      "grad_norm": 1.0906836986541748,
+      "learning_rate": 1.4133744022822527e-05,
+      "loss": 0.1811,
+      "step": 1010
+    },
+    {
+      "epoch": 2.7294751009421265,
+      "grad_norm": 1.0168821811676025,
+      "learning_rate": 1.400065820852758e-05,
+      "loss": 0.1633,
+      "step": 1015
+    },
+    {
+      "epoch": 2.742934051144011,
+      "grad_norm": 1.0989868640899658,
+      "learning_rate": 1.3867651367088692e-05,
+      "loss": 0.1655,
+      "step": 1020
+    },
+    {
+      "epoch": 2.756393001345895,
+      "grad_norm": 1.2408825159072876,
+      "learning_rate": 1.373473400935433e-05,
+      "loss": 0.1898,
+      "step": 1025
+    },
+    {
+      "epoch": 2.7698519515477793,
+      "grad_norm": 0.952360212802887,
+      "learning_rate": 1.360191663910153e-05,
+      "loss": 0.1727,
+      "step": 1030
+    },
+    {
+      "epoch": 2.7833109017496636,
+      "grad_norm": 1.1368646621704102,
+      "learning_rate": 1.3469209752205834e-05,
+      "loss": 0.1884,
+      "step": 1035
+    },
+    {
+      "epoch": 2.796769851951548,
+      "grad_norm": 0.9177131652832031,
+      "learning_rate": 1.3336623835811851e-05,
+      "loss": 0.1697,
+      "step": 1040
+    },
+    {
+      "epoch": 2.810228802153432,
+      "grad_norm": 1.218162178993225,
+      "learning_rate": 1.3204169367504494e-05,
+      "loss": 0.1875,
+      "step": 1045
+    },
+    {
+      "epoch": 2.8236877523553163,
+      "grad_norm": 1.0008909702301025,
+      "learning_rate": 1.3071856814481031e-05,
+      "loss": 0.1757,
+      "step": 1050
+    },
+    {
+      "epoch": 2.8371467025572006,
+      "grad_norm": 0.9754431843757629,
+      "learning_rate": 1.2939696632723877e-05,
+      "loss": 0.167,
+      "step": 1055
+    },
+    {
+      "epoch": 2.850605652759085,
+      "grad_norm": 1.071413516998291,
+      "learning_rate": 1.2807699266174337e-05,
+      "loss": 0.1689,
+      "step": 1060
+    },
+    {
+      "epoch": 2.864064602960969,
+      "grad_norm": 0.8607226610183716,
+      "learning_rate": 1.2675875145907268e-05,
+      "loss": 0.1641,
+      "step": 1065
+    },
+    {
+      "epoch": 2.8775235531628534,
+      "grad_norm": 1.0274795293807983,
+      "learning_rate": 1.2544234689306743e-05,
+      "loss": 0.1752,
+      "step": 1070
+    },
+    {
+      "epoch": 2.8909825033647376,
+      "grad_norm": 1.068657636642456,
+      "learning_rate": 1.2412788299242854e-05,
+      "loss": 0.181,
+      "step": 1075
+    },
+    {
+      "epoch": 2.904441453566622,
+      "grad_norm": 1.0901589393615723,
+      "learning_rate": 1.228154636324961e-05,
+      "loss": 0.1718,
+      "step": 1080
+    },
+    {
+      "epoch": 2.917900403768506,
+      "grad_norm": 0.9238830804824829,
+      "learning_rate": 1.215051925270406e-05,
+      "loss": 0.1532,
+      "step": 1085
+    },
+    {
+      "epoch": 2.9313593539703904,
+      "grad_norm": 1.0537570714950562,
+      "learning_rate": 1.2019717322006694e-05,
+      "loss": 0.185,
+      "step": 1090
+    },
+    {
+      "epoch": 2.9448183041722746,
+      "grad_norm": 0.9827740788459778,
+      "learning_rate": 1.1889150907763202e-05,
+      "loss": 0.1363,
+      "step": 1095
+    },
+    {
+      "epoch": 2.958277254374159,
+      "grad_norm": 1.0735872983932495,
+      "learning_rate": 1.1758830327967614e-05,
+      "loss": 0.1479,
+      "step": 1100
+    },
+    {
+      "epoch": 2.971736204576043,
+      "grad_norm": 0.90640789270401,
+      "learning_rate": 1.1628765881186924e-05,
+      "loss": 0.1612,
+      "step": 1105
+    },
+    {
+      "epoch": 2.9851951547779274,
+      "grad_norm": 0.9942119717597961,
+      "learning_rate": 1.1498967845747244e-05,
+      "loss": 0.1405,
+      "step": 1110
+    },
+    {
+      "epoch": 2.9986541049798117,
+      "grad_norm": 1.0052928924560547,
+      "learning_rate": 1.136944647892157e-05,
+      "loss": 0.1385,
+      "step": 1115
+    },
+    {
+      "epoch": 3.0107671601615076,
+      "grad_norm": 0.8724901676177979,
+      "learning_rate": 1.124021201611919e-05,
+      "loss": 0.1302,
+      "step": 1120
+    },
+    {
+      "epoch": 3.024226110363392,
+      "grad_norm": 0.9571249485015869,
+      "learning_rate": 1.1111274670076859e-05,
+      "loss": 0.1148,
+      "step": 1125
+    },
+    {
+      "epoch": 3.037685060565276,
+      "grad_norm": 0.9960916638374329,
+      "learning_rate": 1.0982644630051677e-05,
+      "loss": 0.1249,
+      "step": 1130
+    },
+    {
+      "epoch": 3.0511440107671604,
+      "grad_norm": 1.057931900024414,
+      "learning_rate": 1.085433206101597e-05,
+      "loss": 0.1269,
+      "step": 1135
+    },
+    {
+      "epoch": 3.0646029609690446,
+      "grad_norm": 0.8812566995620728,
+      "learning_rate": 1.0726347102853935e-05,
+      "loss": 0.1352,
+      "step": 1140
+    },
+    {
+      "epoch": 3.078061911170929,
+      "grad_norm": 0.8664368987083435,
+      "learning_rate": 1.0598699869560371e-05,
+      "loss": 0.1048,
+      "step": 1145
+    },
+    {
+      "epoch": 3.091520861372813,
+      "grad_norm": 0.8292278051376343,
+      "learning_rate": 1.0471400448441403e-05,
+      "loss": 0.1167,
+      "step": 1150
+    },
+    {
+      "epoch": 3.1049798115746974,
+      "grad_norm": 0.7671526074409485,
+      "learning_rate": 1.0344458899317353e-05,
+      "loss": 0.1248,
+      "step": 1155
+    },
+    {
+      "epoch": 3.1184387617765816,
+      "grad_norm": 0.8903627395629883,
+      "learning_rate": 1.021788525372775e-05,
+      "loss": 0.1109,
+      "step": 1160
+    },
+    {
+      "epoch": 3.131897711978466,
+      "grad_norm": 1.0957462787628174,
+      "learning_rate": 1.009168951413861e-05,
+      "loss": 0.1266,
+      "step": 1165
+    },
+    {
+      "epoch": 3.14535666218035,
+      "grad_norm": 1.0637989044189453,
+      "learning_rate": 9.965881653151946e-06,
+      "loss": 0.1122,
+      "step": 1170
+    },
+    {
+      "epoch": 3.1588156123822344,
+      "grad_norm": 0.9090270400047302,
+      "learning_rate": 9.84047161271773e-06,
+      "loss": 0.1163,
+      "step": 1175
+    },
+    {
+      "epoch": 3.1722745625841187,
+      "grad_norm": 0.759506344795227,
+      "learning_rate": 9.715469303348228e-06,
+      "loss": 0.118,
+      "step": 1180
+    },
+    {
+      "epoch": 3.185733512786003,
+      "grad_norm": 0.9539629817008972,
+      "learning_rate": 9.590884603334797e-06,
+      "loss": 0.1263,
+      "step": 1185
+    },
+    {
+      "epoch": 3.199192462987887,
+      "grad_norm": 1.0537840127944946,
+      "learning_rate": 9.466727357967273e-06,
+      "loss": 0.1181,
+      "step": 1190
+    },
+    {
+      "epoch": 3.2126514131897714,
+      "grad_norm": 0.8968941569328308,
+      "learning_rate": 9.343007378755933e-06,
+      "loss": 0.1174,
+      "step": 1195
+    },
+    {
+      "epoch": 3.2261103633916552,
+      "grad_norm": 0.865739107131958,
+      "learning_rate": 9.219734442656171e-06,
+      "loss": 0.1154,
+      "step": 1200
+    },
+    {
+      "epoch": 3.2395693135935395,
+      "grad_norm": 1.0829298496246338,
+      "learning_rate": 9.096918291295842e-06,
+      "loss": 0.1054,
+      "step": 1205
+    },
+    {
+      "epoch": 3.253028263795424,
+      "grad_norm": 1.0235211849212646,
+      "learning_rate": 8.974568630205462e-06,
+      "loss": 0.1203,
+      "step": 1210
+    },
+    {
+      "epoch": 3.2664872139973085,
+      "grad_norm": 0.8741592168807983,
+      "learning_rate": 8.852695128051192e-06,
+      "loss": 0.1228,
+      "step": 1215
+    },
+    {
+      "epoch": 3.2799461641991923,
+      "grad_norm": 0.9431548714637756,
+      "learning_rate": 8.731307415870807e-06,
+      "loss": 0.1099,
+      "step": 1220
+    },
+    {
+      "epoch": 3.2934051144010765,
+      "grad_norm": 0.9976819753646851,
+      "learning_rate": 8.610415086312595e-06,
+      "loss": 0.1171,
+      "step": 1225
+    },
+    {
+      "epoch": 3.306864064602961,
+      "grad_norm": 0.817635715007782,
+      "learning_rate": 8.49002769287729e-06,
+      "loss": 0.1236,
+      "step": 1230
+    },
+    {
+      "epoch": 3.320323014804845,
+      "grad_norm": 0.8399416208267212,
+      "learning_rate": 8.370154749163117e-06,
+      "loss": 0.1211,
+      "step": 1235
+    },
+    {
+      "epoch": 3.3337819650067293,
+      "grad_norm": 0.9999458193778992,
+      "learning_rate": 8.25080572811398e-06,
+      "loss": 0.1174,
+      "step": 1240
+    },
+    {
+      "epoch": 3.3472409152086136,
+      "grad_norm": 0.8543193340301514,
+      "learning_rate": 8.131990061270854e-06,
+      "loss": 0.1047,
+      "step": 1245
+    },
+    {
+      "epoch": 3.360699865410498,
+      "grad_norm": 0.9358410239219666,
+      "learning_rate": 8.013717138026478e-06,
+      "loss": 0.1057,
+      "step": 1250
+    },
+    {
+      "epoch": 3.374158815612382,
+      "grad_norm": 0.925000786781311,
+      "learning_rate": 7.895996304883337e-06,
+      "loss": 0.11,
+      "step": 1255
+    },
+    {
+      "epoch": 3.3876177658142663,
+      "grad_norm": 0.9305988550186157,
+      "learning_rate": 7.778836864715076e-06,
+      "loss": 0.1082,
+      "step": 1260
+    },
+    {
+      "epoch": 3.4010767160161506,
+      "grad_norm": 0.9377626776695251,
+      "learning_rate": 7.662248076031326e-06,
+      "loss": 0.1036,
+      "step": 1265
+    },
+    {
+      "epoch": 3.414535666218035,
+      "grad_norm": 0.8918657898902893,
+      "learning_rate": 7.54623915224607e-06,
+      "loss": 0.0999,
+      "step": 1270
+    },
+    {
+      "epoch": 3.427994616419919,
+      "grad_norm": 0.970435619354248,
+      "learning_rate": 7.43081926094952e-06,
+      "loss": 0.1191,
+      "step": 1275
+    },
+    {
+      "epoch": 3.4414535666218034,
+      "grad_norm": 0.8666785955429077,
+      "learning_rate": 7.3159975231836905e-06,
+      "loss": 0.1115,
+      "step": 1280
+    },
+    {
+      "epoch": 3.4549125168236876,
+      "grad_norm": 1.0095986127853394,
+      "learning_rate": 7.20178301272158e-06,
+      "loss": 0.1125,
+      "step": 1285
+    },
+    {
+      "epoch": 3.468371467025572,
+      "grad_norm": 1.0707430839538574,
+      "learning_rate": 7.088184755350149e-06,
+      "loss": 0.1153,
+      "step": 1290
+    },
+    {
+      "epoch": 3.481830417227456,
+      "grad_norm": 0.8752557635307312,
+      "learning_rate": 6.975211728157007e-06,
+      "loss": 0.1062,
+      "step": 1295
+    },
+    {
+      "epoch": 3.4952893674293404,
+      "grad_norm": 0.9246574640274048,
+      "learning_rate": 6.862872858821044e-06,
+      "loss": 0.0944,
+      "step": 1300
+    },
+    {
+      "epoch": 3.5087483176312246,
+      "grad_norm": 0.9313935041427612,
+      "learning_rate": 6.7511770249069075e-06,
+      "loss": 0.1095,
+      "step": 1305
+    },
+    {
+      "epoch": 3.522207267833109,
+      "grad_norm": 1.132778286933899,
+      "learning_rate": 6.640133053163455e-06,
+      "loss": 0.0906,
+      "step": 1310
+    },
+    {
+      "epoch": 3.535666218034993,
+      "grad_norm": 0.7846031188964844,
+      "learning_rate": 6.529749718826216e-06,
+      "loss": 0.1022,
+      "step": 1315
+    },
+    {
+      "epoch": 3.5491251682368774,
+      "grad_norm": 0.9846947193145752,
+      "learning_rate": 6.420035744923936e-06,
+      "loss": 0.1171,
+      "step": 1320
+    },
+    {
+      "epoch": 3.5625841184387617,
+      "grad_norm": 0.9583669304847717,
+      "learning_rate": 6.310999801589238e-06,
+      "loss": 0.0951,
+      "step": 1325
+    },
+    {
+      "epoch": 3.576043068640646,
+      "grad_norm": 0.8905978202819824,
+      "learning_rate": 6.202650505373471e-06,
+      "loss": 0.1113,
+      "step": 1330
+    },
+    {
+      "epoch": 3.58950201884253,
+      "grad_norm": 0.9217025637626648,
+      "learning_rate": 6.0949964185657795e-06,
+      "loss": 0.108,
+      "step": 1335
+    },
+    {
+      "epoch": 3.6029609690444144,
+      "grad_norm": 0.8374770879745483,
+      "learning_rate": 5.988046048516465e-06,
+      "loss": 0.0974,
+      "step": 1340
+    },
+    {
+      "epoch": 3.6164199192462987,
+      "grad_norm": 0.8375626802444458,
+      "learning_rate": 5.8818078469647265e-06,
+      "loss": 0.1055,
+      "step": 1345
+    },
+    {
+      "epoch": 3.629878869448183,
+      "grad_norm": 0.8715273141860962,
+      "learning_rate": 5.776290209370721e-06,
+      "loss": 0.0995,
+      "step": 1350
+    },
+    {
+      "epoch": 3.643337819650067,
+      "grad_norm": 0.9106976985931396,
+      "learning_rate": 5.6715014742521455e-06,
+      "loss": 0.0927,
+      "step": 1355
+    },
+    {
+      "epoch": 3.6567967698519515,
+      "grad_norm": 0.8811970949172974,
+      "learning_rate": 5.567449922525249e-06,
+      "loss": 0.0894,
+      "step": 1360
+    },
+    {
+      "epoch": 3.6702557200538357,
+      "grad_norm": 0.7267932295799255,
+      "learning_rate": 5.464143776850483e-06,
+      "loss": 0.0891,
+      "step": 1365
+    },
+    {
+      "epoch": 3.68371467025572,
+      "grad_norm": 0.855301558971405,
+      "learning_rate": 5.361591200982665e-06,
+      "loss": 0.0945,
+      "step": 1370
+    },
+    {
+      "epoch": 3.6971736204576042,
+      "grad_norm": 0.9464206099510193,
+      "learning_rate": 5.25980029912588e-06,
+      "loss": 0.1059,
+      "step": 1375
+    },
+    {
+      "epoch": 3.7106325706594885,
+      "grad_norm": 1.0975927114486694,
+      "learning_rate": 5.15877911529299e-06,
+      "loss": 0.1017,
+      "step": 1380
+    },
+    {
+      "epoch": 3.7240915208613727,
+      "grad_norm": 0.8893710970878601,
+      "learning_rate": 5.05853563267001e-06,
+      "loss": 0.1205,
+      "step": 1385
+    },
+    {
+      "epoch": 3.737550471063257,
+      "grad_norm": 0.880164384841919,
+      "learning_rate": 4.959077772985215e-06,
+      "loss": 0.1,
+      "step": 1390
+    },
+    {
+      "epoch": 3.7510094212651413,
+      "grad_norm": 1.031550407409668,
+      "learning_rate": 4.860413395883125e-06,
+      "loss": 0.097,
+      "step": 1395
+    },
+    {
+      "epoch": 3.7644683714670255,
+      "grad_norm": 0.9714354872703552,
+      "learning_rate": 4.762550298303403e-06,
+      "loss": 0.1017,
+      "step": 1400
+    },
+    {
+      "epoch": 3.7779273216689098,
+      "grad_norm": 0.816198468208313,
+      "learning_rate": 4.6654962138647005e-06,
+      "loss": 0.0914,
+      "step": 1405
+    },
+    {
+      "epoch": 3.791386271870794,
+      "grad_norm": 0.8618583083152771,
+      "learning_rate": 4.569258812253513e-06,
+      "loss": 0.0944,
+      "step": 1410
+    },
+    {
+      "epoch": 3.8048452220726783,
+      "grad_norm": 0.8077223300933838,
+      "learning_rate": 4.473845698618082e-06,
+      "loss": 0.088,
+      "step": 1415
+    },
+    {
+      "epoch": 3.8183041722745625,
+      "grad_norm": 0.805504560470581,
+      "learning_rate": 4.3792644129673834e-06,
+      "loss": 0.0852,
+      "step": 1420
+    },
+    {
+      "epoch": 3.831763122476447,
+      "grad_norm": 0.7852038741111755,
+      "learning_rate": 4.285522429575311e-06,
+      "loss": 0.0976,
+      "step": 1425
+    },
+    {
+      "epoch": 3.845222072678331,
+      "grad_norm": 0.730466365814209,
+      "learning_rate": 4.192627156389999e-06,
+      "loss": 0.096,
+      "step": 1430
+    },
+    {
+      "epoch": 3.8586810228802153,
+      "grad_norm": 0.8835721611976624,
+      "learning_rate": 4.100585934448427e-06,
+      "loss": 0.086,
+      "step": 1435
+    },
+    {
+      "epoch": 3.8721399730820996,
+      "grad_norm": 0.8414895534515381,
+      "learning_rate": 4.009406037296266e-06,
+      "loss": 0.0896,
+      "step": 1440
+    },
+    {
+      "epoch": 3.885598923283984,
+      "grad_norm": 0.7949928641319275,
+      "learning_rate": 3.919094670413119e-06,
+      "loss": 0.0975,
+      "step": 1445
+    },
+    {
+      "epoch": 3.899057873485868,
+      "grad_norm": 0.8585901856422424,
+      "learning_rate": 3.829658970643095e-06,
+      "loss": 0.0841,
+      "step": 1450
+    },
+    {
+      "epoch": 3.9125168236877523,
+      "grad_norm": 0.9198737144470215,
+      "learning_rate": 3.7411060056308262e-06,
+      "loss": 0.0874,
+      "step": 1455
+    },
+    {
+      "epoch": 3.9259757738896366,
+      "grad_norm": 0.8871223330497742,
+      "learning_rate": 3.6534427732629365e-06,
+      "loss": 0.0923,
+      "step": 1460
+    },
+    {
+      "epoch": 3.939434724091521,
+      "grad_norm": 0.9608241319656372,
+      "learning_rate": 3.566676201115047e-06,
+      "loss": 0.088,
+      "step": 1465
+    },
+    {
+      "epoch": 3.952893674293405,
+      "grad_norm": 0.9304593205451965,
+      "learning_rate": 3.4808131459043202e-06,
+      "loss": 0.0956,
+      "step": 1470
+    },
+    {
+      "epoch": 3.9663526244952894,
+      "grad_norm": 0.892681360244751,
+      "learning_rate": 3.3958603929476108e-06,
+      "loss": 0.0867,
+      "step": 1475
+    },
+    {
+      "epoch": 3.9798115746971736,
+      "grad_norm": 0.820776641368866,
+      "learning_rate": 3.311824655625257e-06,
+      "loss": 0.0834,
+      "step": 1480
+    },
+    {
+      "epoch": 3.993270524899058,
+      "grad_norm": 0.9003862738609314,
+      "learning_rate": 3.228712574850538e-06,
+      "loss": 0.0924,
+      "step": 1485
+    }
+  ],
+  "logging_steps": 5,
+  "max_steps": 1860,
+  "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": 2.1176579011190456e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}

49_128_e5_3e-5/checkpoint-1488/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:172c033ee011733dcab4f83b130b3ce88d4ead4e1c616867187f00f92a816ae0
+size 7736

49_128_e5_3e-5/checkpoint-1488/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)

49_128_e5_3e-5/checkpoint-1860/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

49_128_e5_3e-5/checkpoint-1860/adapter_config.json

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

49_128_e5_3e-5/checkpoint-1860/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:80554b76d9846502b67c3631929174bd0f73fa6b8134d050a80061bab87cf443
+size 791751704

49_128_e5_3e-5/checkpoint-1860/latest

@@ -0,0 +1 @@
+global_step1860

49_128_e5_3e-5/checkpoint-1860/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5e0488a50e70ab88f9fd6516a2f24985a8e67087bb4c680f995b726da1886f62
+size 15920

49_128_e5_3e-5/checkpoint-1860/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:467f6bf0e4987240db78a615e8b6effe4b298c593a2872ab30acbe0f5c0714cf
+size 15920

49_128_e5_3e-5/checkpoint-1860/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a9f6351c1aaf7f580daa6110e10f808c641eac8e017bb78038e843bbe2741595
+size 15920