Upload folder using huggingface_hub

README.md

@@ -0,0 +1,60 @@
+---
+library_name: peft
+license: other
+base_model: unsloth/mistral-7b-instruct-v0.3-bnb-4bit
+tags:
+- llama-factory
+- lora
+- unsloth
+- generated_from_trainer
+model-index:
+- name: mistral-treinado
+  results: []
+---
+
+<!-- This model card has been generated automatically according to the information the Trainer had access to. You
+should probably proofread and complete it, then remove this comment. -->
+
+# mistral-treinado
+
+This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.3](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3) on the treino_pt_rde dataset.
+
+## Model description
+
+More information needed
+
+## Intended uses & limitations
+
+More information needed
+
+## Training and evaluation data
+
+More information needed
+
+## Training procedure
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+- learning_rate: 3e-05
+- train_batch_size: 4
+- eval_batch_size: 8
+- seed: 42
+- gradient_accumulation_steps: 2
+- total_train_batch_size: 8
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- lr_scheduler_type: cosine
+- num_epochs: 3.0
+- mixed_precision_training: Native AMP
+
+### Training results
+
+
+
+### Framework versions
+
+- PEFT 0.15.2
+- Transformers 4.52.4
+- Pytorch 2.8.0+cu126
+- Datasets 3.6.0
+- Tokenizers 0.21.1

adapter_config.json

@@ -0,0 +1,43 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": {
+    "base_model_class": "MistralForCausalLM",
+    "parent_library": "transformers.models.mistral.modeling_mistral",
+    "unsloth_fixed": true
+  },
+  "base_model_name_or_path": "unsloth/mistral-7b-instruct-v0.3-bnb-4bit",
+  "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": 16,
+  "lora_bias": false,
+  "lora_dropout": 0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "o_proj",
+    "gate_proj",
+    "q_proj",
+    "k_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a84ee66abe7f9a3129784babc9b46656fbe76623efe7131228f04bdcd0f3f973
+size 83945296

all_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "num_input_tokens_seen": 5107072,
+    "total_flos": 2.186269295986606e+17,
+    "train_loss": 0.10808860245144068,
+    "train_runtime": 1785.3385,
+    "train_samples_per_second": 6.244,
+    "train_steps_per_second": 0.781
+}

chat_template.jinja

@@ -0,0 +1,9 @@
+{% set system_message = 'Below is an instruction that describes a task. Write a response that appropriately completes the request.
+
+' %}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '### Instruction:
+' + content + '
+
+### Response:
+' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' + '
+
+' }}{% endif %}{% endfor %}

checkpoint-1000/README.md

@@ -0,0 +1,202 @@
+---
+base_model: unsloth/mistral-7b-instruct-v0.3-bnb-4bit
+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

checkpoint-1000/adapter_config.json

@@ -0,0 +1,43 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": {
+    "base_model_class": "MistralForCausalLM",
+    "parent_library": "transformers.models.mistral.modeling_mistral",
+    "unsloth_fixed": true
+  },
+  "base_model_name_or_path": "unsloth/mistral-7b-instruct-v0.3-bnb-4bit",
+  "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": 16,
+  "lora_bias": false,
+  "lora_dropout": 0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "o_proj",
+    "gate_proj",
+    "q_proj",
+    "k_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

checkpoint-1000/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4df1f6b0a6d08b217b23c9b4bf6ac39c63a7a6596edc0d76bc9167ba82cfb243
+size 83945296

checkpoint-1000/chat_template.jinja

@@ -0,0 +1,9 @@
+{% set system_message = 'Below is an instruction that describes a task. Write a response that appropriately completes the request.
+
+' %}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '### Instruction:
+' + content + '
+
+### Response:
+' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' + '
+
+' }}{% endif %}{% endfor %}

checkpoint-1000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b7c261994bcdf9245bb690632e5cd19f23a3a27c49f6298350622190e81742de
+size 168149539

checkpoint-1000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2534e434cd5abbb8f7668d3eab0549db0ef95d6a797a3efa86b712e8e32266a7
+size 14645

checkpoint-1000/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:14ae2a2128444abab378aa06c09a61a84665f758fcc19fc46f5789b0bc1b5665
+size 1383

checkpoint-1000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:532ee5f4d9f2124464d731be496cbed9ce4a0bf450586fd2cbd04728b54d34bc
+size 1465

checkpoint-1000/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-1000/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37f00374dea48658ee8f5d0f21895b9bc55cb0103939607c8185bfd1c6ca1f89
+size 587404

checkpoint-1000/trainer_state.json

@@ -0,0 +1,1034 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.1506996770721205,
+  "eval_steps": 500,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.021528525296017224,
+      "grad_norm": 0.000117536197649315,
+      "learning_rate": 2.9996919063517692e-05,
+      "loss": 0.8869,
+      "num_input_tokens_seen": 36480,
+      "step": 10,
+      "train_runtime": 21.2665,
+      "train_tokens_per_second": 1715.375
+    },
+    {
+      "epoch": 0.04305705059203445,
+      "grad_norm": 0.00022027654631529003,
+      "learning_rate": 2.9986270537572827e-05,
+      "loss": 0.199,
+      "num_input_tokens_seen": 72672,
+      "step": 20,
+      "train_runtime": 33.9024,
+      "train_tokens_per_second": 2143.565
+    },
+    {
+      "epoch": 0.06458557588805167,
+      "grad_norm": 5.31423429492861e-05,
+      "learning_rate": 2.9968021784983436e-05,
+      "loss": 0.1809,
+      "num_input_tokens_seen": 109472,
+      "step": 30,
+      "train_runtime": 47.2118,
+      "train_tokens_per_second": 2318.741
+    },
+    {
+      "epoch": 0.0861141011840689,
+      "grad_norm": 5.581648292718455e-05,
+      "learning_rate": 2.994218206053077e-05,
+      "loss": 0.1766,
+      "num_input_tokens_seen": 146656,
+      "step": 40,
+      "train_runtime": 60.0165,
+      "train_tokens_per_second": 2443.597
+    },
+    {
+      "epoch": 0.10764262648008611,
+      "grad_norm": 4.6867833589203656e-05,
+      "learning_rate": 2.9908764468726846e-05,
+      "loss": 0.1784,
+      "num_input_tokens_seen": 183200,
+      "step": 50,
+      "train_runtime": 72.7117,
+      "train_tokens_per_second": 2519.54
+    },
+    {
+      "epoch": 0.12917115177610333,
+      "grad_norm": 7.248102338053286e-05,
+      "learning_rate": 2.986778595716853e-05,
+      "loss": 0.165,
+      "num_input_tokens_seen": 220352,
+      "step": 60,
+      "train_runtime": 85.5372,
+      "train_tokens_per_second": 2576.096
+    },
+    {
+      "epoch": 0.15069967707212056,
+      "grad_norm": 6.31138464086689e-05,
+      "learning_rate": 2.981926730794264e-05,
+      "loss": 0.1884,
+      "num_input_tokens_seen": 256864,
+      "step": 70,
+      "train_runtime": 98.2829,
+      "train_tokens_per_second": 2613.516
+    },
+    {
+      "epoch": 0.1722282023681378,
+      "grad_norm": 8.531038474757224e-05,
+      "learning_rate": 2.9763233127086394e-05,
+      "loss": 0.1668,
+      "num_input_tokens_seen": 293408,
+      "step": 80,
+      "train_runtime": 110.9453,
+      "train_tokens_per_second": 2644.619
+    },
+    {
+      "epoch": 0.193756727664155,
+      "grad_norm": 3.399947308935225e-05,
+      "learning_rate": 2.9699711832108555e-05,
+      "loss": 0.1538,
+      "num_input_tokens_seen": 329152,
+      "step": 90,
+      "train_runtime": 123.3823,
+      "train_tokens_per_second": 2667.742
+    },
+    {
+      "epoch": 0.21528525296017223,
+      "grad_norm": 6.389995542122051e-05,
+      "learning_rate": 2.9628735637577612e-05,
+      "loss": 0.1729,
+      "num_input_tokens_seen": 365632,
+      "step": 100,
+      "train_runtime": 136.0062,
+      "train_tokens_per_second": 2688.348
+    },
+    {
+      "epoch": 0.23681377825618946,
+      "grad_norm": 8.456028444925323e-05,
+      "learning_rate": 2.955034053878428e-05,
+      "loss": 0.1346,
+      "num_input_tokens_seen": 401888,
+      "step": 110,
+      "train_runtime": 148.598,
+      "train_tokens_per_second": 2704.531
+    },
+    {
+      "epoch": 0.25834230355220666,
+      "grad_norm": 6.639358616666868e-05,
+      "learning_rate": 2.9464566293486653e-05,
+      "loss": 0.1622,
+      "num_input_tokens_seen": 438816,
+      "step": 120,
+      "train_runtime": 161.3801,
+      "train_tokens_per_second": 2719.146
+    },
+    {
+      "epoch": 0.2798708288482239,
+      "grad_norm": 6.005055911373347e-05,
+      "learning_rate": 2.9371456401747193e-05,
+      "loss": 0.1357,
+      "num_input_tokens_seen": 475648,
+      "step": 130,
+      "train_runtime": 174.163,
+      "train_tokens_per_second": 2731.051
+    },
+    {
+      "epoch": 0.3013993541442411,
+      "grad_norm": 8.286784577649087e-05,
+      "learning_rate": 2.9271058083871864e-05,
+      "loss": 0.1537,
+      "num_input_tokens_seen": 512256,
+      "step": 140,
+      "train_runtime": 186.8563,
+      "train_tokens_per_second": 2741.444
+    },
+    {
+      "epoch": 0.32292787944025836,
+      "grad_norm": 4.636685480363667e-05,
+      "learning_rate": 2.916342225646254e-05,
+      "loss": 0.1183,
+      "num_input_tokens_seen": 548992,
+      "step": 150,
+      "train_runtime": 199.6287,
+      "train_tokens_per_second": 2750.065
+    },
+    {
+      "epoch": 0.3444564047362756,
+      "grad_norm": 4.919446655549109e-05,
+      "learning_rate": 2.904860350659487e-05,
+      "loss": 0.1336,
+      "num_input_tokens_seen": 585792,
+      "step": 160,
+      "train_runtime": 212.4096,
+      "train_tokens_per_second": 2757.841
+    },
+    {
+      "epoch": 0.36598493003229277,
+      "grad_norm": 8.222900942200795e-05,
+      "learning_rate": 2.8926660064134637e-05,
+      "loss": 0.1661,
+      "num_input_tokens_seen": 621952,
+      "step": 170,
+      "train_runtime": 225.0588,
+      "train_tokens_per_second": 2763.509
+    },
+    {
+      "epoch": 0.38751345532831,
+      "grad_norm": 3.515497155603953e-05,
+      "learning_rate": 2.8797653772206726e-05,
+      "loss": 0.1405,
+      "num_input_tokens_seen": 658240,
+      "step": 180,
+      "train_runtime": 237.6697,
+      "train_tokens_per_second": 2769.557
+    },
+    {
+      "epoch": 0.40904198062432723,
+      "grad_norm": 4.577370418701321e-05,
+      "learning_rate": 2.8661650055831645e-05,
+      "loss": 0.1456,
+      "num_input_tokens_seen": 694880,
+      "step": 190,
+      "train_runtime": 250.3071,
+      "train_tokens_per_second": 2776.11
+    },
+    {
+      "epoch": 0.43057050592034446,
+      "grad_norm": 5.2699797379318625e-05,
+      "learning_rate": 2.8518717888745486e-05,
+      "loss": 0.1158,
+      "num_input_tokens_seen": 731968,
+      "step": 200,
+      "train_runtime": 263.1248,
+      "train_tokens_per_second": 2781.828
+    },
+    {
+      "epoch": 0.4520990312163617,
+      "grad_norm": 4.86937933601439e-05,
+      "learning_rate": 2.836892975842015e-05,
+      "loss": 0.1293,
+      "num_input_tokens_seen": 768416,
+      "step": 210,
+      "train_runtime": 275.7314,
+      "train_tokens_per_second": 2786.828
+    },
+    {
+      "epoch": 0.4736275565123789,
+      "grad_norm": 5.121620051795617e-05,
+      "learning_rate": 2.8212361629301655e-05,
+      "loss": 0.1456,
+      "num_input_tokens_seen": 805376,
+      "step": 220,
+      "train_runtime": 288.4976,
+      "train_tokens_per_second": 2791.621
+    },
+    {
+      "epoch": 0.4951560818083961,
+      "grad_norm": 3.388872210052796e-05,
+      "learning_rate": 2.8049092904285033e-05,
+      "loss": 0.1428,
+      "num_input_tokens_seen": 841472,
+      "step": 230,
+      "train_runtime": 301.0113,
+      "train_tokens_per_second": 2795.483
+    },
+    {
+      "epoch": 0.5166846071044133,
+      "grad_norm": 5.026933649787679e-05,
+      "learning_rate": 2.7879206384445526e-05,
+      "loss": 0.1334,
+      "num_input_tokens_seen": 878464,
+      "step": 240,
+      "train_runtime": 313.7308,
+      "train_tokens_per_second": 2800.057
+    },
+    {
+      "epoch": 0.5382131324004306,
+      "grad_norm": 2.826541276590433e-05,
+      "learning_rate": 2.7702788227046286e-05,
+      "loss": 0.1414,
+      "num_input_tokens_seen": 914432,
+      "step": 250,
+      "train_runtime": 326.1923,
+      "train_tokens_per_second": 2803.352
+    },
+    {
+      "epoch": 0.5597416576964478,
+      "grad_norm": 5.1211100071668625e-05,
+      "learning_rate": 2.751992790184413e-05,
+      "loss": 0.1098,
+      "num_input_tokens_seen": 952320,
+      "step": 260,
+      "train_runtime": 339.127,
+      "train_tokens_per_second": 2808.151
+    },
+    {
+      "epoch": 0.581270182992465,
+      "grad_norm": 4.860994886257686e-05,
+      "learning_rate": 2.733071814571527e-05,
+      "loss": 0.1643,
+      "num_input_tokens_seen": 988800,
+      "step": 270,
+      "train_runtime": 351.788,
+      "train_tokens_per_second": 2810.784
+    },
+    {
+      "epoch": 0.6027987082884823,
+      "grad_norm": 3.192482836311683e-05,
+      "learning_rate": 2.7135254915624213e-05,
+      "loss": 0.1222,
+      "num_input_tokens_seen": 1025952,
+      "step": 280,
+      "train_runtime": 364.6184,
+      "train_tokens_per_second": 2813.769
+    },
+    {
+      "epoch": 0.6243272335844995,
+      "grad_norm": 6.410479545593262e-05,
+      "learning_rate": 2.693363733995954e-05,
+      "loss": 0.1489,
+      "num_input_tokens_seen": 1063360,
+      "step": 290,
+      "train_runtime": 377.4609,
+      "train_tokens_per_second": 2817.139
+    },
+    {
+      "epoch": 0.6458557588805167,
+      "grad_norm": 7.877861935412511e-05,
+      "learning_rate": 2.672596766826138e-05,
+      "loss": 0.1207,
+      "num_input_tokens_seen": 1099648,
+      "step": 300,
+      "train_runtime": 390.138,
+      "train_tokens_per_second": 2818.613
+    },
+    {
+      "epoch": 0.667384284176534,
+      "grad_norm": 5.999449058435857e-05,
+      "learning_rate": 2.6512351219365927e-05,
+      "loss": 0.1319,
+      "num_input_tokens_seen": 1135712,
+      "step": 310,
+      "train_runtime": 402.7639,
+      "train_tokens_per_second": 2819.796
+    },
+    {
+      "epoch": 0.6889128094725512,
+      "grad_norm": 2.934955591626931e-05,
+      "learning_rate": 2.6292896327993467e-05,
+      "loss": 0.1453,
+      "num_input_tokens_seen": 1172384,
+      "step": 320,
+      "train_runtime": 415.4411,
+      "train_tokens_per_second": 2822.022
+    },
+    {
+      "epoch": 0.7104413347685683,
+      "grad_norm": 4.606367292581126e-05,
+      "learning_rate": 2.606771428980681e-05,
+      "loss": 0.1202,
+      "num_input_tokens_seen": 1209440,
+      "step": 330,
+      "train_runtime": 428.2678,
+      "train_tokens_per_second": 2824.027
+    },
+    {
+      "epoch": 0.7319698600645855,
+      "grad_norm": 3.709704469656572e-05,
+      "learning_rate": 2.5836919304968184e-05,
+      "loss": 0.1283,
+      "num_input_tokens_seen": 1245920,
+      "step": 340,
+      "train_runtime": 440.939,
+      "train_tokens_per_second": 2825.606
+    },
+    {
+      "epoch": 0.7534983853606028,
+      "grad_norm": 4.9612910515861586e-05,
+      "learning_rate": 2.5600628420223057e-05,
+      "loss": 0.1374,
+      "num_input_tokens_seen": 1284288,
+      "step": 350,
+      "train_runtime": 454.0391,
+      "train_tokens_per_second": 2828.585
+    },
+    {
+      "epoch": 0.77502691065662,
+      "grad_norm": 3.349883991177194e-05,
+      "learning_rate": 2.535896146954037e-05,
+      "loss": 0.1355,
+      "num_input_tokens_seen": 1320896,
+      "step": 360,
+      "train_runtime": 466.7256,
+      "train_tokens_per_second": 2830.134
+    },
+    {
+      "epoch": 0.7965554359526372,
+      "grad_norm": 3.364904478075914e-05,
+      "learning_rate": 2.5112041013339212e-05,
+      "loss": 0.131,
+      "num_input_tokens_seen": 1359264,
+      "step": 370,
+      "train_runtime": 479.8532,
+      "train_tokens_per_second": 2832.667
+    },
+    {
+      "epoch": 0.8180839612486545,
+      "grad_norm": 3.689135701279156e-05,
+      "learning_rate": 2.48599922763328e-05,
+      "loss": 0.1034,
+      "num_input_tokens_seen": 1396032,
+      "step": 380,
+      "train_runtime": 492.5082,
+      "train_tokens_per_second": 2834.535
+    },
+    {
+      "epoch": 0.8396124865446717,
+      "grad_norm": 3.208614361938089e-05,
+      "learning_rate": 2.4602943084021286e-05,
+      "loss": 0.1465,
+      "num_input_tokens_seen": 1432224,
+      "step": 390,
+      "train_runtime": 505.1852,
+      "train_tokens_per_second": 2835.047
+    },
+    {
+      "epoch": 0.8611410118406889,
+      "grad_norm": 5.1129703933838755e-05,
+      "learning_rate": 2.434102379786553e-05,
+      "loss": 0.1049,
+      "num_input_tokens_seen": 1468224,
+      "step": 400,
+      "train_runtime": 517.7561,
+      "train_tokens_per_second": 2835.744
+    },
+    {
+      "epoch": 0.8826695371367062,
+      "grad_norm": 5.446456270874478e-05,
+      "learning_rate": 2.4074367249174864e-05,
+      "loss": 0.1024,
+      "num_input_tokens_seen": 1505088,
+      "step": 410,
+      "train_runtime": 530.4568,
+      "train_tokens_per_second": 2837.343
+    },
+    {
+      "epoch": 0.9041980624327234,
+      "grad_norm": 8.551870996598154e-05,
+      "learning_rate": 2.380310867174219e-05,
+      "loss": 0.1155,
+      "num_input_tokens_seen": 1542048,
+      "step": 420,
+      "train_runtime": 543.2418,
+      "train_tokens_per_second": 2838.603
+    },
+    {
+      "epoch": 0.9257265877287406,
+      "grad_norm": 2.7210369808017276e-05,
+      "learning_rate": 2.3527385633260724e-05,
+      "loss": 0.117,
+      "num_input_tokens_seen": 1578496,
+      "step": 430,
+      "train_runtime": 555.7658,
+      "train_tokens_per_second": 2840.218
+    },
+    {
+      "epoch": 0.9472551130247578,
+      "grad_norm": 3.511784234433435e-05,
+      "learning_rate": 2.3247337965557097e-05,
+      "loss": 0.1023,
+      "num_input_tokens_seen": 1613440,
+      "step": 440,
+      "train_runtime": 568.0483,
+      "train_tokens_per_second": 2840.322
+    },
+    {
+      "epoch": 0.9687836383207751,
+      "grad_norm": 2.4012464564293623e-05,
+      "learning_rate": 2.2963107693676192e-05,
+      "loss": 0.1046,
+      "num_input_tokens_seen": 1649760,
+      "step": 450,
+      "train_runtime": 580.7665,
+      "train_tokens_per_second": 2840.66
+    },
+    {
+      "epoch": 0.9903121636167922,
+      "grad_norm": 4.6500768803525716e-05,
+      "learning_rate": 2.2674838963853738e-05,
+      "loss": 0.1106,
+      "num_input_tokens_seen": 1686016,
+      "step": 460,
+      "train_runtime": 593.3342,
+      "train_tokens_per_second": 2841.596
+    },
+    {
+      "epoch": 1.0107642626480087,
+      "grad_norm": 0.00010867212404264137,
+      "learning_rate": 2.2382677970413097e-05,
+      "loss": 0.0859,
+      "num_input_tokens_seen": 1721152,
+      "step": 470,
+      "train_runtime": 605.4653,
+      "train_tokens_per_second": 2842.693
+    },
+    {
+      "epoch": 1.0322927879440258,
+      "grad_norm": 3.351278428453952e-05,
+      "learning_rate": 2.2086772881623455e-05,
+      "loss": 0.1089,
+      "num_input_tokens_seen": 1757568,
+      "step": 480,
+      "train_runtime": 618.0558,
+      "train_tokens_per_second": 2843.705
+    },
+    {
+      "epoch": 1.0538213132400431,
+      "grad_norm": 4.3039726733695716e-05,
+      "learning_rate": 2.1787273764556827e-05,
+      "loss": 0.0852,
+      "num_input_tokens_seen": 1794848,
+      "step": 490,
+      "train_runtime": 630.8577,
+      "train_tokens_per_second": 2845.092
+    },
+    {
+      "epoch": 1.0753498385360603,
+      "grad_norm": 4.8467773012816906e-05,
+      "learning_rate": 2.1484332508982167e-05,
+      "loss": 0.0988,
+      "num_input_tokens_seen": 1832256,
+      "step": 500,
+      "train_runtime": 643.6632,
+      "train_tokens_per_second": 2846.607
+    },
+    {
+      "epoch": 1.0968783638320776,
+      "grad_norm": 3.706678398884833e-05,
+      "learning_rate": 2.11781027503351e-05,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 1868928,
+      "step": 510,
+      "train_runtime": 656.3448,
+      "train_tokens_per_second": 2847.479
+    },
+    {
+      "epoch": 1.1184068891280947,
+      "grad_norm": 4.0160564822144806e-05,
+      "learning_rate": 2.0868739791802312e-05,
+      "loss": 0.1107,
+      "num_input_tokens_seen": 1905824,
+      "step": 520,
+      "train_runtime": 669.0663,
+      "train_tokens_per_second": 2848.483
+    },
+    {
+      "epoch": 1.1399354144241118,
+      "grad_norm": 7.405655196635053e-05,
+      "learning_rate": 2.0556400525560168e-05,
+      "loss": 0.0894,
+      "num_input_tokens_seen": 1941760,
+      "step": 530,
+      "train_runtime": 681.5699,
+      "train_tokens_per_second": 2848.952
+    },
+    {
+      "epoch": 1.1614639397201292,
+      "grad_norm": 4.077348785358481e-05,
+      "learning_rate": 2.0241243353207484e-05,
+      "loss": 0.1016,
+      "num_input_tokens_seen": 1978208,
+      "step": 540,
+      "train_runtime": 694.297,
+      "train_tokens_per_second": 2849.224
+    },
+    {
+      "epoch": 1.1829924650161463,
+      "grad_norm": 4.100809746887535e-05,
+      "learning_rate": 1.9923428105432796e-05,
+      "loss": 0.0932,
+      "num_input_tokens_seen": 2015904,
+      "step": 550,
+      "train_runtime": 707.2242,
+      "train_tokens_per_second": 2850.445
+    },
+    {
+      "epoch": 1.2045209903121636,
+      "grad_norm": 8.349807467311621e-05,
+      "learning_rate": 1.9603115960956852e-05,
+      "loss": 0.1056,
+      "num_input_tokens_seen": 2052672,
+      "step": 560,
+      "train_runtime": 719.9875,
+      "train_tokens_per_second": 2850.983
+    },
+    {
+      "epoch": 1.2260495156081808,
+      "grad_norm": 2.9793711291858926e-05,
+      "learning_rate": 1.9280469364791505e-05,
+      "loss": 0.1033,
+      "num_input_tokens_seen": 2088736,
+      "step": 570,
+      "train_runtime": 732.5942,
+      "train_tokens_per_second": 2851.15
+    },
+    {
+      "epoch": 1.247578040904198,
+      "grad_norm": 6.373477663146332e-05,
+      "learning_rate": 1.8955651945856354e-05,
+      "loss": 0.1035,
+      "num_input_tokens_seen": 2125600,
+      "step": 580,
+      "train_runtime": 745.4583,
+      "train_tokens_per_second": 2851.4
+    },
+    {
+      "epoch": 1.2691065662002152,
+      "grad_norm": 3.1842911994317546e-05,
+      "learning_rate": 1.8628828433995013e-05,
+      "loss": 0.0811,
+      "num_input_tokens_seen": 2161792,
+      "step": 590,
+      "train_runtime": 758.1079,
+      "train_tokens_per_second": 2851.562
+    },
+    {
+      "epoch": 1.2906350914962326,
+      "grad_norm": 4.18422459915746e-05,
+      "learning_rate": 1.8300164576433035e-05,
+      "loss": 0.0888,
+      "num_input_tokens_seen": 2197920,
+      "step": 600,
+      "train_runtime": 770.7916,
+      "train_tokens_per_second": 2851.51
+    },
+    {
+      "epoch": 1.3121636167922497,
+      "grad_norm": 4.866390736424364e-05,
+      "learning_rate": 1.7969827053719862e-05,
+      "loss": 0.0968,
+      "num_input_tokens_seen": 2235168,
+      "step": 610,
+      "train_runtime": 783.669,
+      "train_tokens_per_second": 2852.184
+    },
+    {
+      "epoch": 1.333692142088267,
+      "grad_norm": 4.293969686841592e-05,
+      "learning_rate": 1.7637983395197465e-05,
+      "loss": 0.1311,
+      "num_input_tokens_seen": 2270816,
+      "step": 620,
+      "train_runtime": 796.2193,
+      "train_tokens_per_second": 2851.998
+    },
+    {
+      "epoch": 1.3552206673842842,
+      "grad_norm": 6.0659429436782375e-05,
+      "learning_rate": 1.7304801894038522e-05,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 2308128,
+      "step": 630,
+      "train_runtime": 809.1462,
+      "train_tokens_per_second": 2852.547
+    },
+    {
+      "epoch": 1.3767491926803013,
+      "grad_norm": 3.0738956411369145e-05,
+      "learning_rate": 1.6970451521897198e-05,
+      "loss": 0.0887,
+      "num_input_tokens_seen": 2344992,
+      "step": 640,
+      "train_runtime": 822.0044,
+      "train_tokens_per_second": 2852.773
+    },
+    {
+      "epoch": 1.3982777179763186,
+      "grad_norm": 8.285656076623127e-05,
+      "learning_rate": 1.663510184321587e-05,
+      "loss": 0.1041,
+      "num_input_tokens_seen": 2381056,
+      "step": 650,
+      "train_runtime": 834.6005,
+      "train_tokens_per_second": 2852.929
+    },
+    {
+      "epoch": 1.419806243272336,
+      "grad_norm": 5.0590566388564184e-05,
+      "learning_rate": 1.6298922929231194e-05,
+      "loss": 0.1022,
+      "num_input_tokens_seen": 2417152,
+      "step": 660,
+      "train_runtime": 847.302,
+      "train_tokens_per_second": 2852.763
+    },
+    {
+      "epoch": 1.441334768568353,
+      "grad_norm": 7.397153240162879e-05,
+      "learning_rate": 1.5962085271723158e-05,
+      "loss": 0.0831,
+      "num_input_tokens_seen": 2454048,
+      "step": 670,
+      "train_runtime": 860.0949,
+      "train_tokens_per_second": 2853.229
+    },
+    {
+      "epoch": 1.4628632938643702,
+      "grad_norm": 4.698023258242756e-05,
+      "learning_rate": 1.5624759696550874e-05,
+      "loss": 0.1055,
+      "num_input_tokens_seen": 2491552,
+      "step": 680,
+      "train_runtime": 873.1207,
+      "train_tokens_per_second": 2853.617
+    },
+    {
+      "epoch": 1.4843918191603875,
+      "grad_norm": 4.3750838813139126e-05,
+      "learning_rate": 1.528711727701893e-05,
+      "loss": 0.0812,
+      "num_input_tokens_seen": 2528128,
+      "step": 690,
+      "train_runtime": 885.9021,
+      "train_tokens_per_second": 2853.733
+    },
+    {
+      "epoch": 1.5059203444564049,
+      "grad_norm": 6.37572884443216e-05,
+      "learning_rate": 1.4949329247118258e-05,
+      "loss": 0.1016,
+      "num_input_tokens_seen": 2564448,
+      "step": 700,
+      "train_runtime": 898.5373,
+      "train_tokens_per_second": 2854.025
+    },
+    {
+      "epoch": 1.527448869752422,
+      "grad_norm": 4.437049574335106e-05,
+      "learning_rate": 1.4611566914685489e-05,
+      "loss": 0.0956,
+      "num_input_tokens_seen": 2601056,
+      "step": 710,
+      "train_runtime": 911.3255,
+      "train_tokens_per_second": 2854.146
+    },
+    {
+      "epoch": 1.5489773950484391,
+      "grad_norm": 5.1571758376667276e-05,
+      "learning_rate": 1.4274001574524894e-05,
+      "loss": 0.0936,
+      "num_input_tokens_seen": 2637888,
+      "step": 720,
+      "train_runtime": 924.1173,
+      "train_tokens_per_second": 2854.495
+    },
+    {
+      "epoch": 1.5705059203444565,
+      "grad_norm": 6.383092113537714e-05,
+      "learning_rate": 1.3936804421536906e-05,
+      "loss": 0.0728,
+      "num_input_tokens_seen": 2673312,
+      "step": 730,
+      "train_runtime": 937.0636,
+      "train_tokens_per_second": 2852.861
+    },
+    {
+      "epoch": 1.5920344456404736,
+      "grad_norm": 2.989832137245685e-05,
+      "learning_rate": 1.3600146463897327e-05,
+      "loss": 0.0839,
+      "num_input_tokens_seen": 2709696,
+      "step": 740,
+      "train_runtime": 949.7913,
+      "train_tokens_per_second": 2852.938
+    },
+    {
+      "epoch": 1.6135629709364907,
+      "grad_norm": 0.00010528949496801943,
+      "learning_rate": 1.3264198436331193e-05,
+      "loss": 0.1308,
+      "num_input_tokens_seen": 2746528,
+      "step": 750,
+      "train_runtime": 962.4815,
+      "train_tokens_per_second": 2853.591
+    },
+    {
+      "epoch": 1.635091496232508,
+      "grad_norm": 3.306402868474834e-05,
+      "learning_rate": 1.2929130713525396e-05,
+      "loss": 0.0927,
+      "num_input_tokens_seen": 2782656,
+      "step": 760,
+      "train_runtime": 975.0888,
+      "train_tokens_per_second": 2853.746
+    },
+    {
+      "epoch": 1.6566200215285254,
+      "grad_norm": 3.532276605255902e-05,
+      "learning_rate": 1.2595113223723788e-05,
+      "loss": 0.1135,
+      "num_input_tokens_seen": 2819744,
+      "step": 770,
+      "train_runtime": 987.9762,
+      "train_tokens_per_second": 2854.061
+    },
+    {
+      "epoch": 1.6781485468245425,
+      "grad_norm": 6.426728941733018e-05,
+      "learning_rate": 1.2262315362548825e-05,
+      "loss": 0.0813,
+      "num_input_tokens_seen": 2857216,
+      "step": 780,
+      "train_runtime": 1001.0086,
+      "train_tokens_per_second": 2854.337
+    },
+    {
+      "epoch": 1.6996770721205596,
+      "grad_norm": 8.145067113218829e-05,
+      "learning_rate": 1.1930905907093196e-05,
+      "loss": 0.0933,
+      "num_input_tokens_seen": 2892000,
+      "step": 790,
+      "train_runtime": 1013.4209,
+      "train_tokens_per_second": 2853.701
+    },
+    {
+      "epoch": 1.721205597416577,
+      "grad_norm": 0.00011963866563746706,
+      "learning_rate": 1.1601052930325236e-05,
+      "loss": 0.0881,
+      "num_input_tokens_seen": 2928832,
+      "step": 800,
+      "train_runtime": 1026.3231,
+      "train_tokens_per_second": 2853.714
+    },
+    {
+      "epoch": 1.7427341227125943,
+      "grad_norm": 5.128612610860728e-05,
+      "learning_rate": 1.127292371585141e-05,
+      "loss": 0.079,
+      "num_input_tokens_seen": 2965632,
+      "step": 810,
+      "train_runtime": 1039.1265,
+      "train_tokens_per_second": 2853.966
+    },
+    {
+      "epoch": 1.7642626480086114,
+      "grad_norm": 7.751452358206734e-05,
+      "learning_rate": 1.0946684673079074e-05,
+      "loss": 0.1132,
+      "num_input_tokens_seen": 3002496,
+      "step": 820,
+      "train_runtime": 1051.9839,
+      "train_tokens_per_second": 2854.127
+    },
+    {
+      "epoch": 1.7857911733046286,
+      "grad_norm": 4.0213129977928475e-05,
+      "learning_rate": 1.0622501252822628e-05,
+      "loss": 0.0793,
+      "num_input_tokens_seen": 3038944,
+      "step": 830,
+      "train_runtime": 1064.6802,
+      "train_tokens_per_second": 2854.326
+    },
+    {
+      "epoch": 1.807319698600646,
+      "grad_norm": 4.586371505865827e-05,
+      "learning_rate": 1.0300537863395823e-05,
+      "loss": 0.0918,
+      "num_input_tokens_seen": 3076032,
+      "step": 840,
+      "train_runtime": 1077.5859,
+      "train_tokens_per_second": 2854.559
+    },
+    {
+      "epoch": 1.8288482238966632,
+      "grad_norm": 3.2480071240570396e-05,
+      "learning_rate": 9.98095778723271e-06,
+      "loss": 0.093,
+      "num_input_tokens_seen": 3112448,
+      "step": 850,
+      "train_runtime": 1090.3056,
+      "train_tokens_per_second": 2854.656
+    },
+    {
+      "epoch": 1.8503767491926801,
+      "grad_norm": 5.541834980249405e-05,
+      "learning_rate": 9.663923098079656e-06,
+      "loss": 0.0973,
+      "num_input_tokens_seen": 3149376,
+      "step": 860,
+      "train_runtime": 1103.1339,
+      "train_tokens_per_second": 2854.935
+    },
+    {
+      "epoch": 1.8719052744886975,
+      "grad_norm": 3.055919296457432e-05,
+      "learning_rate": 9.349594578800259e-06,
+      "loss": 0.0995,
+      "num_input_tokens_seen": 3186304,
+      "step": 870,
+      "train_runtime": 1116.0618,
+      "train_tokens_per_second": 2854.953
+    },
+    {
+      "epoch": 1.8934337997847148,
+      "grad_norm": 4.6285113057820126e-05,
+      "learning_rate": 9.038131639834977e-06,
+      "loss": 0.1238,
+      "num_input_tokens_seen": 3222752,
+      "step": 880,
+      "train_runtime": 1128.7812,
+      "train_tokens_per_second": 2855.072
+    },
+    {
+      "epoch": 1.914962325080732,
+      "grad_norm": 4.483849261305295e-05,
+      "learning_rate": 8.7296922383568e-06,
+      "loss": 0.0912,
+      "num_input_tokens_seen": 3258976,
+      "step": 890,
+      "train_runtime": 1141.5159,
+      "train_tokens_per_second": 2854.954
+    },
+    {
+      "epoch": 1.936490850376749,
+      "grad_norm": 5.5885881010908633e-05,
+      "learning_rate": 8.424432798163838e-06,
+      "loss": 0.0935,
+      "num_input_tokens_seen": 3295040,
+      "step": 900,
+      "train_runtime": 1154.1132,
+      "train_tokens_per_second": 2855.041
+    },
+    {
+      "epoch": 1.9580193756727664,
+      "grad_norm": 3.8836213207105175e-05,
+      "learning_rate": 8.122508130349711e-06,
+      "loss": 0.0722,
+      "num_input_tokens_seen": 3331392,
+      "step": 910,
+      "train_runtime": 1166.7365,
+      "train_tokens_per_second": 2855.308
+    },
+    {
+      "epoch": 1.9795479009687837,
+      "grad_norm": 6.445301551138982e-05,
+      "learning_rate": 7.82407135479164e-06,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 3368032,
+      "step": 920,
+      "train_runtime": 1179.5494,
+      "train_tokens_per_second": 2855.355
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.651374805078376e-05,
+      "learning_rate": 7.529273822496379e-06,
+      "loss": 0.0647,
+      "num_input_tokens_seen": 3403584,
+      "step": 930,
+      "train_runtime": 1191.8451,
+      "train_tokens_per_second": 2855.727
+    },
+    {
+      "epoch": 2.0215285252960173,
+      "grad_norm": 9.509055962553248e-05,
+      "learning_rate": 7.238265038843143e-06,
+      "loss": 0.0881,
+      "num_input_tokens_seen": 3440000,
+      "step": 940,
+      "train_runtime": 1204.5181,
+      "train_tokens_per_second": 2855.914
+    },
+    {
+      "epoch": 2.0430570505920342,
+      "grad_norm": 5.2039082220289856e-05,
+      "learning_rate": 6.951192587762555e-06,
+      "loss": 0.0573,
+      "num_input_tokens_seen": 3476640,
+      "step": 950,
+      "train_runtime": 1217.2855,
+      "train_tokens_per_second": 2856.06
+    },
+    {
+      "epoch": 2.0645855758880516,
+      "grad_norm": 3.832239599432796e-05,
+      "learning_rate": 6.668202056890052e-06,
+      "loss": 0.0646,
+      "num_input_tokens_seen": 3512992,
+      "step": 960,
+      "train_runtime": 1229.9665,
+      "train_tokens_per_second": 2856.169
+    },
+    {
+      "epoch": 2.086114101184069,
+      "grad_norm": 5.945537122897804e-05,
+      "learning_rate": 6.389436963731707e-06,
+      "loss": 0.0682,
+      "num_input_tokens_seen": 3549504,
+      "step": 970,
+      "train_runtime": 1242.6763,
+      "train_tokens_per_second": 2856.338
+    },
+    {
+      "epoch": 2.1076426264800863,
+      "grad_norm": 6.592880527023226e-05,
+      "learning_rate": 6.115038682879904e-06,
+      "loss": 0.0931,
+      "num_input_tokens_seen": 3586592,
+      "step": 980,
+      "train_runtime": 1255.5495,
+      "train_tokens_per_second": 2856.591
+    },
+    {
+      "epoch": 2.129171151776103,
+      "grad_norm": 2.9424012609524652e-05,
+      "learning_rate": 5.845146374315781e-06,
+      "loss": 0.0686,
+      "num_input_tokens_seen": 3622944,
+      "step": 990,
+      "train_runtime": 1268.2309,
+      "train_tokens_per_second": 2856.691
+    },
+    {
+      "epoch": 2.1506996770721205,
+      "grad_norm": 5.797382618766278e-05,
+      "learning_rate": 5.579896912834807e-06,
+      "loss": 0.0917,
+      "num_input_tokens_seen": 3660096,
+      "step": 1000,
+      "train_runtime": 1281.1218,
+      "train_tokens_per_second": 2856.946
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 1395,
+  "num_input_tokens_seen": 3660096,
+  "num_train_epochs": 3,
+  "save_steps": 1000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.5668382010599014e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-1000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0da07f323cd8cadc582879cf09f450ad877537a29fbf8dd1bdb4c30cac53e814
+size 6161

checkpoint-1395/README.md

@@ -0,0 +1,202 @@
+---
+base_model: unsloth/mistral-7b-instruct-v0.3-bnb-4bit
+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

checkpoint-1395/adapter_config.json

@@ -0,0 +1,43 @@
+{
+  "alpha_pattern": {},
+  "auto_mapping": {
+    "base_model_class": "MistralForCausalLM",
+    "parent_library": "transformers.models.mistral.modeling_mistral",
+    "unsloth_fixed": true
+  },
+  "base_model_name_or_path": "unsloth/mistral-7b-instruct-v0.3-bnb-4bit",
+  "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": 16,
+  "lora_bias": false,
+  "lora_dropout": 0,
+  "megatron_config": null,
+  "megatron_core": "megatron.core",
+  "modules_to_save": null,
+  "peft_type": "LORA",
+  "r": 8,
+  "rank_pattern": {},
+  "revision": null,
+  "target_modules": [
+    "down_proj",
+    "o_proj",
+    "gate_proj",
+    "q_proj",
+    "k_proj",
+    "up_proj",
+    "v_proj"
+  ],
+  "task_type": "CAUSAL_LM",
+  "trainable_token_indices": null,
+  "use_dora": false,
+  "use_rslora": false
+}

checkpoint-1395/adapter_model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a84ee66abe7f9a3129784babc9b46656fbe76623efe7131228f04bdcd0f3f973
+size 83945296

checkpoint-1395/chat_template.jinja

@@ -0,0 +1,9 @@
+{% set system_message = 'Below is an instruction that describes a task. Write a response that appropriately completes the request.
+
+' %}{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% endif %}{% if system_message is defined %}{{ system_message }}{% endif %}{% for message in loop_messages %}{% set content = message['content'] %}{% if message['role'] == 'user' %}{{ '### Instruction:
+' + content + '
+
+### Response:
+' }}{% elif message['role'] == 'assistant' %}{{ content + '</s>' + '
+
+' }}{% endif %}{% endfor %}

checkpoint-1395/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f50137bde7ca7f800f18eccc9c9f723dd7c4da2983251a0894120f9c7867873a
+size 168149539

checkpoint-1395/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2534e434cd5abbb8f7668d3eab0549db0ef95d6a797a3efa86b712e8e32266a7
+size 14645

checkpoint-1395/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3fb96e2e902c31267abe8b8c46628fd1600502041ca6301bd8a44b7861297fd
+size 1383

checkpoint-1395/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04444d631cd637645204283b0be3b130f0d32a6e952a616746c838579a7e024c
+size 1465

checkpoint-1395/special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

checkpoint-1395/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37f00374dea48658ee8f5d0f21895b9bc55cb0103939607c8185bfd1c6ca1f89
+size 587404

checkpoint-1395/trainer_state.json

@@ -0,0 +1,1424 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1395,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.021528525296017224,
+      "grad_norm": 0.000117536197649315,
+      "learning_rate": 2.9996919063517692e-05,
+      "loss": 0.8869,
+      "num_input_tokens_seen": 36480,
+      "step": 10,
+      "train_runtime": 21.2665,
+      "train_tokens_per_second": 1715.375
+    },
+    {
+      "epoch": 0.04305705059203445,
+      "grad_norm": 0.00022027654631529003,
+      "learning_rate": 2.9986270537572827e-05,
+      "loss": 0.199,
+      "num_input_tokens_seen": 72672,
+      "step": 20,
+      "train_runtime": 33.9024,
+      "train_tokens_per_second": 2143.565
+    },
+    {
+      "epoch": 0.06458557588805167,
+      "grad_norm": 5.31423429492861e-05,
+      "learning_rate": 2.9968021784983436e-05,
+      "loss": 0.1809,
+      "num_input_tokens_seen": 109472,
+      "step": 30,
+      "train_runtime": 47.2118,
+      "train_tokens_per_second": 2318.741
+    },
+    {
+      "epoch": 0.0861141011840689,
+      "grad_norm": 5.581648292718455e-05,
+      "learning_rate": 2.994218206053077e-05,
+      "loss": 0.1766,
+      "num_input_tokens_seen": 146656,
+      "step": 40,
+      "train_runtime": 60.0165,
+      "train_tokens_per_second": 2443.597
+    },
+    {
+      "epoch": 0.10764262648008611,
+      "grad_norm": 4.6867833589203656e-05,
+      "learning_rate": 2.9908764468726846e-05,
+      "loss": 0.1784,
+      "num_input_tokens_seen": 183200,
+      "step": 50,
+      "train_runtime": 72.7117,
+      "train_tokens_per_second": 2519.54
+    },
+    {
+      "epoch": 0.12917115177610333,
+      "grad_norm": 7.248102338053286e-05,
+      "learning_rate": 2.986778595716853e-05,
+      "loss": 0.165,
+      "num_input_tokens_seen": 220352,
+      "step": 60,
+      "train_runtime": 85.5372,
+      "train_tokens_per_second": 2576.096
+    },
+    {
+      "epoch": 0.15069967707212056,
+      "grad_norm": 6.31138464086689e-05,
+      "learning_rate": 2.981926730794264e-05,
+      "loss": 0.1884,
+      "num_input_tokens_seen": 256864,
+      "step": 70,
+      "train_runtime": 98.2829,
+      "train_tokens_per_second": 2613.516
+    },
+    {
+      "epoch": 0.1722282023681378,
+      "grad_norm": 8.531038474757224e-05,
+      "learning_rate": 2.9763233127086394e-05,
+      "loss": 0.1668,
+      "num_input_tokens_seen": 293408,
+      "step": 80,
+      "train_runtime": 110.9453,
+      "train_tokens_per_second": 2644.619
+    },
+    {
+      "epoch": 0.193756727664155,
+      "grad_norm": 3.399947308935225e-05,
+      "learning_rate": 2.9699711832108555e-05,
+      "loss": 0.1538,
+      "num_input_tokens_seen": 329152,
+      "step": 90,
+      "train_runtime": 123.3823,
+      "train_tokens_per_second": 2667.742
+    },
+    {
+      "epoch": 0.21528525296017223,
+      "grad_norm": 6.389995542122051e-05,
+      "learning_rate": 2.9628735637577612e-05,
+      "loss": 0.1729,
+      "num_input_tokens_seen": 365632,
+      "step": 100,
+      "train_runtime": 136.0062,
+      "train_tokens_per_second": 2688.348
+    },
+    {
+      "epoch": 0.23681377825618946,
+      "grad_norm": 8.456028444925323e-05,
+      "learning_rate": 2.955034053878428e-05,
+      "loss": 0.1346,
+      "num_input_tokens_seen": 401888,
+      "step": 110,
+      "train_runtime": 148.598,
+      "train_tokens_per_second": 2704.531
+    },
+    {
+      "epoch": 0.25834230355220666,
+      "grad_norm": 6.639358616666868e-05,
+      "learning_rate": 2.9464566293486653e-05,
+      "loss": 0.1622,
+      "num_input_tokens_seen": 438816,
+      "step": 120,
+      "train_runtime": 161.3801,
+      "train_tokens_per_second": 2719.146
+    },
+    {
+      "epoch": 0.2798708288482239,
+      "grad_norm": 6.005055911373347e-05,
+      "learning_rate": 2.9371456401747193e-05,
+      "loss": 0.1357,
+      "num_input_tokens_seen": 475648,
+      "step": 130,
+      "train_runtime": 174.163,
+      "train_tokens_per_second": 2731.051
+    },
+    {
+      "epoch": 0.3013993541442411,
+      "grad_norm": 8.286784577649087e-05,
+      "learning_rate": 2.9271058083871864e-05,
+      "loss": 0.1537,
+      "num_input_tokens_seen": 512256,
+      "step": 140,
+      "train_runtime": 186.8563,
+      "train_tokens_per_second": 2741.444
+    },
+    {
+      "epoch": 0.32292787944025836,
+      "grad_norm": 4.636685480363667e-05,
+      "learning_rate": 2.916342225646254e-05,
+      "loss": 0.1183,
+      "num_input_tokens_seen": 548992,
+      "step": 150,
+      "train_runtime": 199.6287,
+      "train_tokens_per_second": 2750.065
+    },
+    {
+      "epoch": 0.3444564047362756,
+      "grad_norm": 4.919446655549109e-05,
+      "learning_rate": 2.904860350659487e-05,
+      "loss": 0.1336,
+      "num_input_tokens_seen": 585792,
+      "step": 160,
+      "train_runtime": 212.4096,
+      "train_tokens_per_second": 2757.841
+    },
+    {
+      "epoch": 0.36598493003229277,
+      "grad_norm": 8.222900942200795e-05,
+      "learning_rate": 2.8926660064134637e-05,
+      "loss": 0.1661,
+      "num_input_tokens_seen": 621952,
+      "step": 170,
+      "train_runtime": 225.0588,
+      "train_tokens_per_second": 2763.509
+    },
+    {
+      "epoch": 0.38751345532831,
+      "grad_norm": 3.515497155603953e-05,
+      "learning_rate": 2.8797653772206726e-05,
+      "loss": 0.1405,
+      "num_input_tokens_seen": 658240,
+      "step": 180,
+      "train_runtime": 237.6697,
+      "train_tokens_per_second": 2769.557
+    },
+    {
+      "epoch": 0.40904198062432723,
+      "grad_norm": 4.577370418701321e-05,
+      "learning_rate": 2.8661650055831645e-05,
+      "loss": 0.1456,
+      "num_input_tokens_seen": 694880,
+      "step": 190,
+      "train_runtime": 250.3071,
+      "train_tokens_per_second": 2776.11
+    },
+    {
+      "epoch": 0.43057050592034446,
+      "grad_norm": 5.2699797379318625e-05,
+      "learning_rate": 2.8518717888745486e-05,
+      "loss": 0.1158,
+      "num_input_tokens_seen": 731968,
+      "step": 200,
+      "train_runtime": 263.1248,
+      "train_tokens_per_second": 2781.828
+    },
+    {
+      "epoch": 0.4520990312163617,
+      "grad_norm": 4.86937933601439e-05,
+      "learning_rate": 2.836892975842015e-05,
+      "loss": 0.1293,
+      "num_input_tokens_seen": 768416,
+      "step": 210,
+      "train_runtime": 275.7314,
+      "train_tokens_per_second": 2786.828
+    },
+    {
+      "epoch": 0.4736275565123789,
+      "grad_norm": 5.121620051795617e-05,
+      "learning_rate": 2.8212361629301655e-05,
+      "loss": 0.1456,
+      "num_input_tokens_seen": 805376,
+      "step": 220,
+      "train_runtime": 288.4976,
+      "train_tokens_per_second": 2791.621
+    },
+    {
+      "epoch": 0.4951560818083961,
+      "grad_norm": 3.388872210052796e-05,
+      "learning_rate": 2.8049092904285033e-05,
+      "loss": 0.1428,
+      "num_input_tokens_seen": 841472,
+      "step": 230,
+      "train_runtime": 301.0113,
+      "train_tokens_per_second": 2795.483
+    },
+    {
+      "epoch": 0.5166846071044133,
+      "grad_norm": 5.026933649787679e-05,
+      "learning_rate": 2.7879206384445526e-05,
+      "loss": 0.1334,
+      "num_input_tokens_seen": 878464,
+      "step": 240,
+      "train_runtime": 313.7308,
+      "train_tokens_per_second": 2800.057
+    },
+    {
+      "epoch": 0.5382131324004306,
+      "grad_norm": 2.826541276590433e-05,
+      "learning_rate": 2.7702788227046286e-05,
+      "loss": 0.1414,
+      "num_input_tokens_seen": 914432,
+      "step": 250,
+      "train_runtime": 326.1923,
+      "train_tokens_per_second": 2803.352
+    },
+    {
+      "epoch": 0.5597416576964478,
+      "grad_norm": 5.1211100071668625e-05,
+      "learning_rate": 2.751992790184413e-05,
+      "loss": 0.1098,
+      "num_input_tokens_seen": 952320,
+      "step": 260,
+      "train_runtime": 339.127,
+      "train_tokens_per_second": 2808.151
+    },
+    {
+      "epoch": 0.581270182992465,
+      "grad_norm": 4.860994886257686e-05,
+      "learning_rate": 2.733071814571527e-05,
+      "loss": 0.1643,
+      "num_input_tokens_seen": 988800,
+      "step": 270,
+      "train_runtime": 351.788,
+      "train_tokens_per_second": 2810.784
+    },
+    {
+      "epoch": 0.6027987082884823,
+      "grad_norm": 3.192482836311683e-05,
+      "learning_rate": 2.7135254915624213e-05,
+      "loss": 0.1222,
+      "num_input_tokens_seen": 1025952,
+      "step": 280,
+      "train_runtime": 364.6184,
+      "train_tokens_per_second": 2813.769
+    },
+    {
+      "epoch": 0.6243272335844995,
+      "grad_norm": 6.410479545593262e-05,
+      "learning_rate": 2.693363733995954e-05,
+      "loss": 0.1489,
+      "num_input_tokens_seen": 1063360,
+      "step": 290,
+      "train_runtime": 377.4609,
+      "train_tokens_per_second": 2817.139
+    },
+    {
+      "epoch": 0.6458557588805167,
+      "grad_norm": 7.877861935412511e-05,
+      "learning_rate": 2.672596766826138e-05,
+      "loss": 0.1207,
+      "num_input_tokens_seen": 1099648,
+      "step": 300,
+      "train_runtime": 390.138,
+      "train_tokens_per_second": 2818.613
+    },
+    {
+      "epoch": 0.667384284176534,
+      "grad_norm": 5.999449058435857e-05,
+      "learning_rate": 2.6512351219365927e-05,
+      "loss": 0.1319,
+      "num_input_tokens_seen": 1135712,
+      "step": 310,
+      "train_runtime": 402.7639,
+      "train_tokens_per_second": 2819.796
+    },
+    {
+      "epoch": 0.6889128094725512,
+      "grad_norm": 2.934955591626931e-05,
+      "learning_rate": 2.6292896327993467e-05,
+      "loss": 0.1453,
+      "num_input_tokens_seen": 1172384,
+      "step": 320,
+      "train_runtime": 415.4411,
+      "train_tokens_per_second": 2822.022
+    },
+    {
+      "epoch": 0.7104413347685683,
+      "grad_norm": 4.606367292581126e-05,
+      "learning_rate": 2.606771428980681e-05,
+      "loss": 0.1202,
+      "num_input_tokens_seen": 1209440,
+      "step": 330,
+      "train_runtime": 428.2678,
+      "train_tokens_per_second": 2824.027
+    },
+    {
+      "epoch": 0.7319698600645855,
+      "grad_norm": 3.709704469656572e-05,
+      "learning_rate": 2.5836919304968184e-05,
+      "loss": 0.1283,
+      "num_input_tokens_seen": 1245920,
+      "step": 340,
+      "train_runtime": 440.939,
+      "train_tokens_per_second": 2825.606
+    },
+    {
+      "epoch": 0.7534983853606028,
+      "grad_norm": 4.9612910515861586e-05,
+      "learning_rate": 2.5600628420223057e-05,
+      "loss": 0.1374,
+      "num_input_tokens_seen": 1284288,
+      "step": 350,
+      "train_runtime": 454.0391,
+      "train_tokens_per_second": 2828.585
+    },
+    {
+      "epoch": 0.77502691065662,
+      "grad_norm": 3.349883991177194e-05,
+      "learning_rate": 2.535896146954037e-05,
+      "loss": 0.1355,
+      "num_input_tokens_seen": 1320896,
+      "step": 360,
+      "train_runtime": 466.7256,
+      "train_tokens_per_second": 2830.134
+    },
+    {
+      "epoch": 0.7965554359526372,
+      "grad_norm": 3.364904478075914e-05,
+      "learning_rate": 2.5112041013339212e-05,
+      "loss": 0.131,
+      "num_input_tokens_seen": 1359264,
+      "step": 370,
+      "train_runtime": 479.8532,
+      "train_tokens_per_second": 2832.667
+    },
+    {
+      "epoch": 0.8180839612486545,
+      "grad_norm": 3.689135701279156e-05,
+      "learning_rate": 2.48599922763328e-05,
+      "loss": 0.1034,
+      "num_input_tokens_seen": 1396032,
+      "step": 380,
+      "train_runtime": 492.5082,
+      "train_tokens_per_second": 2834.535
+    },
+    {
+      "epoch": 0.8396124865446717,
+      "grad_norm": 3.208614361938089e-05,
+      "learning_rate": 2.4602943084021286e-05,
+      "loss": 0.1465,
+      "num_input_tokens_seen": 1432224,
+      "step": 390,
+      "train_runtime": 505.1852,
+      "train_tokens_per_second": 2835.047
+    },
+    {
+      "epoch": 0.8611410118406889,
+      "grad_norm": 5.1129703933838755e-05,
+      "learning_rate": 2.434102379786553e-05,
+      "loss": 0.1049,
+      "num_input_tokens_seen": 1468224,
+      "step": 400,
+      "train_runtime": 517.7561,
+      "train_tokens_per_second": 2835.744
+    },
+    {
+      "epoch": 0.8826695371367062,
+      "grad_norm": 5.446456270874478e-05,
+      "learning_rate": 2.4074367249174864e-05,
+      "loss": 0.1024,
+      "num_input_tokens_seen": 1505088,
+      "step": 410,
+      "train_runtime": 530.4568,
+      "train_tokens_per_second": 2837.343
+    },
+    {
+      "epoch": 0.9041980624327234,
+      "grad_norm": 8.551870996598154e-05,
+      "learning_rate": 2.380310867174219e-05,
+      "loss": 0.1155,
+      "num_input_tokens_seen": 1542048,
+      "step": 420,
+      "train_runtime": 543.2418,
+      "train_tokens_per_second": 2838.603
+    },
+    {
+      "epoch": 0.9257265877287406,
+      "grad_norm": 2.7210369808017276e-05,
+      "learning_rate": 2.3527385633260724e-05,
+      "loss": 0.117,
+      "num_input_tokens_seen": 1578496,
+      "step": 430,
+      "train_runtime": 555.7658,
+      "train_tokens_per_second": 2840.218
+    },
+    {
+      "epoch": 0.9472551130247578,
+      "grad_norm": 3.511784234433435e-05,
+      "learning_rate": 2.3247337965557097e-05,
+      "loss": 0.1023,
+      "num_input_tokens_seen": 1613440,
+      "step": 440,
+      "train_runtime": 568.0483,
+      "train_tokens_per_second": 2840.322
+    },
+    {
+      "epoch": 0.9687836383207751,
+      "grad_norm": 2.4012464564293623e-05,
+      "learning_rate": 2.2963107693676192e-05,
+      "loss": 0.1046,
+      "num_input_tokens_seen": 1649760,
+      "step": 450,
+      "train_runtime": 580.7665,
+      "train_tokens_per_second": 2840.66
+    },
+    {
+      "epoch": 0.9903121636167922,
+      "grad_norm": 4.6500768803525716e-05,
+      "learning_rate": 2.2674838963853738e-05,
+      "loss": 0.1106,
+      "num_input_tokens_seen": 1686016,
+      "step": 460,
+      "train_runtime": 593.3342,
+      "train_tokens_per_second": 2841.596
+    },
+    {
+      "epoch": 1.0107642626480087,
+      "grad_norm": 0.00010867212404264137,
+      "learning_rate": 2.2382677970413097e-05,
+      "loss": 0.0859,
+      "num_input_tokens_seen": 1721152,
+      "step": 470,
+      "train_runtime": 605.4653,
+      "train_tokens_per_second": 2842.693
+    },
+    {
+      "epoch": 1.0322927879440258,
+      "grad_norm": 3.351278428453952e-05,
+      "learning_rate": 2.2086772881623455e-05,
+      "loss": 0.1089,
+      "num_input_tokens_seen": 1757568,
+      "step": 480,
+      "train_runtime": 618.0558,
+      "train_tokens_per_second": 2843.705
+    },
+    {
+      "epoch": 1.0538213132400431,
+      "grad_norm": 4.3039726733695716e-05,
+      "learning_rate": 2.1787273764556827e-05,
+      "loss": 0.0852,
+      "num_input_tokens_seen": 1794848,
+      "step": 490,
+      "train_runtime": 630.8577,
+      "train_tokens_per_second": 2845.092
+    },
+    {
+      "epoch": 1.0753498385360603,
+      "grad_norm": 4.8467773012816906e-05,
+      "learning_rate": 2.1484332508982167e-05,
+      "loss": 0.0988,
+      "num_input_tokens_seen": 1832256,
+      "step": 500,
+      "train_runtime": 643.6632,
+      "train_tokens_per_second": 2846.607
+    },
+    {
+      "epoch": 1.0968783638320776,
+      "grad_norm": 3.706678398884833e-05,
+      "learning_rate": 2.11781027503351e-05,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 1868928,
+      "step": 510,
+      "train_runtime": 656.3448,
+      "train_tokens_per_second": 2847.479
+    },
+    {
+      "epoch": 1.1184068891280947,
+      "grad_norm": 4.0160564822144806e-05,
+      "learning_rate": 2.0868739791802312e-05,
+      "loss": 0.1107,
+      "num_input_tokens_seen": 1905824,
+      "step": 520,
+      "train_runtime": 669.0663,
+      "train_tokens_per_second": 2848.483
+    },
+    {
+      "epoch": 1.1399354144241118,
+      "grad_norm": 7.405655196635053e-05,
+      "learning_rate": 2.0556400525560168e-05,
+      "loss": 0.0894,
+      "num_input_tokens_seen": 1941760,
+      "step": 530,
+      "train_runtime": 681.5699,
+      "train_tokens_per_second": 2848.952
+    },
+    {
+      "epoch": 1.1614639397201292,
+      "grad_norm": 4.077348785358481e-05,
+      "learning_rate": 2.0241243353207484e-05,
+      "loss": 0.1016,
+      "num_input_tokens_seen": 1978208,
+      "step": 540,
+      "train_runtime": 694.297,
+      "train_tokens_per_second": 2849.224
+    },
+    {
+      "epoch": 1.1829924650161463,
+      "grad_norm": 4.100809746887535e-05,
+      "learning_rate": 1.9923428105432796e-05,
+      "loss": 0.0932,
+      "num_input_tokens_seen": 2015904,
+      "step": 550,
+      "train_runtime": 707.2242,
+      "train_tokens_per_second": 2850.445
+    },
+    {
+      "epoch": 1.2045209903121636,
+      "grad_norm": 8.349807467311621e-05,
+      "learning_rate": 1.9603115960956852e-05,
+      "loss": 0.1056,
+      "num_input_tokens_seen": 2052672,
+      "step": 560,
+      "train_runtime": 719.9875,
+      "train_tokens_per_second": 2850.983
+    },
+    {
+      "epoch": 1.2260495156081808,
+      "grad_norm": 2.9793711291858926e-05,
+      "learning_rate": 1.9280469364791505e-05,
+      "loss": 0.1033,
+      "num_input_tokens_seen": 2088736,
+      "step": 570,
+      "train_runtime": 732.5942,
+      "train_tokens_per_second": 2851.15
+    },
+    {
+      "epoch": 1.247578040904198,
+      "grad_norm": 6.373477663146332e-05,
+      "learning_rate": 1.8955651945856354e-05,
+      "loss": 0.1035,
+      "num_input_tokens_seen": 2125600,
+      "step": 580,
+      "train_runtime": 745.4583,
+      "train_tokens_per_second": 2851.4
+    },
+    {
+      "epoch": 1.2691065662002152,
+      "grad_norm": 3.1842911994317546e-05,
+      "learning_rate": 1.8628828433995013e-05,
+      "loss": 0.0811,
+      "num_input_tokens_seen": 2161792,
+      "step": 590,
+      "train_runtime": 758.1079,
+      "train_tokens_per_second": 2851.562
+    },
+    {
+      "epoch": 1.2906350914962326,
+      "grad_norm": 4.18422459915746e-05,
+      "learning_rate": 1.8300164576433035e-05,
+      "loss": 0.0888,
+      "num_input_tokens_seen": 2197920,
+      "step": 600,
+      "train_runtime": 770.7916,
+      "train_tokens_per_second": 2851.51
+    },
+    {
+      "epoch": 1.3121636167922497,
+      "grad_norm": 4.866390736424364e-05,
+      "learning_rate": 1.7969827053719862e-05,
+      "loss": 0.0968,
+      "num_input_tokens_seen": 2235168,
+      "step": 610,
+      "train_runtime": 783.669,
+      "train_tokens_per_second": 2852.184
+    },
+    {
+      "epoch": 1.333692142088267,
+      "grad_norm": 4.293969686841592e-05,
+      "learning_rate": 1.7637983395197465e-05,
+      "loss": 0.1311,
+      "num_input_tokens_seen": 2270816,
+      "step": 620,
+      "train_runtime": 796.2193,
+      "train_tokens_per_second": 2851.998
+    },
+    {
+      "epoch": 1.3552206673842842,
+      "grad_norm": 6.0659429436782375e-05,
+      "learning_rate": 1.7304801894038522e-05,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 2308128,
+      "step": 630,
+      "train_runtime": 809.1462,
+      "train_tokens_per_second": 2852.547
+    },
+    {
+      "epoch": 1.3767491926803013,
+      "grad_norm": 3.0738956411369145e-05,
+      "learning_rate": 1.6970451521897198e-05,
+      "loss": 0.0887,
+      "num_input_tokens_seen": 2344992,
+      "step": 640,
+      "train_runtime": 822.0044,
+      "train_tokens_per_second": 2852.773
+    },
+    {
+      "epoch": 1.3982777179763186,
+      "grad_norm": 8.285656076623127e-05,
+      "learning_rate": 1.663510184321587e-05,
+      "loss": 0.1041,
+      "num_input_tokens_seen": 2381056,
+      "step": 650,
+      "train_runtime": 834.6005,
+      "train_tokens_per_second": 2852.929
+    },
+    {
+      "epoch": 1.419806243272336,
+      "grad_norm": 5.0590566388564184e-05,
+      "learning_rate": 1.6298922929231194e-05,
+      "loss": 0.1022,
+      "num_input_tokens_seen": 2417152,
+      "step": 660,
+      "train_runtime": 847.302,
+      "train_tokens_per_second": 2852.763
+    },
+    {
+      "epoch": 1.441334768568353,
+      "grad_norm": 7.397153240162879e-05,
+      "learning_rate": 1.5962085271723158e-05,
+      "loss": 0.0831,
+      "num_input_tokens_seen": 2454048,
+      "step": 670,
+      "train_runtime": 860.0949,
+      "train_tokens_per_second": 2853.229
+    },
+    {
+      "epoch": 1.4628632938643702,
+      "grad_norm": 4.698023258242756e-05,
+      "learning_rate": 1.5624759696550874e-05,
+      "loss": 0.1055,
+      "num_input_tokens_seen": 2491552,
+      "step": 680,
+      "train_runtime": 873.1207,
+      "train_tokens_per_second": 2853.617
+    },
+    {
+      "epoch": 1.4843918191603875,
+      "grad_norm": 4.3750838813139126e-05,
+      "learning_rate": 1.528711727701893e-05,
+      "loss": 0.0812,
+      "num_input_tokens_seen": 2528128,
+      "step": 690,
+      "train_runtime": 885.9021,
+      "train_tokens_per_second": 2853.733
+    },
+    {
+      "epoch": 1.5059203444564049,
+      "grad_norm": 6.37572884443216e-05,
+      "learning_rate": 1.4949329247118258e-05,
+      "loss": 0.1016,
+      "num_input_tokens_seen": 2564448,
+      "step": 700,
+      "train_runtime": 898.5373,
+      "train_tokens_per_second": 2854.025
+    },
+    {
+      "epoch": 1.527448869752422,
+      "grad_norm": 4.437049574335106e-05,
+      "learning_rate": 1.4611566914685489e-05,
+      "loss": 0.0956,
+      "num_input_tokens_seen": 2601056,
+      "step": 710,
+      "train_runtime": 911.3255,
+      "train_tokens_per_second": 2854.146
+    },
+    {
+      "epoch": 1.5489773950484391,
+      "grad_norm": 5.1571758376667276e-05,
+      "learning_rate": 1.4274001574524894e-05,
+      "loss": 0.0936,
+      "num_input_tokens_seen": 2637888,
+      "step": 720,
+      "train_runtime": 924.1173,
+      "train_tokens_per_second": 2854.495
+    },
+    {
+      "epoch": 1.5705059203444565,
+      "grad_norm": 6.383092113537714e-05,
+      "learning_rate": 1.3936804421536906e-05,
+      "loss": 0.0728,
+      "num_input_tokens_seen": 2673312,
+      "step": 730,
+      "train_runtime": 937.0636,
+      "train_tokens_per_second": 2852.861
+    },
+    {
+      "epoch": 1.5920344456404736,
+      "grad_norm": 2.989832137245685e-05,
+      "learning_rate": 1.3600146463897327e-05,
+      "loss": 0.0839,
+      "num_input_tokens_seen": 2709696,
+      "step": 740,
+      "train_runtime": 949.7913,
+      "train_tokens_per_second": 2852.938
+    },
+    {
+      "epoch": 1.6135629709364907,
+      "grad_norm": 0.00010528949496801943,
+      "learning_rate": 1.3264198436331193e-05,
+      "loss": 0.1308,
+      "num_input_tokens_seen": 2746528,
+      "step": 750,
+      "train_runtime": 962.4815,
+      "train_tokens_per_second": 2853.591
+    },
+    {
+      "epoch": 1.635091496232508,
+      "grad_norm": 3.306402868474834e-05,
+      "learning_rate": 1.2929130713525396e-05,
+      "loss": 0.0927,
+      "num_input_tokens_seen": 2782656,
+      "step": 760,
+      "train_runtime": 975.0888,
+      "train_tokens_per_second": 2853.746
+    },
+    {
+      "epoch": 1.6566200215285254,
+      "grad_norm": 3.532276605255902e-05,
+      "learning_rate": 1.2595113223723788e-05,
+      "loss": 0.1135,
+      "num_input_tokens_seen": 2819744,
+      "step": 770,
+      "train_runtime": 987.9762,
+      "train_tokens_per_second": 2854.061
+    },
+    {
+      "epoch": 1.6781485468245425,
+      "grad_norm": 6.426728941733018e-05,
+      "learning_rate": 1.2262315362548825e-05,
+      "loss": 0.0813,
+      "num_input_tokens_seen": 2857216,
+      "step": 780,
+      "train_runtime": 1001.0086,
+      "train_tokens_per_second": 2854.337
+    },
+    {
+      "epoch": 1.6996770721205596,
+      "grad_norm": 8.145067113218829e-05,
+      "learning_rate": 1.1930905907093196e-05,
+      "loss": 0.0933,
+      "num_input_tokens_seen": 2892000,
+      "step": 790,
+      "train_runtime": 1013.4209,
+      "train_tokens_per_second": 2853.701
+    },
+    {
+      "epoch": 1.721205597416577,
+      "grad_norm": 0.00011963866563746706,
+      "learning_rate": 1.1601052930325236e-05,
+      "loss": 0.0881,
+      "num_input_tokens_seen": 2928832,
+      "step": 800,
+      "train_runtime": 1026.3231,
+      "train_tokens_per_second": 2853.714
+    },
+    {
+      "epoch": 1.7427341227125943,
+      "grad_norm": 5.128612610860728e-05,
+      "learning_rate": 1.127292371585141e-05,
+      "loss": 0.079,
+      "num_input_tokens_seen": 2965632,
+      "step": 810,
+      "train_runtime": 1039.1265,
+      "train_tokens_per_second": 2853.966
+    },
+    {
+      "epoch": 1.7642626480086114,
+      "grad_norm": 7.751452358206734e-05,
+      "learning_rate": 1.0946684673079074e-05,
+      "loss": 0.1132,
+      "num_input_tokens_seen": 3002496,
+      "step": 820,
+      "train_runtime": 1051.9839,
+      "train_tokens_per_second": 2854.127
+    },
+    {
+      "epoch": 1.7857911733046286,
+      "grad_norm": 4.0213129977928475e-05,
+      "learning_rate": 1.0622501252822628e-05,
+      "loss": 0.0793,
+      "num_input_tokens_seen": 3038944,
+      "step": 830,
+      "train_runtime": 1064.6802,
+      "train_tokens_per_second": 2854.326
+    },
+    {
+      "epoch": 1.807319698600646,
+      "grad_norm": 4.586371505865827e-05,
+      "learning_rate": 1.0300537863395823e-05,
+      "loss": 0.0918,
+      "num_input_tokens_seen": 3076032,
+      "step": 840,
+      "train_runtime": 1077.5859,
+      "train_tokens_per_second": 2854.559
+    },
+    {
+      "epoch": 1.8288482238966632,
+      "grad_norm": 3.2480071240570396e-05,
+      "learning_rate": 9.98095778723271e-06,
+      "loss": 0.093,
+      "num_input_tokens_seen": 3112448,
+      "step": 850,
+      "train_runtime": 1090.3056,
+      "train_tokens_per_second": 2854.656
+    },
+    {
+      "epoch": 1.8503767491926801,
+      "grad_norm": 5.541834980249405e-05,
+      "learning_rate": 9.663923098079656e-06,
+      "loss": 0.0973,
+      "num_input_tokens_seen": 3149376,
+      "step": 860,
+      "train_runtime": 1103.1339,
+      "train_tokens_per_second": 2854.935
+    },
+    {
+      "epoch": 1.8719052744886975,
+      "grad_norm": 3.055919296457432e-05,
+      "learning_rate": 9.349594578800259e-06,
+      "loss": 0.0995,
+      "num_input_tokens_seen": 3186304,
+      "step": 870,
+      "train_runtime": 1116.0618,
+      "train_tokens_per_second": 2854.953
+    },
+    {
+      "epoch": 1.8934337997847148,
+      "grad_norm": 4.6285113057820126e-05,
+      "learning_rate": 9.038131639834977e-06,
+      "loss": 0.1238,
+      "num_input_tokens_seen": 3222752,
+      "step": 880,
+      "train_runtime": 1128.7812,
+      "train_tokens_per_second": 2855.072
+    },
+    {
+      "epoch": 1.914962325080732,
+      "grad_norm": 4.483849261305295e-05,
+      "learning_rate": 8.7296922383568e-06,
+      "loss": 0.0912,
+      "num_input_tokens_seen": 3258976,
+      "step": 890,
+      "train_runtime": 1141.5159,
+      "train_tokens_per_second": 2854.954
+    },
+    {
+      "epoch": 1.936490850376749,
+      "grad_norm": 5.5885881010908633e-05,
+      "learning_rate": 8.424432798163838e-06,
+      "loss": 0.0935,
+      "num_input_tokens_seen": 3295040,
+      "step": 900,
+      "train_runtime": 1154.1132,
+      "train_tokens_per_second": 2855.041
+    },
+    {
+      "epoch": 1.9580193756727664,
+      "grad_norm": 3.8836213207105175e-05,
+      "learning_rate": 8.122508130349711e-06,
+      "loss": 0.0722,
+      "num_input_tokens_seen": 3331392,
+      "step": 910,
+      "train_runtime": 1166.7365,
+      "train_tokens_per_second": 2855.308
+    },
+    {
+      "epoch": 1.9795479009687837,
+      "grad_norm": 6.445301551138982e-05,
+      "learning_rate": 7.82407135479164e-06,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 3368032,
+      "step": 920,
+      "train_runtime": 1179.5494,
+      "train_tokens_per_second": 2855.355
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.651374805078376e-05,
+      "learning_rate": 7.529273822496379e-06,
+      "loss": 0.0647,
+      "num_input_tokens_seen": 3403584,
+      "step": 930,
+      "train_runtime": 1191.8451,
+      "train_tokens_per_second": 2855.727
+    },
+    {
+      "epoch": 2.0215285252960173,
+      "grad_norm": 9.509055962553248e-05,
+      "learning_rate": 7.238265038843143e-06,
+      "loss": 0.0881,
+      "num_input_tokens_seen": 3440000,
+      "step": 940,
+      "train_runtime": 1204.5181,
+      "train_tokens_per_second": 2855.914
+    },
+    {
+      "epoch": 2.0430570505920342,
+      "grad_norm": 5.2039082220289856e-05,
+      "learning_rate": 6.951192587762555e-06,
+      "loss": 0.0573,
+      "num_input_tokens_seen": 3476640,
+      "step": 950,
+      "train_runtime": 1217.2855,
+      "train_tokens_per_second": 2856.06
+    },
+    {
+      "epoch": 2.0645855758880516,
+      "grad_norm": 3.832239599432796e-05,
+      "learning_rate": 6.668202056890052e-06,
+      "loss": 0.0646,
+      "num_input_tokens_seen": 3512992,
+      "step": 960,
+      "train_runtime": 1229.9665,
+      "train_tokens_per_second": 2856.169
+    },
+    {
+      "epoch": 2.086114101184069,
+      "grad_norm": 5.945537122897804e-05,
+      "learning_rate": 6.389436963731707e-06,
+      "loss": 0.0682,
+      "num_input_tokens_seen": 3549504,
+      "step": 970,
+      "train_runtime": 1242.6763,
+      "train_tokens_per_second": 2856.338
+    },
+    {
+      "epoch": 2.1076426264800863,
+      "grad_norm": 6.592880527023226e-05,
+      "learning_rate": 6.115038682879904e-06,
+      "loss": 0.0931,
+      "num_input_tokens_seen": 3586592,
+      "step": 980,
+      "train_runtime": 1255.5495,
+      "train_tokens_per_second": 2856.591
+    },
+    {
+      "epoch": 2.129171151776103,
+      "grad_norm": 2.9424012609524652e-05,
+      "learning_rate": 5.845146374315781e-06,
+      "loss": 0.0686,
+      "num_input_tokens_seen": 3622944,
+      "step": 990,
+      "train_runtime": 1268.2309,
+      "train_tokens_per_second": 2856.691
+    },
+    {
+      "epoch": 2.1506996770721205,
+      "grad_norm": 5.797382618766278e-05,
+      "learning_rate": 5.579896912834807e-06,
+      "loss": 0.0917,
+      "num_input_tokens_seen": 3660096,
+      "step": 1000,
+      "train_runtime": 1281.1218,
+      "train_tokens_per_second": 2856.946
+    },
+    {
+      "epoch": 2.172228202368138,
+      "grad_norm": 5.529261034098454e-05,
+      "learning_rate": 5.319424818631293e-06,
+      "loss": 0.0931,
+      "num_input_tokens_seen": 3697024,
+      "step": 1010,
+      "train_runtime": 1294.9799,
+      "train_tokens_per_second": 2854.889
+    },
+    {
+      "epoch": 2.193756727664155,
+      "grad_norm": 5.912356573389843e-05,
+      "learning_rate": 5.063862189076988e-06,
+      "loss": 0.0855,
+      "num_input_tokens_seen": 3734240,
+      "step": 1020,
+      "train_runtime": 1307.7689,
+      "train_tokens_per_second": 2855.428
+    },
+    {
+      "epoch": 2.215285252960172,
+      "grad_norm": 5.3946208936395124e-05,
+      "learning_rate": 4.813338631728471e-06,
+      "loss": 0.0909,
+      "num_input_tokens_seen": 3771232,
+      "step": 1030,
+      "train_runtime": 1320.5384,
+      "train_tokens_per_second": 2855.829
+    },
+    {
+      "epoch": 2.2368137782561894,
+      "grad_norm": 6.292902253335342e-05,
+      "learning_rate": 4.567981198597168e-06,
+      "loss": 0.0817,
+      "num_input_tokens_seen": 3808416,
+      "step": 1040,
+      "train_runtime": 1333.3555,
+      "train_tokens_per_second": 2856.264
+    },
+    {
+      "epoch": 2.2583423035522068,
+      "grad_norm": 4.050394636578858e-05,
+      "learning_rate": 4.327914321715459e-06,
+      "loss": 0.0599,
+      "num_input_tokens_seen": 3844768,
+      "step": 1050,
+      "train_runtime": 1345.9495,
+      "train_tokens_per_second": 2856.547
+    },
+    {
+      "epoch": 2.2798708288482237,
+      "grad_norm": 5.740273991250433e-05,
+      "learning_rate": 4.09325975003146e-06,
+      "loss": 0.0711,
+      "num_input_tokens_seen": 3881088,
+      "step": 1060,
+      "train_runtime": 1358.6041,
+      "train_tokens_per_second": 2856.673
+    },
+    {
+      "epoch": 2.301399354144241,
+      "grad_norm": 7.280943100340664e-05,
+      "learning_rate": 3.864136487664538e-06,
+      "loss": 0.0723,
+      "num_input_tokens_seen": 3916928,
+      "step": 1070,
+      "train_runtime": 1371.1849,
+      "train_tokens_per_second": 2856.601
+    },
+    {
+      "epoch": 2.3229278794402584,
+      "grad_norm": 4.93960760650225e-05,
+      "learning_rate": 3.640660733552863e-06,
+      "loss": 0.0524,
+      "num_input_tokens_seen": 3952928,
+      "step": 1080,
+      "train_runtime": 1383.7869,
+      "train_tokens_per_second": 2856.602
+    },
+    {
+      "epoch": 2.3444564047362757,
+      "grad_norm": 7.801376341376454e-05,
+      "learning_rate": 3.4229458225235696e-06,
+      "loss": 0.0513,
+      "num_input_tokens_seen": 3990720,
+      "step": 1090,
+      "train_runtime": 1396.9725,
+      "train_tokens_per_second": 2856.692
+    },
+    {
+      "epoch": 2.3659849300322926,
+      "grad_norm": 4.576309584081173e-05,
+      "learning_rate": 3.211102167815469e-06,
+      "loss": 0.0773,
+      "num_input_tokens_seen": 4026656,
+      "step": 1100,
+      "train_runtime": 1409.5674,
+      "train_tokens_per_second": 2856.661
+    },
+    {
+      "epoch": 2.38751345532831,
+      "grad_norm": 0.0001336225977865979,
+      "learning_rate": 3.0052372050834025e-06,
+      "loss": 0.0896,
+      "num_input_tokens_seen": 4064192,
+      "step": 1110,
+      "train_runtime": 1422.5111,
+      "train_tokens_per_second": 2857.055
+    },
+    {
+      "epoch": 2.4090419806243273,
+      "grad_norm": 8.807591075310484e-05,
+      "learning_rate": 2.805455337912722e-06,
+      "loss": 0.0719,
+      "num_input_tokens_seen": 4101632,
+      "step": 1120,
+      "train_runtime": 1435.3867,
+      "train_tokens_per_second": 2857.51
+    },
+    {
+      "epoch": 2.4305705059203446,
+      "grad_norm": 3.673565879580565e-05,
+      "learning_rate": 2.611857884871379e-06,
+      "loss": 0.0801,
+      "num_input_tokens_seen": 4138208,
+      "step": 1130,
+      "train_runtime": 1448.0115,
+      "train_tokens_per_second": 2857.856
+    },
+    {
+      "epoch": 2.4520990312163615,
+      "grad_norm": 2.5486542654107325e-05,
+      "learning_rate": 2.424543028126661e-06,
+      "loss": 0.0821,
+      "num_input_tokens_seen": 4174592,
+      "step": 1140,
+      "train_runtime": 1460.6997,
+      "train_tokens_per_second": 2857.94
+    },
+    {
+      "epoch": 2.473627556512379,
+      "grad_norm": 6.899920845171437e-05,
+      "learning_rate": 2.243605763652489e-06,
+      "loss": 0.0654,
+      "num_input_tokens_seen": 4211616,
+      "step": 1150,
+      "train_runtime": 1473.5308,
+      "train_tokens_per_second": 2858.18
+    },
+    {
+      "epoch": 2.495156081808396,
+      "grad_norm": 4.8038589739007875e-05,
+      "learning_rate": 2.0691378530525933e-06,
+      "loss": 0.0767,
+      "num_input_tokens_seen": 4247744,
+      "step": 1160,
+      "train_runtime": 1486.1222,
+      "train_tokens_per_second": 2858.274
+    },
+    {
+      "epoch": 2.5166846071044136,
+      "grad_norm": 7.49904866097495e-05,
+      "learning_rate": 1.9012277770239829e-06,
+      "loss": 0.0655,
+      "num_input_tokens_seen": 4284224,
+      "step": 1170,
+      "train_runtime": 1498.696,
+      "train_tokens_per_second": 2858.634
+    },
+    {
+      "epoch": 2.5382131324004304,
+      "grad_norm": 6.3829000282567e-05,
+      "learning_rate": 1.7399606904843224e-06,
+      "loss": 0.0584,
+      "num_input_tokens_seen": 4320672,
+      "step": 1180,
+      "train_runtime": 1511.4477,
+      "train_tokens_per_second": 2858.631
+    },
+    {
+      "epoch": 2.559741657696448,
+      "grad_norm": 5.6509288697270676e-05,
+      "learning_rate": 1.5854183793859577e-06,
+      "loss": 0.0601,
+      "num_input_tokens_seen": 4358080,
+      "step": 1190,
+      "train_runtime": 1524.3591,
+      "train_tokens_per_second": 2858.959
+    },
+    {
+      "epoch": 2.581270182992465,
+      "grad_norm": 4.297928171581589e-05,
+      "learning_rate": 1.437679219238498e-06,
+      "loss": 0.0665,
+      "num_input_tokens_seen": 4394752,
+      "step": 1200,
+      "train_runtime": 1537.0429,
+      "train_tokens_per_second": 2859.225
+    },
+    {
+      "epoch": 2.602798708288482,
+      "grad_norm": 8.681658800924197e-05,
+      "learning_rate": 1.2968181353609854e-06,
+      "loss": 0.0658,
+      "num_input_tokens_seen": 4431808,
+      "step": 1210,
+      "train_runtime": 1549.8548,
+      "train_tokens_per_second": 2859.499
+    },
+    {
+      "epoch": 2.6243272335844994,
+      "grad_norm": 3.503639891277999e-05,
+      "learning_rate": 1.1629065648838495e-06,
+      "loss": 0.0898,
+      "num_input_tokens_seen": 4467328,
+      "step": 1220,
+      "train_runtime": 1562.3415,
+      "train_tokens_per_second": 2859.38
+    },
+    {
+      "epoch": 2.6458557588805167,
+      "grad_norm": 3.0625513318227604e-05,
+      "learning_rate": 1.0360124205198102e-06,
+      "loss": 0.0619,
+      "num_input_tokens_seen": 4503808,
+      "step": 1230,
+      "train_runtime": 1575.0206,
+      "train_tokens_per_second": 2859.523
+    },
+    {
+      "epoch": 2.667384284176534,
+      "grad_norm": 6.452269735746086e-05,
+      "learning_rate": 9.162000561222672e-07,
+      "loss": 0.0763,
+      "num_input_tokens_seen": 4541536,
+      "step": 1240,
+      "train_runtime": 1588.024,
+      "train_tokens_per_second": 2859.866
+    },
+    {
+      "epoch": 2.6889128094725514,
+      "grad_norm": 2.9151902708690614e-05,
+      "learning_rate": 8.035302340484646e-07,
+      "loss": 0.0909,
+      "num_input_tokens_seen": 4577664,
+      "step": 1250,
+      "train_runtime": 1600.6009,
+      "train_tokens_per_second": 2859.966
+    },
+    {
+      "epoch": 2.7104413347685683,
+      "grad_norm": 6.271711754379794e-05,
+      "learning_rate": 6.980600943441273e-07,
+      "loss": 0.0683,
+      "num_input_tokens_seen": 4614176,
+      "step": 1260,
+      "train_runtime": 1613.2327,
+      "train_tokens_per_second": 2860.205
+    },
+    {
+      "epoch": 2.7319698600645856,
+      "grad_norm": 5.843062899657525e-05,
+      "learning_rate": 5.998431257651132e-07,
+      "loss": 0.0867,
+      "num_input_tokens_seen": 4650848,
+      "step": 1270,
+      "train_runtime": 1625.9328,
+      "train_tokens_per_second": 2860.418
+    },
+    {
+      "epoch": 2.7534983853606025,
+      "grad_norm": 5.628829603665508e-05,
+      "learning_rate": 5.089291386508055e-07,
+      "loss": 0.0648,
+      "num_input_tokens_seen": 4687456,
+      "step": 1280,
+      "train_runtime": 1638.6467,
+      "train_tokens_per_second": 2860.565
+    },
+    {
+      "epoch": 2.77502691065662,
+      "grad_norm": 4.340551095083356e-05,
+      "learning_rate": 4.2536423966299256e-07,
+      "loss": 0.0672,
+      "num_input_tokens_seen": 4723840,
+      "step": 1290,
+      "train_runtime": 1651.315,
+      "train_tokens_per_second": 2860.653
+    },
+    {
+      "epoch": 2.7965554359526372,
+      "grad_norm": 7.965038093971089e-05,
+      "learning_rate": 3.491908084030776e-07,
+      "loss": 0.089,
+      "num_input_tokens_seen": 4761056,
+      "step": 1300,
+      "train_runtime": 1664.1238,
+      "train_tokens_per_second": 2860.999
+    },
+    {
+      "epoch": 2.8180839612486546,
+      "grad_norm": 5.358497583074495e-05,
+      "learning_rate": 2.804474759194176e-07,
+      "loss": 0.07,
+      "num_input_tokens_seen": 4798240,
+      "step": 1310,
+      "train_runtime": 1676.8916,
+      "train_tokens_per_second": 2861.39
+    },
+    {
+      "epoch": 2.839612486544672,
+      "grad_norm": 3.2774096325738356e-05,
+      "learning_rate": 2.1916910511575627e-07,
+      "loss": 0.0807,
+      "num_input_tokens_seen": 4834560,
+      "step": 1320,
+      "train_runtime": 1689.486,
+      "train_tokens_per_second": 2861.557
+    },
+    {
+      "epoch": 2.861141011840689,
+      "grad_norm": 6.476557609857991e-05,
+      "learning_rate": 1.6538677307062933e-07,
+      "loss": 0.0669,
+      "num_input_tokens_seen": 4871328,
+      "step": 1330,
+      "train_runtime": 1702.2714,
+      "train_tokens_per_second": 2861.664
+    },
+    {
+      "epoch": 2.882669537136706,
+      "grad_norm": 8.176828123396263e-05,
+      "learning_rate": 1.1912775527676278e-07,
+      "loss": 0.0552,
+      "num_input_tokens_seen": 4907744,
+      "step": 1340,
+      "train_runtime": 1714.8671,
+      "train_tokens_per_second": 2861.88
+    },
+    {
+      "epoch": 2.9041980624327235,
+      "grad_norm": 7.269597699632868e-05,
+      "learning_rate": 8.04155118084099e-08,
+      "loss": 0.0917,
+      "num_input_tokens_seen": 4943712,
+      "step": 1350,
+      "train_runtime": 1727.4271,
+      "train_tokens_per_second": 2861.893
+    },
+    {
+      "epoch": 2.9257265877287404,
+      "grad_norm": 3.572047353372909e-05,
+      "learning_rate": 4.9269675423671556e-08,
+      "loss": 0.074,
+      "num_input_tokens_seen": 4980480,
+      "step": 1360,
+      "train_runtime": 1740.2346,
+      "train_tokens_per_second": 2861.959
+    },
+    {
+      "epoch": 2.9472551130247577,
+      "grad_norm": 8.610687655163929e-05,
+      "learning_rate": 2.570604160781753e-08,
+      "loss": 0.0599,
+      "num_input_tokens_seen": 5016800,
+      "step": 1370,
+      "train_runtime": 1752.9074,
+      "train_tokens_per_second": 2861.988
+    },
+    {
+      "epoch": 2.968783638320775,
+      "grad_norm": 5.934851287747733e-05,
+      "learning_rate": 9.736560562675912e-09,
+      "loss": 0.0593,
+      "num_input_tokens_seen": 5053440,
+      "step": 1380,
+      "train_runtime": 1765.6959,
+      "train_tokens_per_second": 2862.01
+    },
+    {
+      "epoch": 2.9903121636167924,
+      "grad_norm": 3.7563720979960635e-05,
+      "learning_rate": 1.3693311461326908e-09,
+      "loss": 0.0763,
+      "num_input_tokens_seen": 5090592,
+      "step": 1390,
+      "train_runtime": 1778.5239,
+      "train_tokens_per_second": 2862.257
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 1395,
+  "num_input_tokens_seen": 5107072,
+  "num_train_epochs": 3,
+  "save_steps": 1000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.186269295986606e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-1395/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0da07f323cd8cadc582879cf09f450ad877537a29fbf8dd1bdb4c30cac53e814
+size 6161

llamaboard_config.yaml

@@ -0,0 +1,85 @@
+top.booster: unsloth
+top.checkpoint_path: []
+top.finetuning_type: lora
+top.model_name: Mistral-7B-Instruct-v0.3
+top.quantization_bit: '4'
+top.quantization_method: bnb
+top.rope_scaling: none
+top.template: alpaca
+train.additional_target: ''
+train.apollo_rank: 16
+train.apollo_scale: 32
+train.apollo_target: all
+train.apollo_update_interval: 200
+train.badam_mode: layer
+train.badam_switch_interval: 50
+train.badam_switch_mode: ascending
+train.badam_update_ratio: 0.05
+train.batch_size: 4
+train.compute_type: fp16
+train.create_new_adapter: false
+train.cutoff_len: 2048
+train.dataset:
+- treino_pt_rde
+train.dataset_dir: data
+train.ds_offload: false
+train.ds_stage: none
+train.enable_thinking: true
+train.extra_args: '{"optim": "adamw_torch"}'
+train.freeze_extra_modules: ''
+train.freeze_language_model: false
+train.freeze_multi_modal_projector: true
+train.freeze_trainable_layers: 2
+train.freeze_trainable_modules: all
+train.freeze_vision_tower: true
+train.galore_rank: 16
+train.galore_scale: 2
+train.galore_target: all
+train.galore_update_interval: 200
+train.gradient_accumulation_steps: 2
+train.image_max_pixels: 768*768
+train.image_min_pixels: 32*32
+train.learning_rate: 3e-5
+train.logging_steps: 10
+train.lora_alpha: 16
+train.lora_dropout: 0
+train.lora_rank: 8
+train.lora_target: ''
+train.loraplus_lr_ratio: 0
+train.lr_scheduler_type: cosine
+train.mask_history: false
+train.max_grad_norm: '1.0'
+train.max_samples: '3716'
+train.neat_packing: false
+train.neftune_alpha: 0
+train.num_train_epochs: '3.0'
+train.packing: false
+train.ppo_score_norm: false
+train.ppo_whiten_rewards: false
+train.pref_beta: 0.1
+train.pref_ftx: 0
+train.pref_loss: sigmoid
+train.report_to: none
+train.resize_vocab: false
+train.reward_model: []
+train.save_steps: 1000
+train.swanlab_api_key: ''
+train.swanlab_link: ''
+train.swanlab_mode: cloud
+train.swanlab_project: llamafactory
+train.swanlab_run_name: ''
+train.swanlab_workspace: ''
+train.train_on_prompt: false
+train.training_stage: Supervised Fine-Tuning
+train.use_apollo: false
+train.use_badam: false
+train.use_dora: false
+train.use_galore: false
+train.use_llama_pro: false
+train.use_pissa: false
+train.use_rslora: false
+train.use_swanlab: false
+train.val_size: 0
+train.video_max_pixels: 256*256
+train.video_min_pixels: 16*16
+train.warmup_steps: 0

running_log.txt

@@ -0,0 +1,726 @@
+[INFO|2025-09-23 13:33:40] tokenization_utils_base.py:2023 >> loading file tokenizer.model from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/tokenizer.model
+
+[INFO|2025-09-23 13:33:40] tokenization_utils_base.py:2023 >> loading file tokenizer.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/tokenizer.json
+
+[INFO|2025-09-23 13:33:40] tokenization_utils_base.py:2023 >> loading file added_tokens.json from cache at None
+
+[INFO|2025-09-23 13:33:40] tokenization_utils_base.py:2023 >> loading file special_tokens_map.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/special_tokens_map.json
+
+[INFO|2025-09-23 13:33:40] tokenization_utils_base.py:2023 >> loading file tokenizer_config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/tokenizer_config.json
+
+[INFO|2025-09-23 13:33:40] tokenization_utils_base.py:2023 >> loading file chat_template.jinja from cache at None
+
+[INFO|2025-09-23 13:33:43] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/config.json
+
+[INFO|2025-09-23 13:33:43] configuration_utils.py:770 >> Model config MistralConfig {
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": null,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "use_cache": true,
+  "vocab_size": 32768
+}
+
+
+[INFO|2025-09-23 13:33:43] tokenization_utils_base.py:2023 >> loading file tokenizer.model from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/tokenizer.model
+
+[INFO|2025-09-23 13:33:43] tokenization_utils_base.py:2023 >> loading file tokenizer.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/tokenizer.json
+
+[INFO|2025-09-23 13:33:43] tokenization_utils_base.py:2023 >> loading file added_tokens.json from cache at None
+
+[INFO|2025-09-23 13:33:43] tokenization_utils_base.py:2023 >> loading file special_tokens_map.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/special_tokens_map.json
+
+[INFO|2025-09-23 13:33:43] tokenization_utils_base.py:2023 >> loading file tokenizer_config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/tokenizer_config.json
+
+[INFO|2025-09-23 13:33:43] tokenization_utils_base.py:2023 >> loading file chat_template.jinja from cache at None
+
+[INFO|2025-09-23 13:33:43] logging.py:143 >> Add pad token: </s>
+
+[INFO|2025-09-23 13:33:43] logging.py:143 >> Loading dataset treino_pt_rde.json...
+
+[INFO|2025-09-23 13:33:47] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--mistralai--Mistral-7B-Instruct-v0.3/snapshots/0d4b76e1efeb5eb6f6b5e757c79870472e04bd3a/config.json
+
+[INFO|2025-09-23 13:33:47] configuration_utils.py:770 >> Model config MistralConfig {
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": null,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "use_cache": true,
+  "vocab_size": 32768
+}
+
+
+[INFO|2025-09-23 13:33:47] logging.py:143 >> Quantizing model to 4 bit with bitsandbytes.
+
+[INFO|2025-09-23 13:33:47] logging.py:143 >> KV cache is disabled during training.
+
+[INFO|2025-09-23 13:33:57] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unsloth--mistral-7b-instruct-v0.3-bnb-4bit/snapshots/d5f623888f1415cf89b5c208d09cb620694618ee/config.json
+
+[INFO|2025-09-23 13:33:57] configuration_utils.py:770 >> Model config MistralConfig {
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 770,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "unsloth_version": "2024.9",
+  "use_cache": true,
+  "vocab_size": 32768
+}
+
+
+[INFO|2025-09-23 13:34:00] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unslothai--colabpro/snapshots/234f33d5f3e1d9ad83421f33640cd88474a25025/config.json
+
+[INFO|2025-09-23 13:34:01] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unslothai--repeat/snapshots/7c48478c02f84ed89f149b0815cc0216ee831fb0/config.json
+
+[INFO|2025-09-23 13:34:02] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unslothai--vram-80/snapshots/830433cec98f0aa155f72baba38dc9fe6831d3f3/config.json
+
+[INFO|2025-09-23 13:34:04] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unslothai--1/snapshots/7ec782b7604cd9ea0781c23a4270f031650f5617/config.json
+
+[INFO|2025-09-23 13:34:04] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unsloth--mistral-7b-instruct-v0.3-bnb-4bit/snapshots/d5f623888f1415cf89b5c208d09cb620694618ee/config.json
+
+[INFO|2025-09-23 13:34:04] configuration_utils.py:770 >> Model config MistralConfig {
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 770,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "unsloth_version": "2024.9",
+  "use_cache": true,
+  "vocab_size": 32768
+}
+
+
+[INFO|2025-09-23 13:34:04] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unsloth--mistral-7b-instruct-v0.3-bnb-4bit/snapshots/d5f623888f1415cf89b5c208d09cb620694618ee/config.json
+
+[INFO|2025-09-23 13:34:04] configuration_utils.py:770 >> Model config MistralConfig {
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 770,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "float16",
+  "transformers_version": "4.52.4",
+  "unsloth_version": "2024.9",
+  "use_cache": true,
+  "vocab_size": 32768
+}
+
+
+[INFO|2025-09-23 13:34:21] modeling_utils.py:1151 >> loading weights file model.safetensors from cache at /root/.cache/huggingface/hub/models--unsloth--mistral-7b-instruct-v0.3-bnb-4bit/snapshots/d5f623888f1415cf89b5c208d09cb620694618ee/model.safetensors
+
+[INFO|2025-09-23 13:34:21] modeling_utils.py:2241 >> Instantiating MistralForCausalLM model under default dtype torch.float16.
+
+[INFO|2025-09-23 13:34:21] configuration_utils.py:1135 >> Generate config GenerationConfig {
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 770
+}
+
+
+[INFO|2025-09-23 13:34:23] modeling_utils.py:5131 >> All model checkpoint weights were used when initializing MistralForCausalLM.
+
+
+[INFO|2025-09-23 13:34:23] modeling_utils.py:5139 >> All the weights of MistralForCausalLM were initialized from the model checkpoint at unsloth/mistral-7b-instruct-v0.3-bnb-4bit.
+If your task is similar to the task the model of the checkpoint was trained on, you can already use MistralForCausalLM for predictions without further training.
+
+[INFO|2025-09-23 13:34:24] configuration_utils.py:1090 >> loading configuration file generation_config.json from cache at /root/.cache/huggingface/hub/models--unsloth--mistral-7b-instruct-v0.3-bnb-4bit/snapshots/d5f623888f1415cf89b5c208d09cb620694618ee/generation_config.json
+
+[INFO|2025-09-23 13:34:24] configuration_utils.py:1135 >> Generate config GenerationConfig {
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "max_length": 32768,
+  "pad_token_id": 770
+}
+
+
+[INFO|2025-09-23 13:34:31] logging.py:143 >> Gradient checkpointing enabled.
+
+[INFO|2025-09-23 13:34:31] logging.py:143 >> Upcasting trainable params to float32.
+
+[INFO|2025-09-23 13:34:31] logging.py:143 >> Fine-tuning method: LoRA
+
+[INFO|2025-09-23 13:34:31] logging.py:143 >> Found linear modules: down_proj,gate_proj,o_proj,q_proj,k_proj,up_proj,v_proj
+
+[WARNING|2025-09-23 13:34:34] logging.py:328 >> Unsloth 2025.9.7 patched 32 layers with 32 QKV layers, 32 O layers and 32 MLP layers.
+
+[INFO|2025-09-23 13:34:36] logging.py:143 >> trainable params: 20,971,520 || all params: 7,268,995,072 || trainable%: 0.2885
+
+[INFO|2025-09-23 13:34:36] trainer.py:756 >> Using auto half precision backend
+
+[WARNING|2025-09-23 13:34:37] <string>:173 >> ==((====))==  Unsloth - 2x faster free finetuning | Num GPUs used = 1
+   \\   /|    Num examples = 3,716 | Num Epochs = 3 | Total steps = 1,395
+O^O/ \_/ \    Batch size per device = 4 | Gradient accumulation steps = 2
+\        /    Data Parallel GPUs = 1 | Total batch size (4 x 2 x 1) = 8
+ "-____-"     Trainable parameters = 20,971,520 of 7,268,995,072 (0.29% trained)
+
+[INFO|2025-09-23 13:34:59] logging.py:143 >> {'loss': 0.8869, 'learning_rate': 2.9997e-05, 'epoch': 0.02, 'throughput': 1715.83}
+
+[INFO|2025-09-23 13:35:12] logging.py:143 >> {'loss': 0.1990, 'learning_rate': 2.9986e-05, 'epoch': 0.04, 'throughput': 2143.92}
+
+[INFO|2025-09-23 13:35:25] logging.py:143 >> {'loss': 0.1809, 'learning_rate': 2.9968e-05, 'epoch': 0.06, 'throughput': 2319.02}
+
+[INFO|2025-09-23 13:35:38] logging.py:143 >> {'loss': 0.1766, 'learning_rate': 2.9942e-05, 'epoch': 0.09, 'throughput': 2443.83}
+
+[INFO|2025-09-23 13:35:50] logging.py:143 >> {'loss': 0.1784, 'learning_rate': 2.9909e-05, 'epoch': 0.11, 'throughput': 2519.74}
+
+[INFO|2025-09-23 13:36:03] logging.py:143 >> {'loss': 0.1650, 'learning_rate': 2.9868e-05, 'epoch': 0.13, 'throughput': 2576.27}
+
+[INFO|2025-09-23 13:36:16] logging.py:143 >> {'loss': 0.1884, 'learning_rate': 2.9819e-05, 'epoch': 0.15, 'throughput': 2613.67}
+
+[INFO|2025-09-23 13:36:29] logging.py:143 >> {'loss': 0.1668, 'learning_rate': 2.9763e-05, 'epoch': 0.17, 'throughput': 2644.76}
+
+[INFO|2025-09-23 13:36:41] logging.py:143 >> {'loss': 0.1538, 'learning_rate': 2.9700e-05, 'epoch': 0.19, 'throughput': 2667.87}
+
+[INFO|2025-09-23 13:36:54] logging.py:143 >> {'loss': 0.1729, 'learning_rate': 2.9629e-05, 'epoch': 0.22, 'throughput': 2688.46}
+
+[INFO|2025-09-23 13:37:06] logging.py:143 >> {'loss': 0.1346, 'learning_rate': 2.9550e-05, 'epoch': 0.24, 'throughput': 2704.64}
+
+[INFO|2025-09-23 13:37:19] logging.py:143 >> {'loss': 0.1622, 'learning_rate': 2.9465e-05, 'epoch': 0.26, 'throughput': 2719.24}
+
+[INFO|2025-09-23 13:37:32] logging.py:143 >> {'loss': 0.1357, 'learning_rate': 2.9371e-05, 'epoch': 0.28, 'throughput': 2731.14}
+
+[INFO|2025-09-23 13:37:45] logging.py:143 >> {'loss': 0.1537, 'learning_rate': 2.9271e-05, 'epoch': 0.30, 'throughput': 2741.53}
+
+[INFO|2025-09-23 13:37:57] logging.py:143 >> {'loss': 0.1183, 'learning_rate': 2.9163e-05, 'epoch': 0.32, 'throughput': 2750.14}
+
+[INFO|2025-09-23 13:38:10] logging.py:143 >> {'loss': 0.1336, 'learning_rate': 2.9049e-05, 'epoch': 0.34, 'throughput': 2757.92}
+
+[INFO|2025-09-23 13:38:23] logging.py:143 >> {'loss': 0.1661, 'learning_rate': 2.8927e-05, 'epoch': 0.37, 'throughput': 2763.58}
+
+[INFO|2025-09-23 13:38:35] logging.py:143 >> {'loss': 0.1405, 'learning_rate': 2.8798e-05, 'epoch': 0.39, 'throughput': 2769.62}
+
+[INFO|2025-09-23 13:38:48] logging.py:143 >> {'loss': 0.1456, 'learning_rate': 2.8662e-05, 'epoch': 0.41, 'throughput': 2776.17}
+
+[INFO|2025-09-23 13:39:01] logging.py:143 >> {'loss': 0.1158, 'learning_rate': 2.8519e-05, 'epoch': 0.43, 'throughput': 2781.89}
+
+[INFO|2025-09-23 13:39:13] logging.py:143 >> {'loss': 0.1293, 'learning_rate': 2.8369e-05, 'epoch': 0.45, 'throughput': 2786.89}
+
+[INFO|2025-09-23 13:39:26] logging.py:143 >> {'loss': 0.1456, 'learning_rate': 2.8212e-05, 'epoch': 0.47, 'throughput': 2791.68}
+
+[INFO|2025-09-23 13:39:39] logging.py:143 >> {'loss': 0.1428, 'learning_rate': 2.8049e-05, 'epoch': 0.50, 'throughput': 2795.54}
+
+[INFO|2025-09-23 13:39:51] logging.py:143 >> {'loss': 0.1334, 'learning_rate': 2.7879e-05, 'epoch': 0.52, 'throughput': 2800.11}
+
+[INFO|2025-09-23 13:40:04] logging.py:143 >> {'loss': 0.1414, 'learning_rate': 2.7703e-05, 'epoch': 0.54, 'throughput': 2803.40}
+
+[INFO|2025-09-23 13:40:17] logging.py:143 >> {'loss': 0.1098, 'learning_rate': 2.7520e-05, 'epoch': 0.56, 'throughput': 2808.20}
+
+[INFO|2025-09-23 13:40:29] logging.py:143 >> {'loss': 0.1643, 'learning_rate': 2.7331e-05, 'epoch': 0.58, 'throughput': 2810.83}
+
+[INFO|2025-09-23 13:40:42] logging.py:143 >> {'loss': 0.1222, 'learning_rate': 2.7135e-05, 'epoch': 0.60, 'throughput': 2813.81}
+
+[INFO|2025-09-23 13:40:55] logging.py:143 >> {'loss': 0.1489, 'learning_rate': 2.6934e-05, 'epoch': 0.62, 'throughput': 2817.18}
+
+[INFO|2025-09-23 13:41:08] logging.py:143 >> {'loss': 0.1207, 'learning_rate': 2.6726e-05, 'epoch': 0.65, 'throughput': 2818.65}
+
+[INFO|2025-09-23 13:41:20] logging.py:143 >> {'loss': 0.1319, 'learning_rate': 2.6512e-05, 'epoch': 0.67, 'throughput': 2819.84}
+
+[INFO|2025-09-23 13:41:33] logging.py:143 >> {'loss': 0.1453, 'learning_rate': 2.6293e-05, 'epoch': 0.69, 'throughput': 2822.06}
+
+[INFO|2025-09-23 13:41:46] logging.py:143 >> {'loss': 0.1202, 'learning_rate': 2.6068e-05, 'epoch': 0.71, 'throughput': 2824.06}
+
+[INFO|2025-09-23 13:41:59] logging.py:143 >> {'loss': 0.1283, 'learning_rate': 2.5837e-05, 'epoch': 0.73, 'throughput': 2825.64}
+
+[INFO|2025-09-23 13:42:12] logging.py:143 >> {'loss': 0.1374, 'learning_rate': 2.5601e-05, 'epoch': 0.75, 'throughput': 2828.62}
+
+[INFO|2025-09-23 13:42:24] logging.py:143 >> {'loss': 0.1355, 'learning_rate': 2.5359e-05, 'epoch': 0.78, 'throughput': 2830.17}
+
+[INFO|2025-09-23 13:42:38] logging.py:143 >> {'loss': 0.1310, 'learning_rate': 2.5112e-05, 'epoch': 0.80, 'throughput': 2832.70}
+
+[INFO|2025-09-23 13:42:50] logging.py:143 >> {'loss': 0.1034, 'learning_rate': 2.4860e-05, 'epoch': 0.82, 'throughput': 2834.57}
+
+[INFO|2025-09-23 13:43:03] logging.py:143 >> {'loss': 0.1465, 'learning_rate': 2.4603e-05, 'epoch': 0.84, 'throughput': 2835.08}
+
+[INFO|2025-09-23 13:43:15] logging.py:143 >> {'loss': 0.1049, 'learning_rate': 2.4341e-05, 'epoch': 0.86, 'throughput': 2835.78}
+
+[INFO|2025-09-23 13:43:28] logging.py:143 >> {'loss': 0.1024, 'learning_rate': 2.4074e-05, 'epoch': 0.88, 'throughput': 2837.37}
+
+[INFO|2025-09-23 13:43:41] logging.py:143 >> {'loss': 0.1155, 'learning_rate': 2.3803e-05, 'epoch': 0.90, 'throughput': 2838.63}
+
+[INFO|2025-09-23 13:43:53] logging.py:143 >> {'loss': 0.1170, 'learning_rate': 2.3527e-05, 'epoch': 0.93, 'throughput': 2840.25}
+
+[INFO|2025-09-23 13:44:06] logging.py:143 >> {'loss': 0.1023, 'learning_rate': 2.3247e-05, 'epoch': 0.95, 'throughput': 2840.35}
+
+[INFO|2025-09-23 13:44:18] logging.py:143 >> {'loss': 0.1046, 'learning_rate': 2.2963e-05, 'epoch': 0.97, 'throughput': 2840.69}
+
+[INFO|2025-09-23 13:44:31] logging.py:143 >> {'loss': 0.1106, 'learning_rate': 2.2675e-05, 'epoch': 0.99, 'throughput': 2841.62}
+
+[INFO|2025-09-23 13:44:43] logging.py:143 >> {'loss': 0.0859, 'learning_rate': 2.2383e-05, 'epoch': 1.01, 'throughput': 2842.72}
+
+[INFO|2025-09-23 13:44:56] logging.py:143 >> {'loss': 0.1089, 'learning_rate': 2.2087e-05, 'epoch': 1.03, 'throughput': 2843.73}
+
+[INFO|2025-09-23 13:45:09] logging.py:143 >> {'loss': 0.0852, 'learning_rate': 2.1787e-05, 'epoch': 1.05, 'throughput': 2845.12}
+
+[INFO|2025-09-23 13:45:21] logging.py:143 >> {'loss': 0.0988, 'learning_rate': 2.1484e-05, 'epoch': 1.08, 'throughput': 2846.63}
+
+[INFO|2025-09-23 13:45:34] logging.py:143 >> {'loss': 0.1001, 'learning_rate': 2.1178e-05, 'epoch': 1.10, 'throughput': 2847.50}
+
+[INFO|2025-09-23 13:45:47] logging.py:143 >> {'loss': 0.1107, 'learning_rate': 2.0869e-05, 'epoch': 1.12, 'throughput': 2848.51}
+
+[INFO|2025-09-23 13:45:59] logging.py:143 >> {'loss': 0.0894, 'learning_rate': 2.0556e-05, 'epoch': 1.14, 'throughput': 2848.98}
+
+[INFO|2025-09-23 13:46:12] logging.py:143 >> {'loss': 0.1016, 'learning_rate': 2.0241e-05, 'epoch': 1.16, 'throughput': 2849.25}
+
+[INFO|2025-09-23 13:46:25] logging.py:143 >> {'loss': 0.0932, 'learning_rate': 1.9923e-05, 'epoch': 1.18, 'throughput': 2850.47}
+
+[INFO|2025-09-23 13:46:38] logging.py:143 >> {'loss': 0.1056, 'learning_rate': 1.9603e-05, 'epoch': 1.20, 'throughput': 2851.01}
+
+[INFO|2025-09-23 13:46:50] logging.py:143 >> {'loss': 0.1033, 'learning_rate': 1.9280e-05, 'epoch': 1.23, 'throughput': 2851.17}
+
+[INFO|2025-09-23 13:47:03] logging.py:143 >> {'loss': 0.1035, 'learning_rate': 1.8956e-05, 'epoch': 1.25, 'throughput': 2851.42}
+
+[INFO|2025-09-23 13:47:16] logging.py:143 >> {'loss': 0.0811, 'learning_rate': 1.8629e-05, 'epoch': 1.27, 'throughput': 2851.58}
+
+[INFO|2025-09-23 13:47:28] logging.py:143 >> {'loss': 0.0888, 'learning_rate': 1.8300e-05, 'epoch': 1.29, 'throughput': 2851.53}
+
+[INFO|2025-09-23 13:47:41] logging.py:143 >> {'loss': 0.0968, 'learning_rate': 1.7970e-05, 'epoch': 1.31, 'throughput': 2852.20}
+
+[INFO|2025-09-23 13:47:54] logging.py:143 >> {'loss': 0.1311, 'learning_rate': 1.7638e-05, 'epoch': 1.33, 'throughput': 2852.02}
+
+[INFO|2025-09-23 13:48:07] logging.py:143 >> {'loss': 0.1001, 'learning_rate': 1.7305e-05, 'epoch': 1.36, 'throughput': 2852.57}
+
+[INFO|2025-09-23 13:48:20] logging.py:143 >> {'loss': 0.0887, 'learning_rate': 1.6970e-05, 'epoch': 1.38, 'throughput': 2852.79}
+
+[INFO|2025-09-23 13:48:32] logging.py:143 >> {'loss': 0.1041, 'learning_rate': 1.6635e-05, 'epoch': 1.40, 'throughput': 2852.95}
+
+[INFO|2025-09-23 13:48:45] logging.py:143 >> {'loss': 0.1022, 'learning_rate': 1.6299e-05, 'epoch': 1.42, 'throughput': 2852.78}
+
+[INFO|2025-09-23 13:48:58] logging.py:143 >> {'loss': 0.0831, 'learning_rate': 1.5962e-05, 'epoch': 1.44, 'throughput': 2853.25}
+
+[INFO|2025-09-23 13:49:11] logging.py:143 >> {'loss': 0.1055, 'learning_rate': 1.5625e-05, 'epoch': 1.46, 'throughput': 2853.64}
+
+[INFO|2025-09-23 13:49:24] logging.py:143 >> {'loss': 0.0812, 'learning_rate': 1.5287e-05, 'epoch': 1.48, 'throughput': 2853.75}
+
+[INFO|2025-09-23 13:49:36] logging.py:143 >> {'loss': 0.1016, 'learning_rate': 1.4949e-05, 'epoch': 1.51, 'throughput': 2854.04}
+
+[INFO|2025-09-23 13:49:49] logging.py:143 >> {'loss': 0.0956, 'learning_rate': 1.4612e-05, 'epoch': 1.53, 'throughput': 2854.16}
+
+[INFO|2025-09-23 13:50:02] logging.py:143 >> {'loss': 0.0936, 'learning_rate': 1.4274e-05, 'epoch': 1.55, 'throughput': 2854.51}
+
+[INFO|2025-09-23 13:50:15] logging.py:143 >> {'loss': 0.0728, 'learning_rate': 1.3937e-05, 'epoch': 1.57, 'throughput': 2852.88}
+
+[INFO|2025-09-23 13:50:27] logging.py:143 >> {'loss': 0.0839, 'learning_rate': 1.3600e-05, 'epoch': 1.59, 'throughput': 2852.96}
+
+[INFO|2025-09-23 13:50:40] logging.py:143 >> {'loss': 0.1308, 'learning_rate': 1.3264e-05, 'epoch': 1.61, 'throughput': 2853.61}
+
+[INFO|2025-09-23 13:50:53] logging.py:143 >> {'loss': 0.0927, 'learning_rate': 1.2929e-05, 'epoch': 1.64, 'throughput': 2853.76}
+
+[INFO|2025-09-23 13:51:06] logging.py:143 >> {'loss': 0.1135, 'learning_rate': 1.2595e-05, 'epoch': 1.66, 'throughput': 2854.08}
+
+[INFO|2025-09-23 13:51:19] logging.py:143 >> {'loss': 0.0813, 'learning_rate': 1.2262e-05, 'epoch': 1.68, 'throughput': 2854.35}
+
+[INFO|2025-09-23 13:51:31] logging.py:143 >> {'loss': 0.0933, 'learning_rate': 1.1931e-05, 'epoch': 1.70, 'throughput': 2853.72}
+
+[INFO|2025-09-23 13:51:44] logging.py:143 >> {'loss': 0.0881, 'learning_rate': 1.1601e-05, 'epoch': 1.72, 'throughput': 2853.73}
+
+[INFO|2025-09-23 13:51:57] logging.py:143 >> {'loss': 0.0790, 'learning_rate': 1.1273e-05, 'epoch': 1.74, 'throughput': 2853.98}
+
+[INFO|2025-09-23 13:52:10] logging.py:143 >> {'loss': 0.1132, 'learning_rate': 1.0947e-05, 'epoch': 1.76, 'throughput': 2854.14}
+
+[INFO|2025-09-23 13:52:22] logging.py:143 >> {'loss': 0.0793, 'learning_rate': 1.0623e-05, 'epoch': 1.79, 'throughput': 2854.34}
+
+[INFO|2025-09-23 13:52:35] logging.py:143 >> {'loss': 0.0918, 'learning_rate': 1.0301e-05, 'epoch': 1.81, 'throughput': 2854.57}
+
+[INFO|2025-09-23 13:52:48] logging.py:143 >> {'loss': 0.0930, 'learning_rate': 9.9810e-06, 'epoch': 1.83, 'throughput': 2854.67}
+
+[INFO|2025-09-23 13:53:01] logging.py:143 >> {'loss': 0.0973, 'learning_rate': 9.6639e-06, 'epoch': 1.85, 'throughput': 2854.95}
+
+[INFO|2025-09-23 13:53:14] logging.py:143 >> {'loss': 0.0995, 'learning_rate': 9.3496e-06, 'epoch': 1.87, 'throughput': 2854.97}
+
+[INFO|2025-09-23 13:53:26] logging.py:143 >> {'loss': 0.1238, 'learning_rate': 9.0381e-06, 'epoch': 1.89, 'throughput': 2855.09}
+
+[INFO|2025-09-23 13:53:39] logging.py:143 >> {'loss': 0.0912, 'learning_rate': 8.7297e-06, 'epoch': 1.91, 'throughput': 2854.97}
+
+[INFO|2025-09-23 13:53:52] logging.py:143 >> {'loss': 0.0935, 'learning_rate': 8.4244e-06, 'epoch': 1.94, 'throughput': 2855.05}
+
+[INFO|2025-09-23 13:54:04] logging.py:143 >> {'loss': 0.0722, 'learning_rate': 8.1225e-06, 'epoch': 1.96, 'throughput': 2855.32}
+
+[INFO|2025-09-23 13:54:17] logging.py:143 >> {'loss': 0.1001, 'learning_rate': 7.8241e-06, 'epoch': 1.98, 'throughput': 2855.37}
+
+[INFO|2025-09-23 13:54:30] logging.py:143 >> {'loss': 0.0647, 'learning_rate': 7.5293e-06, 'epoch': 2.00, 'throughput': 2855.74}
+
+[INFO|2025-09-23 13:54:42] logging.py:143 >> {'loss': 0.0881, 'learning_rate': 7.2383e-06, 'epoch': 2.02, 'throughput': 2855.93}
+
+[INFO|2025-09-23 13:54:55] logging.py:143 >> {'loss': 0.0573, 'learning_rate': 6.9512e-06, 'epoch': 2.04, 'throughput': 2856.07}
+
+[INFO|2025-09-23 13:55:08] logging.py:143 >> {'loss': 0.0646, 'learning_rate': 6.6682e-06, 'epoch': 2.06, 'throughput': 2856.18}
+
+[INFO|2025-09-23 13:55:20] logging.py:143 >> {'loss': 0.0682, 'learning_rate': 6.3894e-06, 'epoch': 2.09, 'throughput': 2856.35}
+
+[INFO|2025-09-23 13:55:33] logging.py:143 >> {'loss': 0.0931, 'learning_rate': 6.1150e-06, 'epoch': 2.11, 'throughput': 2856.60}
+
+[INFO|2025-09-23 13:55:46] logging.py:143 >> {'loss': 0.0686, 'learning_rate': 5.8451e-06, 'epoch': 2.13, 'throughput': 2856.70}
+
+[INFO|2025-09-23 13:55:59] logging.py:143 >> {'loss': 0.0917, 'learning_rate': 5.5799e-06, 'epoch': 2.15, 'throughput': 2856.96}
+
+[INFO|2025-09-23 13:55:59] trainer.py:3993 >> Saving model checkpoint to saves/Mistral-7B-Instruct-v0.3/lora/mistral-treinado/checkpoint-1000
+
+[INFO|2025-09-23 13:55:59] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unsloth--mistral-7b-instruct-v0.3-bnb-4bit/snapshots/d5f623888f1415cf89b5c208d09cb620694618ee/config.json
+
+[INFO|2025-09-23 13:55:59] configuration_utils.py:770 >> Model config MistralConfig {
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 770,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "unsloth_version": "2024.9",
+  "use_cache": true,
+  "vocab_size": 32768
+}
+
+
+[INFO|2025-09-23 13:56:13] logging.py:143 >> {'loss': 0.0931, 'learning_rate': 5.3194e-06, 'epoch': 2.17, 'throughput': 2854.90}
+
+[INFO|2025-09-23 13:56:25] logging.py:143 >> {'loss': 0.0855, 'learning_rate': 5.0639e-06, 'epoch': 2.19, 'throughput': 2855.44}
+
+[INFO|2025-09-23 13:56:38] logging.py:143 >> {'loss': 0.0909, 'learning_rate': 4.8133e-06, 'epoch': 2.22, 'throughput': 2855.84}
+
+[INFO|2025-09-23 13:56:51] logging.py:143 >> {'loss': 0.0817, 'learning_rate': 4.5680e-06, 'epoch': 2.24, 'throughput': 2856.28}
+
+[INFO|2025-09-23 13:57:04] logging.py:143 >> {'loss': 0.0599, 'learning_rate': 4.3279e-06, 'epoch': 2.26, 'throughput': 2856.56}
+
+[INFO|2025-09-23 13:57:16] logging.py:143 >> {'loss': 0.0711, 'learning_rate': 4.0933e-06, 'epoch': 2.28, 'throughput': 2856.69}
+
+[INFO|2025-09-23 13:57:29] logging.py:143 >> {'loss': 0.0723, 'learning_rate': 3.8641e-06, 'epoch': 2.30, 'throughput': 2856.61}
+
+[INFO|2025-09-23 13:57:41] logging.py:143 >> {'loss': 0.0524, 'learning_rate': 3.6407e-06, 'epoch': 2.32, 'throughput': 2856.61}
+
+[INFO|2025-09-23 13:57:55] logging.py:143 >> {'loss': 0.0513, 'learning_rate': 3.4229e-06, 'epoch': 2.34, 'throughput': 2856.70}
+
+[INFO|2025-09-23 13:58:07] logging.py:143 >> {'loss': 0.0773, 'learning_rate': 3.2111e-06, 'epoch': 2.37, 'throughput': 2856.67}
+
+[INFO|2025-09-23 13:58:20] logging.py:143 >> {'loss': 0.0896, 'learning_rate': 3.0052e-06, 'epoch': 2.39, 'throughput': 2857.07}
+
+[INFO|2025-09-23 13:58:33] logging.py:143 >> {'loss': 0.0719, 'learning_rate': 2.8055e-06, 'epoch': 2.41, 'throughput': 2857.52}
+
+[INFO|2025-09-23 13:58:46] logging.py:143 >> {'loss': 0.0801, 'learning_rate': 2.6119e-06, 'epoch': 2.43, 'throughput': 2857.87}
+
+[INFO|2025-09-23 13:58:58] logging.py:143 >> {'loss': 0.0821, 'learning_rate': 2.4245e-06, 'epoch': 2.45, 'throughput': 2857.95}
+
+[INFO|2025-09-23 13:59:11] logging.py:143 >> {'loss': 0.0654, 'learning_rate': 2.2436e-06, 'epoch': 2.47, 'throughput': 2858.19}
+
+[INFO|2025-09-23 13:59:24] logging.py:143 >> {'loss': 0.0767, 'learning_rate': 2.0691e-06, 'epoch': 2.50, 'throughput': 2858.28}
+
+[INFO|2025-09-23 13:59:36] logging.py:143 >> {'loss': 0.0655, 'learning_rate': 1.9012e-06, 'epoch': 2.52, 'throughput': 2858.65}
+
+[INFO|2025-09-23 13:59:49] logging.py:143 >> {'loss': 0.0584, 'learning_rate': 1.7400e-06, 'epoch': 2.54, 'throughput': 2858.64}
+
+[INFO|2025-09-23 14:00:02] logging.py:143 >> {'loss': 0.0601, 'learning_rate': 1.5854e-06, 'epoch': 2.56, 'throughput': 2858.97}
+
+[INFO|2025-09-23 14:00:15] logging.py:143 >> {'loss': 0.0665, 'learning_rate': 1.4377e-06, 'epoch': 2.58, 'throughput': 2859.24}
+
+[INFO|2025-09-23 14:00:28] logging.py:143 >> {'loss': 0.0658, 'learning_rate': 1.2968e-06, 'epoch': 2.60, 'throughput': 2859.51}
+
+[INFO|2025-09-23 14:00:40] logging.py:143 >> {'loss': 0.0898, 'learning_rate': 1.1629e-06, 'epoch': 2.62, 'throughput': 2859.39}
+
+[INFO|2025-09-23 14:00:53] logging.py:143 >> {'loss': 0.0619, 'learning_rate': 1.0360e-06, 'epoch': 2.65, 'throughput': 2859.53}
+
+[INFO|2025-09-23 14:01:06] logging.py:143 >> {'loss': 0.0763, 'learning_rate': 9.1620e-07, 'epoch': 2.67, 'throughput': 2859.88}
+
+[INFO|2025-09-23 14:01:18] logging.py:143 >> {'loss': 0.0909, 'learning_rate': 8.0353e-07, 'epoch': 2.69, 'throughput': 2859.98}
+
+[INFO|2025-09-23 14:01:31] logging.py:143 >> {'loss': 0.0683, 'learning_rate': 6.9806e-07, 'epoch': 2.71, 'throughput': 2860.22}
+
+[INFO|2025-09-23 14:01:44] logging.py:143 >> {'loss': 0.0867, 'learning_rate': 5.9984e-07, 'epoch': 2.73, 'throughput': 2860.43}
+
+[INFO|2025-09-23 14:01:56] logging.py:143 >> {'loss': 0.0648, 'learning_rate': 5.0893e-07, 'epoch': 2.75, 'throughput': 2860.58}
+
+[INFO|2025-09-23 14:02:09] logging.py:143 >> {'loss': 0.0672, 'learning_rate': 4.2536e-07, 'epoch': 2.78, 'throughput': 2860.66}
+
+[INFO|2025-09-23 14:02:22] logging.py:143 >> {'loss': 0.0890, 'learning_rate': 3.4919e-07, 'epoch': 2.80, 'throughput': 2861.01}
+
+[INFO|2025-09-23 14:02:35] logging.py:143 >> {'loss': 0.0700, 'learning_rate': 2.8045e-07, 'epoch': 2.82, 'throughput': 2861.40}
+
+[INFO|2025-09-23 14:02:47] logging.py:143 >> {'loss': 0.0807, 'learning_rate': 2.1917e-07, 'epoch': 2.84, 'throughput': 2861.57}
+
+[INFO|2025-09-23 14:03:00] logging.py:143 >> {'loss': 0.0669, 'learning_rate': 1.6539e-07, 'epoch': 2.86, 'throughput': 2861.67}
+
+[INFO|2025-09-23 14:03:13] logging.py:143 >> {'loss': 0.0552, 'learning_rate': 1.1913e-07, 'epoch': 2.88, 'throughput': 2861.89}
+
+[INFO|2025-09-23 14:03:25] logging.py:143 >> {'loss': 0.0917, 'learning_rate': 8.0416e-08, 'epoch': 2.90, 'throughput': 2861.90}
+
+[INFO|2025-09-23 14:03:38] logging.py:143 >> {'loss': 0.0740, 'learning_rate': 4.9270e-08, 'epoch': 2.93, 'throughput': 2861.97}
+
+[INFO|2025-09-23 14:03:51] logging.py:143 >> {'loss': 0.0599, 'learning_rate': 2.5706e-08, 'epoch': 2.95, 'throughput': 2862.00}
+
+[INFO|2025-09-23 14:04:03] logging.py:143 >> {'loss': 0.0593, 'learning_rate': 9.7366e-09, 'epoch': 2.97, 'throughput': 2862.02}
+
+[INFO|2025-09-23 14:04:16] logging.py:143 >> {'loss': 0.0763, 'learning_rate': 1.3693e-09, 'epoch': 2.99, 'throughput': 2862.27}
+
+[INFO|2025-09-23 14:04:22] trainer.py:3993 >> Saving model checkpoint to saves/Mistral-7B-Instruct-v0.3/lora/mistral-treinado/checkpoint-1395
+
+[INFO|2025-09-23 14:04:23] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unsloth--mistral-7b-instruct-v0.3-bnb-4bit/snapshots/d5f623888f1415cf89b5c208d09cb620694618ee/config.json
+
+[INFO|2025-09-23 14:04:23] configuration_utils.py:770 >> Model config MistralConfig {
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 770,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "unsloth_version": "2024.9",
+  "use_cache": true,
+  "vocab_size": 32768
+}
+
+
+[INFO|2025-09-23 14:04:23] <string>:438 >> 
+
+Training completed. Do not forget to share your model on huggingface.co/models =)
+
+
+
+[INFO|2025-09-23 14:04:23] trainer.py:3993 >> Saving model checkpoint to saves/Mistral-7B-Instruct-v0.3/lora/mistral-treinado
+
+[INFO|2025-09-23 14:04:24] configuration_utils.py:698 >> loading configuration file config.json from cache at /root/.cache/huggingface/hub/models--unsloth--mistral-7b-instruct-v0.3-bnb-4bit/snapshots/d5f623888f1415cf89b5c208d09cb620694618ee/config.json
+
+[INFO|2025-09-23 14:04:24] configuration_utils.py:770 >> Model config MistralConfig {
+  "architectures": [
+    "MistralForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "head_dim": 128,
+  "hidden_act": "silu",
+  "hidden_size": 4096,
+  "initializer_range": 0.02,
+  "intermediate_size": 14336,
+  "max_position_embeddings": 32768,
+  "model_type": "mistral",
+  "num_attention_heads": 32,
+  "num_hidden_layers": 32,
+  "num_key_value_heads": 8,
+  "pad_token_id": 770,
+  "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
+    "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
+    "bnb_4bit_quant_type": "nf4",
+    "bnb_4bit_use_double_quant": true,
+    "llm_int8_enable_fp32_cpu_offload": false,
+    "llm_int8_has_fp16_weight": false,
+    "llm_int8_skip_modules": null,
+    "llm_int8_threshold": 6.0,
+    "load_in_4bit": true,
+    "load_in_8bit": false,
+    "quant_method": "bitsandbytes"
+  },
+  "rms_norm_eps": 1e-05,
+  "rope_theta": 1000000.0,
+  "sliding_window": null,
+  "tie_word_embeddings": false,
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.52.4",
+  "unsloth_version": "2024.9",
+  "use_cache": true,
+  "vocab_size": 32768
+}
+
+
+[WARNING|2025-09-23 14:04:24] logging.py:148 >> No metric eval_loss to plot.
+
+[WARNING|2025-09-23 14:04:24] logging.py:148 >> No metric eval_accuracy to plot.
+
+[INFO|2025-09-23 14:04:24] modelcard.py:450 >> Dropping the following result as it does not have all the necessary fields:
+{'task': {'name': 'Causal Language Modeling', 'type': 'text-generation'}}
+

special_tokens_map.json

@@ -0,0 +1,24 @@
+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": "</s>",
+  "unk_token": {
+    "content": "<unk>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37f00374dea48658ee8f5d0f21895b9bc55cb0103939607c8185bfd1c6ca1f89
+size 587404

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 3.0,
+    "num_input_tokens_seen": 5107072,
+    "total_flos": 2.186269295986606e+17,
+    "train_loss": 0.10808860245144068,
+    "train_runtime": 1785.3385,
+    "train_samples_per_second": 6.244,
+    "train_steps_per_second": 0.781
+}

trainer_log.jsonl

@@ -0,0 +1,140 @@
+{"current_steps": 10, "total_steps": 1395, "loss": 0.8869, "lr": 2.9996919063517692e-05, "epoch": 0.021528525296017224, "percentage": 0.72, "elapsed_time": "0:00:21", "remaining_time": "0:49:04", "throughput": 1715.83, "total_tokens": 36480}
+{"current_steps": 20, "total_steps": 1395, "loss": 0.199, "lr": 2.9986270537572827e-05, "epoch": 0.04305705059203445, "percentage": 1.43, "elapsed_time": "0:00:33", "remaining_time": "0:38:50", "throughput": 2143.92, "total_tokens": 72672}
+{"current_steps": 30, "total_steps": 1395, "loss": 0.1809, "lr": 2.9968021784983436e-05, "epoch": 0.06458557588805167, "percentage": 2.15, "elapsed_time": "0:00:47", "remaining_time": "0:35:47", "throughput": 2319.02, "total_tokens": 109472}
+{"current_steps": 40, "total_steps": 1395, "loss": 0.1766, "lr": 2.994218206053077e-05, "epoch": 0.0861141011840689, "percentage": 2.87, "elapsed_time": "0:01:00", "remaining_time": "0:33:52", "throughput": 2443.83, "total_tokens": 146656}
+{"current_steps": 50, "total_steps": 1395, "loss": 0.1784, "lr": 2.9908764468726846e-05, "epoch": 0.10764262648008611, "percentage": 3.58, "elapsed_time": "0:01:12", "remaining_time": "0:32:35", "throughput": 2519.74, "total_tokens": 183200}
+{"current_steps": 60, "total_steps": 1395, "loss": 0.165, "lr": 2.986778595716853e-05, "epoch": 0.12917115177610333, "percentage": 4.3, "elapsed_time": "0:01:25", "remaining_time": "0:31:43", "throughput": 2576.27, "total_tokens": 220352}
+{"current_steps": 70, "total_steps": 1395, "loss": 0.1884, "lr": 2.981926730794264e-05, "epoch": 0.15069967707212056, "percentage": 5.02, "elapsed_time": "0:01:38", "remaining_time": "0:31:00", "throughput": 2613.67, "total_tokens": 256864}
+{"current_steps": 80, "total_steps": 1395, "loss": 0.1668, "lr": 2.9763233127086394e-05, "epoch": 0.1722282023681378, "percentage": 5.73, "elapsed_time": "0:01:50", "remaining_time": "0:30:23", "throughput": 2644.76, "total_tokens": 293408}
+{"current_steps": 90, "total_steps": 1395, "loss": 0.1538, "lr": 2.9699711832108555e-05, "epoch": 0.193756727664155, "percentage": 6.45, "elapsed_time": "0:02:03", "remaining_time": "0:29:48", "throughput": 2667.87, "total_tokens": 329152}
+{"current_steps": 100, "total_steps": 1395, "loss": 0.1729, "lr": 2.9628735637577612e-05, "epoch": 0.21528525296017223, "percentage": 7.17, "elapsed_time": "0:02:16", "remaining_time": "0:29:21", "throughput": 2688.46, "total_tokens": 365632}
+{"current_steps": 110, "total_steps": 1395, "loss": 0.1346, "lr": 2.955034053878428e-05, "epoch": 0.23681377825618946, "percentage": 7.89, "elapsed_time": "0:02:28", "remaining_time": "0:28:55", "throughput": 2704.64, "total_tokens": 401888}
+{"current_steps": 120, "total_steps": 1395, "loss": 0.1622, "lr": 2.9464566293486653e-05, "epoch": 0.25834230355220666, "percentage": 8.6, "elapsed_time": "0:02:41", "remaining_time": "0:28:34", "throughput": 2719.24, "total_tokens": 438816}
+{"current_steps": 130, "total_steps": 1395, "loss": 0.1357, "lr": 2.9371456401747193e-05, "epoch": 0.2798708288482239, "percentage": 9.32, "elapsed_time": "0:02:54", "remaining_time": "0:28:14", "throughput": 2731.14, "total_tokens": 475648}
+{"current_steps": 140, "total_steps": 1395, "loss": 0.1537, "lr": 2.9271058083871864e-05, "epoch": 0.3013993541442411, "percentage": 10.04, "elapsed_time": "0:03:06", "remaining_time": "0:27:54", "throughput": 2741.53, "total_tokens": 512256}
+{"current_steps": 150, "total_steps": 1395, "loss": 0.1183, "lr": 2.916342225646254e-05, "epoch": 0.32292787944025836, "percentage": 10.75, "elapsed_time": "0:03:19", "remaining_time": "0:27:36", "throughput": 2750.14, "total_tokens": 548992}
+{"current_steps": 160, "total_steps": 1395, "loss": 0.1336, "lr": 2.904860350659487e-05, "epoch": 0.3444564047362756, "percentage": 11.47, "elapsed_time": "0:03:32", "remaining_time": "0:27:19", "throughput": 2757.92, "total_tokens": 585792}
+{"current_steps": 170, "total_steps": 1395, "loss": 0.1661, "lr": 2.8926660064134637e-05, "epoch": 0.36598493003229277, "percentage": 12.19, "elapsed_time": "0:03:45", "remaining_time": "0:27:01", "throughput": 2763.58, "total_tokens": 621952}
+{"current_steps": 180, "total_steps": 1395, "loss": 0.1405, "lr": 2.8797653772206726e-05, "epoch": 0.38751345532831, "percentage": 12.9, "elapsed_time": "0:03:57", "remaining_time": "0:26:44", "throughput": 2769.62, "total_tokens": 658240}
+{"current_steps": 190, "total_steps": 1395, "loss": 0.1456, "lr": 2.8661650055831645e-05, "epoch": 0.40904198062432723, "percentage": 13.62, "elapsed_time": "0:04:10", "remaining_time": "0:26:27", "throughput": 2776.17, "total_tokens": 694880}
+{"current_steps": 200, "total_steps": 1395, "loss": 0.1158, "lr": 2.8518717888745486e-05, "epoch": 0.43057050592034446, "percentage": 14.34, "elapsed_time": "0:04:23", "remaining_time": "0:26:12", "throughput": 2781.89, "total_tokens": 731968}
+{"current_steps": 210, "total_steps": 1395, "loss": 0.1293, "lr": 2.836892975842015e-05, "epoch": 0.4520990312163617, "percentage": 15.05, "elapsed_time": "0:04:35", "remaining_time": "0:25:55", "throughput": 2786.89, "total_tokens": 768416}
+{"current_steps": 220, "total_steps": 1395, "loss": 0.1456, "lr": 2.8212361629301655e-05, "epoch": 0.4736275565123789, "percentage": 15.77, "elapsed_time": "0:04:48", "remaining_time": "0:25:40", "throughput": 2791.68, "total_tokens": 805376}
+{"current_steps": 230, "total_steps": 1395, "loss": 0.1428, "lr": 2.8049092904285033e-05, "epoch": 0.4951560818083961, "percentage": 16.49, "elapsed_time": "0:05:01", "remaining_time": "0:25:24", "throughput": 2795.54, "total_tokens": 841472}
+{"current_steps": 240, "total_steps": 1395, "loss": 0.1334, "lr": 2.7879206384445526e-05, "epoch": 0.5166846071044133, "percentage": 17.2, "elapsed_time": "0:05:13", "remaining_time": "0:25:09", "throughput": 2800.11, "total_tokens": 878464}
+{"current_steps": 250, "total_steps": 1395, "loss": 0.1414, "lr": 2.7702788227046286e-05, "epoch": 0.5382131324004306, "percentage": 17.92, "elapsed_time": "0:05:26", "remaining_time": "0:24:53", "throughput": 2803.4, "total_tokens": 914432}
+{"current_steps": 260, "total_steps": 1395, "loss": 0.1098, "lr": 2.751992790184413e-05, "epoch": 0.5597416576964478, "percentage": 18.64, "elapsed_time": "0:05:39", "remaining_time": "0:24:40", "throughput": 2808.2, "total_tokens": 952320}
+{"current_steps": 270, "total_steps": 1395, "loss": 0.1643, "lr": 2.733071814571527e-05, "epoch": 0.581270182992465, "percentage": 19.35, "elapsed_time": "0:05:51", "remaining_time": "0:24:25", "throughput": 2810.83, "total_tokens": 988800}
+{"current_steps": 280, "total_steps": 1395, "loss": 0.1222, "lr": 2.7135254915624213e-05, "epoch": 0.6027987082884823, "percentage": 20.07, "elapsed_time": "0:06:04", "remaining_time": "0:24:11", "throughput": 2813.81, "total_tokens": 1025952}
+{"current_steps": 290, "total_steps": 1395, "loss": 0.1489, "lr": 2.693363733995954e-05, "epoch": 0.6243272335844995, "percentage": 20.79, "elapsed_time": "0:06:17", "remaining_time": "0:23:58", "throughput": 2817.18, "total_tokens": 1063360}
+{"current_steps": 300, "total_steps": 1395, "loss": 0.1207, "lr": 2.672596766826138e-05, "epoch": 0.6458557588805167, "percentage": 21.51, "elapsed_time": "0:06:30", "remaining_time": "0:23:43", "throughput": 2818.65, "total_tokens": 1099648}
+{"current_steps": 310, "total_steps": 1395, "loss": 0.1319, "lr": 2.6512351219365927e-05, "epoch": 0.667384284176534, "percentage": 22.22, "elapsed_time": "0:06:42", "remaining_time": "0:23:29", "throughput": 2819.84, "total_tokens": 1135712}
+{"current_steps": 320, "total_steps": 1395, "loss": 0.1453, "lr": 2.6292896327993467e-05, "epoch": 0.6889128094725512, "percentage": 22.94, "elapsed_time": "0:06:55", "remaining_time": "0:23:15", "throughput": 2822.06, "total_tokens": 1172384}
+{"current_steps": 330, "total_steps": 1395, "loss": 0.1202, "lr": 2.606771428980681e-05, "epoch": 0.7104413347685683, "percentage": 23.66, "elapsed_time": "0:07:08", "remaining_time": "0:23:02", "throughput": 2824.06, "total_tokens": 1209440}
+{"current_steps": 340, "total_steps": 1395, "loss": 0.1283, "lr": 2.5836919304968184e-05, "epoch": 0.7319698600645855, "percentage": 24.37, "elapsed_time": "0:07:20", "remaining_time": "0:22:48", "throughput": 2825.64, "total_tokens": 1245920}
+{"current_steps": 350, "total_steps": 1395, "loss": 0.1374, "lr": 2.5600628420223057e-05, "epoch": 0.7534983853606028, "percentage": 25.09, "elapsed_time": "0:07:34", "remaining_time": "0:22:35", "throughput": 2828.62, "total_tokens": 1284288}
+{"current_steps": 360, "total_steps": 1395, "loss": 0.1355, "lr": 2.535896146954037e-05, "epoch": 0.77502691065662, "percentage": 25.81, "elapsed_time": "0:07:46", "remaining_time": "0:22:21", "throughput": 2830.17, "total_tokens": 1320896}
+{"current_steps": 370, "total_steps": 1395, "loss": 0.131, "lr": 2.5112041013339212e-05, "epoch": 0.7965554359526372, "percentage": 26.52, "elapsed_time": "0:07:59", "remaining_time": "0:22:09", "throughput": 2832.7, "total_tokens": 1359264}
+{"current_steps": 380, "total_steps": 1395, "loss": 0.1034, "lr": 2.48599922763328e-05, "epoch": 0.8180839612486545, "percentage": 27.24, "elapsed_time": "0:08:12", "remaining_time": "0:21:55", "throughput": 2834.57, "total_tokens": 1396032}
+{"current_steps": 390, "total_steps": 1395, "loss": 0.1465, "lr": 2.4602943084021286e-05, "epoch": 0.8396124865446717, "percentage": 27.96, "elapsed_time": "0:08:25", "remaining_time": "0:21:41", "throughput": 2835.08, "total_tokens": 1432224}
+{"current_steps": 400, "total_steps": 1395, "loss": 0.1049, "lr": 2.434102379786553e-05, "epoch": 0.8611410118406889, "percentage": 28.67, "elapsed_time": "0:08:37", "remaining_time": "0:21:27", "throughput": 2835.78, "total_tokens": 1468224}
+{"current_steps": 410, "total_steps": 1395, "loss": 0.1024, "lr": 2.4074367249174864e-05, "epoch": 0.8826695371367062, "percentage": 29.39, "elapsed_time": "0:08:50", "remaining_time": "0:21:14", "throughput": 2837.37, "total_tokens": 1505088}
+{"current_steps": 420, "total_steps": 1395, "loss": 0.1155, "lr": 2.380310867174219e-05, "epoch": 0.9041980624327234, "percentage": 30.11, "elapsed_time": "0:09:03", "remaining_time": "0:21:01", "throughput": 2838.63, "total_tokens": 1542048}
+{"current_steps": 430, "total_steps": 1395, "loss": 0.117, "lr": 2.3527385633260724e-05, "epoch": 0.9257265877287406, "percentage": 30.82, "elapsed_time": "0:09:15", "remaining_time": "0:20:47", "throughput": 2840.25, "total_tokens": 1578496}
+{"current_steps": 440, "total_steps": 1395, "loss": 0.1023, "lr": 2.3247337965557097e-05, "epoch": 0.9472551130247578, "percentage": 31.54, "elapsed_time": "0:09:28", "remaining_time": "0:20:32", "throughput": 2840.35, "total_tokens": 1613440}
+{"current_steps": 450, "total_steps": 1395, "loss": 0.1046, "lr": 2.2963107693676192e-05, "epoch": 0.9687836383207751, "percentage": 32.26, "elapsed_time": "0:09:40", "remaining_time": "0:20:19", "throughput": 2840.69, "total_tokens": 1649760}
+{"current_steps": 460, "total_steps": 1395, "loss": 0.1106, "lr": 2.2674838963853738e-05, "epoch": 0.9903121636167922, "percentage": 32.97, "elapsed_time": "0:09:53", "remaining_time": "0:20:06", "throughput": 2841.62, "total_tokens": 1686016}
+{"current_steps": 470, "total_steps": 1395, "loss": 0.0859, "lr": 2.2382677970413097e-05, "epoch": 1.0107642626480087, "percentage": 33.69, "elapsed_time": "0:10:05", "remaining_time": "0:19:51", "throughput": 2842.72, "total_tokens": 1721152}
+{"current_steps": 480, "total_steps": 1395, "loss": 0.1089, "lr": 2.2086772881623455e-05, "epoch": 1.0322927879440258, "percentage": 34.41, "elapsed_time": "0:10:18", "remaining_time": "0:19:38", "throughput": 2843.73, "total_tokens": 1757568}
+{"current_steps": 490, "total_steps": 1395, "loss": 0.0852, "lr": 2.1787273764556827e-05, "epoch": 1.0538213132400431, "percentage": 35.13, "elapsed_time": "0:10:30", "remaining_time": "0:19:25", "throughput": 2845.12, "total_tokens": 1794848}
+{"current_steps": 500, "total_steps": 1395, "loss": 0.0988, "lr": 2.1484332508982167e-05, "epoch": 1.0753498385360603, "percentage": 35.84, "elapsed_time": "0:10:43", "remaining_time": "0:19:12", "throughput": 2846.63, "total_tokens": 1832256}
+{"current_steps": 510, "total_steps": 1395, "loss": 0.1001, "lr": 2.11781027503351e-05, "epoch": 1.0968783638320776, "percentage": 36.56, "elapsed_time": "0:10:56", "remaining_time": "0:18:58", "throughput": 2847.5, "total_tokens": 1868928}
+{"current_steps": 520, "total_steps": 1395, "loss": 0.1107, "lr": 2.0868739791802312e-05, "epoch": 1.1184068891280947, "percentage": 37.28, "elapsed_time": "0:11:09", "remaining_time": "0:18:45", "throughput": 2848.51, "total_tokens": 1905824}
+{"current_steps": 530, "total_steps": 1395, "loss": 0.0894, "lr": 2.0556400525560168e-05, "epoch": 1.1399354144241118, "percentage": 37.99, "elapsed_time": "0:11:21", "remaining_time": "0:18:32", "throughput": 2848.98, "total_tokens": 1941760}
+{"current_steps": 540, "total_steps": 1395, "loss": 0.1016, "lr": 2.0241243353207484e-05, "epoch": 1.1614639397201292, "percentage": 38.71, "elapsed_time": "0:11:34", "remaining_time": "0:18:19", "throughput": 2849.25, "total_tokens": 1978208}
+{"current_steps": 550, "total_steps": 1395, "loss": 0.0932, "lr": 1.9923428105432796e-05, "epoch": 1.1829924650161463, "percentage": 39.43, "elapsed_time": "0:11:47", "remaining_time": "0:18:06", "throughput": 2850.47, "total_tokens": 2015904}
+{"current_steps": 560, "total_steps": 1395, "loss": 0.1056, "lr": 1.9603115960956852e-05, "epoch": 1.2045209903121636, "percentage": 40.14, "elapsed_time": "0:11:59", "remaining_time": "0:17:53", "throughput": 2851.01, "total_tokens": 2052672}
+{"current_steps": 570, "total_steps": 1395, "loss": 0.1033, "lr": 1.9280469364791505e-05, "epoch": 1.2260495156081808, "percentage": 40.86, "elapsed_time": "0:12:12", "remaining_time": "0:17:40", "throughput": 2851.17, "total_tokens": 2088736}
+{"current_steps": 580, "total_steps": 1395, "loss": 0.1035, "lr": 1.8955651945856354e-05, "epoch": 1.247578040904198, "percentage": 41.58, "elapsed_time": "0:12:25", "remaining_time": "0:17:27", "throughput": 2851.42, "total_tokens": 2125600}
+{"current_steps": 590, "total_steps": 1395, "loss": 0.0811, "lr": 1.8628828433995013e-05, "epoch": 1.2691065662002152, "percentage": 42.29, "elapsed_time": "0:12:38", "remaining_time": "0:17:14", "throughput": 2851.58, "total_tokens": 2161792}
+{"current_steps": 600, "total_steps": 1395, "loss": 0.0888, "lr": 1.8300164576433035e-05, "epoch": 1.2906350914962326, "percentage": 43.01, "elapsed_time": "0:12:50", "remaining_time": "0:17:01", "throughput": 2851.53, "total_tokens": 2197920}
+{"current_steps": 610, "total_steps": 1395, "loss": 0.0968, "lr": 1.7969827053719862e-05, "epoch": 1.3121636167922497, "percentage": 43.73, "elapsed_time": "0:13:03", "remaining_time": "0:16:48", "throughput": 2852.2, "total_tokens": 2235168}
+{"current_steps": 620, "total_steps": 1395, "loss": 0.1311, "lr": 1.7637983395197465e-05, "epoch": 1.333692142088267, "percentage": 44.44, "elapsed_time": "0:13:16", "remaining_time": "0:16:35", "throughput": 2852.02, "total_tokens": 2270816}
+{"current_steps": 630, "total_steps": 1395, "loss": 0.1001, "lr": 1.7304801894038522e-05, "epoch": 1.3552206673842842, "percentage": 45.16, "elapsed_time": "0:13:29", "remaining_time": "0:16:22", "throughput": 2852.57, "total_tokens": 2308128}
+{"current_steps": 640, "total_steps": 1395, "loss": 0.0887, "lr": 1.6970451521897198e-05, "epoch": 1.3767491926803013, "percentage": 45.88, "elapsed_time": "0:13:41", "remaining_time": "0:16:09", "throughput": 2852.79, "total_tokens": 2344992}
+{"current_steps": 650, "total_steps": 1395, "loss": 0.1041, "lr": 1.663510184321587e-05, "epoch": 1.3982777179763186, "percentage": 46.59, "elapsed_time": "0:13:54", "remaining_time": "0:15:56", "throughput": 2852.95, "total_tokens": 2381056}
+{"current_steps": 660, "total_steps": 1395, "loss": 0.1022, "lr": 1.6298922929231194e-05, "epoch": 1.419806243272336, "percentage": 47.31, "elapsed_time": "0:14:07", "remaining_time": "0:15:43", "throughput": 2852.78, "total_tokens": 2417152}
+{"current_steps": 670, "total_steps": 1395, "loss": 0.0831, "lr": 1.5962085271723158e-05, "epoch": 1.441334768568353, "percentage": 48.03, "elapsed_time": "0:14:20", "remaining_time": "0:15:30", "throughput": 2853.25, "total_tokens": 2454048}
+{"current_steps": 680, "total_steps": 1395, "loss": 0.1055, "lr": 1.5624759696550874e-05, "epoch": 1.4628632938643702, "percentage": 48.75, "elapsed_time": "0:14:33", "remaining_time": "0:15:18", "throughput": 2853.64, "total_tokens": 2491552}
+{"current_steps": 690, "total_steps": 1395, "loss": 0.0812, "lr": 1.528711727701893e-05, "epoch": 1.4843918191603875, "percentage": 49.46, "elapsed_time": "0:14:45", "remaining_time": "0:15:05", "throughput": 2853.75, "total_tokens": 2528128}
+{"current_steps": 700, "total_steps": 1395, "loss": 0.1016, "lr": 1.4949329247118258e-05, "epoch": 1.5059203444564049, "percentage": 50.18, "elapsed_time": "0:14:58", "remaining_time": "0:14:52", "throughput": 2854.04, "total_tokens": 2564448}
+{"current_steps": 710, "total_steps": 1395, "loss": 0.0956, "lr": 1.4611566914685489e-05, "epoch": 1.527448869752422, "percentage": 50.9, "elapsed_time": "0:15:11", "remaining_time": "0:14:39", "throughput": 2854.16, "total_tokens": 2601056}
+{"current_steps": 720, "total_steps": 1395, "loss": 0.0936, "lr": 1.4274001574524894e-05, "epoch": 1.5489773950484391, "percentage": 51.61, "elapsed_time": "0:15:24", "remaining_time": "0:14:26", "throughput": 2854.51, "total_tokens": 2637888}
+{"current_steps": 730, "total_steps": 1395, "loss": 0.0728, "lr": 1.3936804421536906e-05, "epoch": 1.5705059203444565, "percentage": 52.33, "elapsed_time": "0:15:37", "remaining_time": "0:14:13", "throughput": 2852.88, "total_tokens": 2673312}
+{"current_steps": 740, "total_steps": 1395, "loss": 0.0839, "lr": 1.3600146463897327e-05, "epoch": 1.5920344456404736, "percentage": 53.05, "elapsed_time": "0:15:49", "remaining_time": "0:14:00", "throughput": 2852.96, "total_tokens": 2709696}
+{"current_steps": 750, "total_steps": 1395, "loss": 0.1308, "lr": 1.3264198436331193e-05, "epoch": 1.6135629709364907, "percentage": 53.76, "elapsed_time": "0:16:02", "remaining_time": "0:13:47", "throughput": 2853.61, "total_tokens": 2746528}
+{"current_steps": 760, "total_steps": 1395, "loss": 0.0927, "lr": 1.2929130713525396e-05, "epoch": 1.635091496232508, "percentage": 54.48, "elapsed_time": "0:16:15", "remaining_time": "0:13:34", "throughput": 2853.76, "total_tokens": 2782656}
+{"current_steps": 770, "total_steps": 1395, "loss": 0.1135, "lr": 1.2595113223723788e-05, "epoch": 1.6566200215285254, "percentage": 55.2, "elapsed_time": "0:16:27", "remaining_time": "0:13:21", "throughput": 2854.08, "total_tokens": 2819744}
+{"current_steps": 780, "total_steps": 1395, "loss": 0.0813, "lr": 1.2262315362548825e-05, "epoch": 1.6781485468245425, "percentage": 55.91, "elapsed_time": "0:16:41", "remaining_time": "0:13:09", "throughput": 2854.35, "total_tokens": 2857216}
+{"current_steps": 790, "total_steps": 1395, "loss": 0.0933, "lr": 1.1930905907093196e-05, "epoch": 1.6996770721205596, "percentage": 56.63, "elapsed_time": "0:16:53", "remaining_time": "0:12:56", "throughput": 2853.72, "total_tokens": 2892000}
+{"current_steps": 800, "total_steps": 1395, "loss": 0.0881, "lr": 1.1601052930325236e-05, "epoch": 1.721205597416577, "percentage": 57.35, "elapsed_time": "0:17:06", "remaining_time": "0:12:43", "throughput": 2853.73, "total_tokens": 2928832}
+{"current_steps": 810, "total_steps": 1395, "loss": 0.079, "lr": 1.127292371585141e-05, "epoch": 1.7427341227125943, "percentage": 58.06, "elapsed_time": "0:17:19", "remaining_time": "0:12:30", "throughput": 2853.98, "total_tokens": 2965632}
+{"current_steps": 820, "total_steps": 1395, "loss": 0.1132, "lr": 1.0946684673079074e-05, "epoch": 1.7642626480086114, "percentage": 58.78, "elapsed_time": "0:17:31", "remaining_time": "0:12:17", "throughput": 2854.14, "total_tokens": 3002496}
+{"current_steps": 830, "total_steps": 1395, "loss": 0.0793, "lr": 1.0622501252822628e-05, "epoch": 1.7857911733046286, "percentage": 59.5, "elapsed_time": "0:17:44", "remaining_time": "0:12:04", "throughput": 2854.34, "total_tokens": 3038944}
+{"current_steps": 840, "total_steps": 1395, "loss": 0.0918, "lr": 1.0300537863395823e-05, "epoch": 1.807319698600646, "percentage": 60.22, "elapsed_time": "0:17:57", "remaining_time": "0:11:51", "throughput": 2854.57, "total_tokens": 3076032}
+{"current_steps": 850, "total_steps": 1395, "loss": 0.093, "lr": 9.98095778723271e-06, "epoch": 1.8288482238966632, "percentage": 60.93, "elapsed_time": "0:18:10", "remaining_time": "0:11:39", "throughput": 2854.67, "total_tokens": 3112448}
+{"current_steps": 860, "total_steps": 1395, "loss": 0.0973, "lr": 9.663923098079656e-06, "epoch": 1.8503767491926801, "percentage": 61.65, "elapsed_time": "0:18:23", "remaining_time": "0:11:26", "throughput": 2854.95, "total_tokens": 3149376}
+{"current_steps": 870, "total_steps": 1395, "loss": 0.0995, "lr": 9.349594578800259e-06, "epoch": 1.8719052744886975, "percentage": 62.37, "elapsed_time": "0:18:36", "remaining_time": "0:11:13", "throughput": 2854.97, "total_tokens": 3186304}
+{"current_steps": 880, "total_steps": 1395, "loss": 0.1238, "lr": 9.038131639834977e-06, "epoch": 1.8934337997847148, "percentage": 63.08, "elapsed_time": "0:18:48", "remaining_time": "0:11:00", "throughput": 2855.09, "total_tokens": 3222752}
+{"current_steps": 890, "total_steps": 1395, "loss": 0.0912, "lr": 8.7296922383568e-06, "epoch": 1.914962325080732, "percentage": 63.8, "elapsed_time": "0:19:01", "remaining_time": "0:10:47", "throughput": 2854.97, "total_tokens": 3258976}
+{"current_steps": 900, "total_steps": 1395, "loss": 0.0935, "lr": 8.424432798163838e-06, "epoch": 1.936490850376749, "percentage": 64.52, "elapsed_time": "0:19:14", "remaining_time": "0:10:34", "throughput": 2855.05, "total_tokens": 3295040}
+{"current_steps": 910, "total_steps": 1395, "loss": 0.0722, "lr": 8.122508130349711e-06, "epoch": 1.9580193756727664, "percentage": 65.23, "elapsed_time": "0:19:26", "remaining_time": "0:10:21", "throughput": 2855.32, "total_tokens": 3331392}
+{"current_steps": 920, "total_steps": 1395, "loss": 0.1001, "lr": 7.82407135479164e-06, "epoch": 1.9795479009687837, "percentage": 65.95, "elapsed_time": "0:19:39", "remaining_time": "0:10:09", "throughput": 2855.37, "total_tokens": 3368032}
+{"current_steps": 930, "total_steps": 1395, "loss": 0.0647, "lr": 7.529273822496379e-06, "epoch": 2.0, "percentage": 66.67, "elapsed_time": "0:19:51", "remaining_time": "0:09:55", "throughput": 2855.74, "total_tokens": 3403584}
+{"current_steps": 940, "total_steps": 1395, "loss": 0.0881, "lr": 7.238265038843143e-06, "epoch": 2.0215285252960173, "percentage": 67.38, "elapsed_time": "0:20:04", "remaining_time": "0:09:43", "throughput": 2855.93, "total_tokens": 3440000}
+{"current_steps": 950, "total_steps": 1395, "loss": 0.0573, "lr": 6.951192587762555e-06, "epoch": 2.0430570505920342, "percentage": 68.1, "elapsed_time": "0:20:17", "remaining_time": "0:09:30", "throughput": 2856.07, "total_tokens": 3476640}
+{"current_steps": 960, "total_steps": 1395, "loss": 0.0646, "lr": 6.668202056890052e-06, "epoch": 2.0645855758880516, "percentage": 68.82, "elapsed_time": "0:20:29", "remaining_time": "0:09:17", "throughput": 2856.18, "total_tokens": 3512992}
+{"current_steps": 970, "total_steps": 1395, "loss": 0.0682, "lr": 6.389436963731707e-06, "epoch": 2.086114101184069, "percentage": 69.53, "elapsed_time": "0:20:42", "remaining_time": "0:09:04", "throughput": 2856.35, "total_tokens": 3549504}
+{"current_steps": 980, "total_steps": 1395, "loss": 0.0931, "lr": 6.115038682879904e-06, "epoch": 2.1076426264800863, "percentage": 70.25, "elapsed_time": "0:20:55", "remaining_time": "0:08:51", "throughput": 2856.6, "total_tokens": 3586592}
+{"current_steps": 990, "total_steps": 1395, "loss": 0.0686, "lr": 5.845146374315781e-06, "epoch": 2.129171151776103, "percentage": 70.97, "elapsed_time": "0:21:08", "remaining_time": "0:08:38", "throughput": 2856.7, "total_tokens": 3622944}
+{"current_steps": 1000, "total_steps": 1395, "loss": 0.0917, "lr": 5.579896912834807e-06, "epoch": 2.1506996770721205, "percentage": 71.68, "elapsed_time": "0:21:21", "remaining_time": "0:08:26", "throughput": 2856.96, "total_tokens": 3660096}
+{"current_steps": 1010, "total_steps": 1395, "loss": 0.0931, "lr": 5.319424818631293e-06, "epoch": 2.172228202368138, "percentage": 72.4, "elapsed_time": "0:21:34", "remaining_time": "0:08:13", "throughput": 2854.9, "total_tokens": 3697024}
+{"current_steps": 1020, "total_steps": 1395, "loss": 0.0855, "lr": 5.063862189076988e-06, "epoch": 2.193756727664155, "percentage": 73.12, "elapsed_time": "0:21:47", "remaining_time": "0:08:00", "throughput": 2855.44, "total_tokens": 3734240}
+{"current_steps": 1030, "total_steps": 1395, "loss": 0.0909, "lr": 4.813338631728471e-06, "epoch": 2.215285252960172, "percentage": 73.84, "elapsed_time": "0:22:00", "remaining_time": "0:07:47", "throughput": 2855.84, "total_tokens": 3771232}
+{"current_steps": 1040, "total_steps": 1395, "loss": 0.0817, "lr": 4.567981198597168e-06, "epoch": 2.2368137782561894, "percentage": 74.55, "elapsed_time": "0:22:13", "remaining_time": "0:07:35", "throughput": 2856.28, "total_tokens": 3808416}
+{"current_steps": 1050, "total_steps": 1395, "loss": 0.0599, "lr": 4.327914321715459e-06, "epoch": 2.2583423035522068, "percentage": 75.27, "elapsed_time": "0:22:25", "remaining_time": "0:07:22", "throughput": 2856.56, "total_tokens": 3844768}
+{"current_steps": 1060, "total_steps": 1395, "loss": 0.0711, "lr": 4.09325975003146e-06, "epoch": 2.2798708288482237, "percentage": 75.99, "elapsed_time": "0:22:38", "remaining_time": "0:07:09", "throughput": 2856.69, "total_tokens": 3881088}
+{"current_steps": 1070, "total_steps": 1395, "loss": 0.0723, "lr": 3.864136487664538e-06, "epoch": 2.301399354144241, "percentage": 76.7, "elapsed_time": "0:22:51", "remaining_time": "0:06:56", "throughput": 2856.61, "total_tokens": 3916928}
+{"current_steps": 1080, "total_steps": 1395, "loss": 0.0524, "lr": 3.640660733552863e-06, "epoch": 2.3229278794402584, "percentage": 77.42, "elapsed_time": "0:23:03", "remaining_time": "0:06:43", "throughput": 2856.61, "total_tokens": 3952928}
+{"current_steps": 1090, "total_steps": 1395, "loss": 0.0513, "lr": 3.4229458225235696e-06, "epoch": 2.3444564047362757, "percentage": 78.14, "elapsed_time": "0:23:16", "remaining_time": "0:06:30", "throughput": 2856.7, "total_tokens": 3990720}
+{"current_steps": 1100, "total_steps": 1395, "loss": 0.0773, "lr": 3.211102167815469e-06, "epoch": 2.3659849300322926, "percentage": 78.85, "elapsed_time": "0:23:29", "remaining_time": "0:06:18", "throughput": 2856.67, "total_tokens": 4026656}
+{"current_steps": 1110, "total_steps": 1395, "loss": 0.0896, "lr": 3.0052372050834025e-06, "epoch": 2.38751345532831, "percentage": 79.57, "elapsed_time": "0:23:42", "remaining_time": "0:06:05", "throughput": 2857.07, "total_tokens": 4064192}
+{"current_steps": 1120, "total_steps": 1395, "loss": 0.0719, "lr": 2.805455337912722e-06, "epoch": 2.4090419806243273, "percentage": 80.29, "elapsed_time": "0:23:55", "remaining_time": "0:05:52", "throughput": 2857.52, "total_tokens": 4101632}
+{"current_steps": 1130, "total_steps": 1395, "loss": 0.0801, "lr": 2.611857884871379e-06, "epoch": 2.4305705059203446, "percentage": 81.0, "elapsed_time": "0:24:08", "remaining_time": "0:05:39", "throughput": 2857.87, "total_tokens": 4138208}
+{"current_steps": 1140, "total_steps": 1395, "loss": 0.0821, "lr": 2.424543028126661e-06, "epoch": 2.4520990312163615, "percentage": 81.72, "elapsed_time": "0:24:20", "remaining_time": "0:05:26", "throughput": 2857.95, "total_tokens": 4174592}
+{"current_steps": 1150, "total_steps": 1395, "loss": 0.0654, "lr": 2.243605763652489e-06, "epoch": 2.473627556512379, "percentage": 82.44, "elapsed_time": "0:24:33", "remaining_time": "0:05:13", "throughput": 2858.19, "total_tokens": 4211616}
+{"current_steps": 1160, "total_steps": 1395, "loss": 0.0767, "lr": 2.0691378530525933e-06, "epoch": 2.495156081808396, "percentage": 83.15, "elapsed_time": "0:24:46", "remaining_time": "0:05:01", "throughput": 2858.28, "total_tokens": 4247744}
+{"current_steps": 1170, "total_steps": 1395, "loss": 0.0655, "lr": 1.9012277770239829e-06, "epoch": 2.5166846071044136, "percentage": 83.87, "elapsed_time": "0:24:58", "remaining_time": "0:04:48", "throughput": 2858.65, "total_tokens": 4284224}
+{"current_steps": 1180, "total_steps": 1395, "loss": 0.0584, "lr": 1.7399606904843224e-06, "epoch": 2.5382131324004304, "percentage": 84.59, "elapsed_time": "0:25:11", "remaining_time": "0:04:35", "throughput": 2858.64, "total_tokens": 4320672}
+{"current_steps": 1190, "total_steps": 1395, "loss": 0.0601, "lr": 1.5854183793859577e-06, "epoch": 2.559741657696448, "percentage": 85.3, "elapsed_time": "0:25:24", "remaining_time": "0:04:22", "throughput": 2858.97, "total_tokens": 4358080}
+{"current_steps": 1200, "total_steps": 1395, "loss": 0.0665, "lr": 1.437679219238498e-06, "epoch": 2.581270182992465, "percentage": 86.02, "elapsed_time": "0:25:37", "remaining_time": "0:04:09", "throughput": 2859.24, "total_tokens": 4394752}
+{"current_steps": 1210, "total_steps": 1395, "loss": 0.0658, "lr": 1.2968181353609854e-06, "epoch": 2.602798708288482, "percentage": 86.74, "elapsed_time": "0:25:49", "remaining_time": "0:03:56", "throughput": 2859.51, "total_tokens": 4431808}
+{"current_steps": 1220, "total_steps": 1395, "loss": 0.0898, "lr": 1.1629065648838495e-06, "epoch": 2.6243272335844994, "percentage": 87.46, "elapsed_time": "0:26:02", "remaining_time": "0:03:44", "throughput": 2859.39, "total_tokens": 4467328}
+{"current_steps": 1230, "total_steps": 1395, "loss": 0.0619, "lr": 1.0360124205198102e-06, "epoch": 2.6458557588805167, "percentage": 88.17, "elapsed_time": "0:26:15", "remaining_time": "0:03:31", "throughput": 2859.53, "total_tokens": 4503808}
+{"current_steps": 1240, "total_steps": 1395, "loss": 0.0763, "lr": 9.162000561222672e-07, "epoch": 2.667384284176534, "percentage": 88.89, "elapsed_time": "0:26:28", "remaining_time": "0:03:18", "throughput": 2859.88, "total_tokens": 4541536}
+{"current_steps": 1250, "total_steps": 1395, "loss": 0.0909, "lr": 8.035302340484646e-07, "epoch": 2.6889128094725514, "percentage": 89.61, "elapsed_time": "0:26:40", "remaining_time": "0:03:05", "throughput": 2859.98, "total_tokens": 4577664}
+{"current_steps": 1260, "total_steps": 1395, "loss": 0.0683, "lr": 6.980600943441273e-07, "epoch": 2.7104413347685683, "percentage": 90.32, "elapsed_time": "0:26:53", "remaining_time": "0:02:52", "throughput": 2860.22, "total_tokens": 4614176}
+{"current_steps": 1270, "total_steps": 1395, "loss": 0.0867, "lr": 5.998431257651132e-07, "epoch": 2.7319698600645856, "percentage": 91.04, "elapsed_time": "0:27:05", "remaining_time": "0:02:40", "throughput": 2860.43, "total_tokens": 4650848}
+{"current_steps": 1280, "total_steps": 1395, "loss": 0.0648, "lr": 5.089291386508055e-07, "epoch": 2.7534983853606025, "percentage": 91.76, "elapsed_time": "0:27:18", "remaining_time": "0:02:27", "throughput": 2860.58, "total_tokens": 4687456}
+{"current_steps": 1290, "total_steps": 1395, "loss": 0.0672, "lr": 4.2536423966299256e-07, "epoch": 2.77502691065662, "percentage": 92.47, "elapsed_time": "0:27:31", "remaining_time": "0:02:14", "throughput": 2860.66, "total_tokens": 4723840}
+{"current_steps": 1300, "total_steps": 1395, "loss": 0.089, "lr": 3.491908084030776e-07, "epoch": 2.7965554359526372, "percentage": 93.19, "elapsed_time": "0:27:44", "remaining_time": "0:02:01", "throughput": 2861.01, "total_tokens": 4761056}
+{"current_steps": 1310, "total_steps": 1395, "loss": 0.07, "lr": 2.804474759194176e-07, "epoch": 2.8180839612486546, "percentage": 93.91, "elapsed_time": "0:27:56", "remaining_time": "0:01:48", "throughput": 2861.4, "total_tokens": 4798240}
+{"current_steps": 1320, "total_steps": 1395, "loss": 0.0807, "lr": 2.1916910511575627e-07, "epoch": 2.839612486544672, "percentage": 94.62, "elapsed_time": "0:28:09", "remaining_time": "0:01:35", "throughput": 2861.57, "total_tokens": 4834560}
+{"current_steps": 1330, "total_steps": 1395, "loss": 0.0669, "lr": 1.6538677307062933e-07, "epoch": 2.861141011840689, "percentage": 95.34, "elapsed_time": "0:28:22", "remaining_time": "0:01:23", "throughput": 2861.67, "total_tokens": 4871328}
+{"current_steps": 1340, "total_steps": 1395, "loss": 0.0552, "lr": 1.1912775527676278e-07, "epoch": 2.882669537136706, "percentage": 96.06, "elapsed_time": "0:28:34", "remaining_time": "0:01:10", "throughput": 2861.89, "total_tokens": 4907744}
+{"current_steps": 1350, "total_steps": 1395, "loss": 0.0917, "lr": 8.04155118084099e-08, "epoch": 2.9041980624327235, "percentage": 96.77, "elapsed_time": "0:28:47", "remaining_time": "0:00:57", "throughput": 2861.9, "total_tokens": 4943712}
+{"current_steps": 1360, "total_steps": 1395, "loss": 0.074, "lr": 4.9269675423671556e-08, "epoch": 2.9257265877287404, "percentage": 97.49, "elapsed_time": "0:29:00", "remaining_time": "0:00:44", "throughput": 2861.97, "total_tokens": 4980480}
+{"current_steps": 1370, "total_steps": 1395, "loss": 0.0599, "lr": 2.570604160781753e-08, "epoch": 2.9472551130247577, "percentage": 98.21, "elapsed_time": "0:29:12", "remaining_time": "0:00:31", "throughput": 2862.0, "total_tokens": 5016800}
+{"current_steps": 1380, "total_steps": 1395, "loss": 0.0593, "lr": 9.736560562675912e-09, "epoch": 2.968783638320775, "percentage": 98.92, "elapsed_time": "0:29:25", "remaining_time": "0:00:19", "throughput": 2862.02, "total_tokens": 5053440}
+{"current_steps": 1390, "total_steps": 1395, "loss": 0.0763, "lr": 1.3693311461326908e-09, "epoch": 2.9903121636167924, "percentage": 99.64, "elapsed_time": "0:29:38", "remaining_time": "0:00:06", "throughput": 2862.27, "total_tokens": 5090592}
+{"current_steps": 1395, "total_steps": 1395, "epoch": 3.0, "percentage": 100.0, "elapsed_time": "0:29:45", "remaining_time": "0:00:00", "throughput": 2860.57, "total_tokens": 5107072}

trainer_state.json

@@ -0,0 +1,1434 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 1395,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.021528525296017224,
+      "grad_norm": 0.000117536197649315,
+      "learning_rate": 2.9996919063517692e-05,
+      "loss": 0.8869,
+      "num_input_tokens_seen": 36480,
+      "step": 10,
+      "train_runtime": 21.2665,
+      "train_tokens_per_second": 1715.375
+    },
+    {
+      "epoch": 0.04305705059203445,
+      "grad_norm": 0.00022027654631529003,
+      "learning_rate": 2.9986270537572827e-05,
+      "loss": 0.199,
+      "num_input_tokens_seen": 72672,
+      "step": 20,
+      "train_runtime": 33.9024,
+      "train_tokens_per_second": 2143.565
+    },
+    {
+      "epoch": 0.06458557588805167,
+      "grad_norm": 5.31423429492861e-05,
+      "learning_rate": 2.9968021784983436e-05,
+      "loss": 0.1809,
+      "num_input_tokens_seen": 109472,
+      "step": 30,
+      "train_runtime": 47.2118,
+      "train_tokens_per_second": 2318.741
+    },
+    {
+      "epoch": 0.0861141011840689,
+      "grad_norm": 5.581648292718455e-05,
+      "learning_rate": 2.994218206053077e-05,
+      "loss": 0.1766,
+      "num_input_tokens_seen": 146656,
+      "step": 40,
+      "train_runtime": 60.0165,
+      "train_tokens_per_second": 2443.597
+    },
+    {
+      "epoch": 0.10764262648008611,
+      "grad_norm": 4.6867833589203656e-05,
+      "learning_rate": 2.9908764468726846e-05,
+      "loss": 0.1784,
+      "num_input_tokens_seen": 183200,
+      "step": 50,
+      "train_runtime": 72.7117,
+      "train_tokens_per_second": 2519.54
+    },
+    {
+      "epoch": 0.12917115177610333,
+      "grad_norm": 7.248102338053286e-05,
+      "learning_rate": 2.986778595716853e-05,
+      "loss": 0.165,
+      "num_input_tokens_seen": 220352,
+      "step": 60,
+      "train_runtime": 85.5372,
+      "train_tokens_per_second": 2576.096
+    },
+    {
+      "epoch": 0.15069967707212056,
+      "grad_norm": 6.31138464086689e-05,
+      "learning_rate": 2.981926730794264e-05,
+      "loss": 0.1884,
+      "num_input_tokens_seen": 256864,
+      "step": 70,
+      "train_runtime": 98.2829,
+      "train_tokens_per_second": 2613.516
+    },
+    {
+      "epoch": 0.1722282023681378,
+      "grad_norm": 8.531038474757224e-05,
+      "learning_rate": 2.9763233127086394e-05,
+      "loss": 0.1668,
+      "num_input_tokens_seen": 293408,
+      "step": 80,
+      "train_runtime": 110.9453,
+      "train_tokens_per_second": 2644.619
+    },
+    {
+      "epoch": 0.193756727664155,
+      "grad_norm": 3.399947308935225e-05,
+      "learning_rate": 2.9699711832108555e-05,
+      "loss": 0.1538,
+      "num_input_tokens_seen": 329152,
+      "step": 90,
+      "train_runtime": 123.3823,
+      "train_tokens_per_second": 2667.742
+    },
+    {
+      "epoch": 0.21528525296017223,
+      "grad_norm": 6.389995542122051e-05,
+      "learning_rate": 2.9628735637577612e-05,
+      "loss": 0.1729,
+      "num_input_tokens_seen": 365632,
+      "step": 100,
+      "train_runtime": 136.0062,
+      "train_tokens_per_second": 2688.348
+    },
+    {
+      "epoch": 0.23681377825618946,
+      "grad_norm": 8.456028444925323e-05,
+      "learning_rate": 2.955034053878428e-05,
+      "loss": 0.1346,
+      "num_input_tokens_seen": 401888,
+      "step": 110,
+      "train_runtime": 148.598,
+      "train_tokens_per_second": 2704.531
+    },
+    {
+      "epoch": 0.25834230355220666,
+      "grad_norm": 6.639358616666868e-05,
+      "learning_rate": 2.9464566293486653e-05,
+      "loss": 0.1622,
+      "num_input_tokens_seen": 438816,
+      "step": 120,
+      "train_runtime": 161.3801,
+      "train_tokens_per_second": 2719.146
+    },
+    {
+      "epoch": 0.2798708288482239,
+      "grad_norm": 6.005055911373347e-05,
+      "learning_rate": 2.9371456401747193e-05,
+      "loss": 0.1357,
+      "num_input_tokens_seen": 475648,
+      "step": 130,
+      "train_runtime": 174.163,
+      "train_tokens_per_second": 2731.051
+    },
+    {
+      "epoch": 0.3013993541442411,
+      "grad_norm": 8.286784577649087e-05,
+      "learning_rate": 2.9271058083871864e-05,
+      "loss": 0.1537,
+      "num_input_tokens_seen": 512256,
+      "step": 140,
+      "train_runtime": 186.8563,
+      "train_tokens_per_second": 2741.444
+    },
+    {
+      "epoch": 0.32292787944025836,
+      "grad_norm": 4.636685480363667e-05,
+      "learning_rate": 2.916342225646254e-05,
+      "loss": 0.1183,
+      "num_input_tokens_seen": 548992,
+      "step": 150,
+      "train_runtime": 199.6287,
+      "train_tokens_per_second": 2750.065
+    },
+    {
+      "epoch": 0.3444564047362756,
+      "grad_norm": 4.919446655549109e-05,
+      "learning_rate": 2.904860350659487e-05,
+      "loss": 0.1336,
+      "num_input_tokens_seen": 585792,
+      "step": 160,
+      "train_runtime": 212.4096,
+      "train_tokens_per_second": 2757.841
+    },
+    {
+      "epoch": 0.36598493003229277,
+      "grad_norm": 8.222900942200795e-05,
+      "learning_rate": 2.8926660064134637e-05,
+      "loss": 0.1661,
+      "num_input_tokens_seen": 621952,
+      "step": 170,
+      "train_runtime": 225.0588,
+      "train_tokens_per_second": 2763.509
+    },
+    {
+      "epoch": 0.38751345532831,
+      "grad_norm": 3.515497155603953e-05,
+      "learning_rate": 2.8797653772206726e-05,
+      "loss": 0.1405,
+      "num_input_tokens_seen": 658240,
+      "step": 180,
+      "train_runtime": 237.6697,
+      "train_tokens_per_second": 2769.557
+    },
+    {
+      "epoch": 0.40904198062432723,
+      "grad_norm": 4.577370418701321e-05,
+      "learning_rate": 2.8661650055831645e-05,
+      "loss": 0.1456,
+      "num_input_tokens_seen": 694880,
+      "step": 190,
+      "train_runtime": 250.3071,
+      "train_tokens_per_second": 2776.11
+    },
+    {
+      "epoch": 0.43057050592034446,
+      "grad_norm": 5.2699797379318625e-05,
+      "learning_rate": 2.8518717888745486e-05,
+      "loss": 0.1158,
+      "num_input_tokens_seen": 731968,
+      "step": 200,
+      "train_runtime": 263.1248,
+      "train_tokens_per_second": 2781.828
+    },
+    {
+      "epoch": 0.4520990312163617,
+      "grad_norm": 4.86937933601439e-05,
+      "learning_rate": 2.836892975842015e-05,
+      "loss": 0.1293,
+      "num_input_tokens_seen": 768416,
+      "step": 210,
+      "train_runtime": 275.7314,
+      "train_tokens_per_second": 2786.828
+    },
+    {
+      "epoch": 0.4736275565123789,
+      "grad_norm": 5.121620051795617e-05,
+      "learning_rate": 2.8212361629301655e-05,
+      "loss": 0.1456,
+      "num_input_tokens_seen": 805376,
+      "step": 220,
+      "train_runtime": 288.4976,
+      "train_tokens_per_second": 2791.621
+    },
+    {
+      "epoch": 0.4951560818083961,
+      "grad_norm": 3.388872210052796e-05,
+      "learning_rate": 2.8049092904285033e-05,
+      "loss": 0.1428,
+      "num_input_tokens_seen": 841472,
+      "step": 230,
+      "train_runtime": 301.0113,
+      "train_tokens_per_second": 2795.483
+    },
+    {
+      "epoch": 0.5166846071044133,
+      "grad_norm": 5.026933649787679e-05,
+      "learning_rate": 2.7879206384445526e-05,
+      "loss": 0.1334,
+      "num_input_tokens_seen": 878464,
+      "step": 240,
+      "train_runtime": 313.7308,
+      "train_tokens_per_second": 2800.057
+    },
+    {
+      "epoch": 0.5382131324004306,
+      "grad_norm": 2.826541276590433e-05,
+      "learning_rate": 2.7702788227046286e-05,
+      "loss": 0.1414,
+      "num_input_tokens_seen": 914432,
+      "step": 250,
+      "train_runtime": 326.1923,
+      "train_tokens_per_second": 2803.352
+    },
+    {
+      "epoch": 0.5597416576964478,
+      "grad_norm": 5.1211100071668625e-05,
+      "learning_rate": 2.751992790184413e-05,
+      "loss": 0.1098,
+      "num_input_tokens_seen": 952320,
+      "step": 260,
+      "train_runtime": 339.127,
+      "train_tokens_per_second": 2808.151
+    },
+    {
+      "epoch": 0.581270182992465,
+      "grad_norm": 4.860994886257686e-05,
+      "learning_rate": 2.733071814571527e-05,
+      "loss": 0.1643,
+      "num_input_tokens_seen": 988800,
+      "step": 270,
+      "train_runtime": 351.788,
+      "train_tokens_per_second": 2810.784
+    },
+    {
+      "epoch": 0.6027987082884823,
+      "grad_norm": 3.192482836311683e-05,
+      "learning_rate": 2.7135254915624213e-05,
+      "loss": 0.1222,
+      "num_input_tokens_seen": 1025952,
+      "step": 280,
+      "train_runtime": 364.6184,
+      "train_tokens_per_second": 2813.769
+    },
+    {
+      "epoch": 0.6243272335844995,
+      "grad_norm": 6.410479545593262e-05,
+      "learning_rate": 2.693363733995954e-05,
+      "loss": 0.1489,
+      "num_input_tokens_seen": 1063360,
+      "step": 290,
+      "train_runtime": 377.4609,
+      "train_tokens_per_second": 2817.139
+    },
+    {
+      "epoch": 0.6458557588805167,
+      "grad_norm": 7.877861935412511e-05,
+      "learning_rate": 2.672596766826138e-05,
+      "loss": 0.1207,
+      "num_input_tokens_seen": 1099648,
+      "step": 300,
+      "train_runtime": 390.138,
+      "train_tokens_per_second": 2818.613
+    },
+    {
+      "epoch": 0.667384284176534,
+      "grad_norm": 5.999449058435857e-05,
+      "learning_rate": 2.6512351219365927e-05,
+      "loss": 0.1319,
+      "num_input_tokens_seen": 1135712,
+      "step": 310,
+      "train_runtime": 402.7639,
+      "train_tokens_per_second": 2819.796
+    },
+    {
+      "epoch": 0.6889128094725512,
+      "grad_norm": 2.934955591626931e-05,
+      "learning_rate": 2.6292896327993467e-05,
+      "loss": 0.1453,
+      "num_input_tokens_seen": 1172384,
+      "step": 320,
+      "train_runtime": 415.4411,
+      "train_tokens_per_second": 2822.022
+    },
+    {
+      "epoch": 0.7104413347685683,
+      "grad_norm": 4.606367292581126e-05,
+      "learning_rate": 2.606771428980681e-05,
+      "loss": 0.1202,
+      "num_input_tokens_seen": 1209440,
+      "step": 330,
+      "train_runtime": 428.2678,
+      "train_tokens_per_second": 2824.027
+    },
+    {
+      "epoch": 0.7319698600645855,
+      "grad_norm": 3.709704469656572e-05,
+      "learning_rate": 2.5836919304968184e-05,
+      "loss": 0.1283,
+      "num_input_tokens_seen": 1245920,
+      "step": 340,
+      "train_runtime": 440.939,
+      "train_tokens_per_second": 2825.606
+    },
+    {
+      "epoch": 0.7534983853606028,
+      "grad_norm": 4.9612910515861586e-05,
+      "learning_rate": 2.5600628420223057e-05,
+      "loss": 0.1374,
+      "num_input_tokens_seen": 1284288,
+      "step": 350,
+      "train_runtime": 454.0391,
+      "train_tokens_per_second": 2828.585
+    },
+    {
+      "epoch": 0.77502691065662,
+      "grad_norm": 3.349883991177194e-05,
+      "learning_rate": 2.535896146954037e-05,
+      "loss": 0.1355,
+      "num_input_tokens_seen": 1320896,
+      "step": 360,
+      "train_runtime": 466.7256,
+      "train_tokens_per_second": 2830.134
+    },
+    {
+      "epoch": 0.7965554359526372,
+      "grad_norm": 3.364904478075914e-05,
+      "learning_rate": 2.5112041013339212e-05,
+      "loss": 0.131,
+      "num_input_tokens_seen": 1359264,
+      "step": 370,
+      "train_runtime": 479.8532,
+      "train_tokens_per_second": 2832.667
+    },
+    {
+      "epoch": 0.8180839612486545,
+      "grad_norm": 3.689135701279156e-05,
+      "learning_rate": 2.48599922763328e-05,
+      "loss": 0.1034,
+      "num_input_tokens_seen": 1396032,
+      "step": 380,
+      "train_runtime": 492.5082,
+      "train_tokens_per_second": 2834.535
+    },
+    {
+      "epoch": 0.8396124865446717,
+      "grad_norm": 3.208614361938089e-05,
+      "learning_rate": 2.4602943084021286e-05,
+      "loss": 0.1465,
+      "num_input_tokens_seen": 1432224,
+      "step": 390,
+      "train_runtime": 505.1852,
+      "train_tokens_per_second": 2835.047
+    },
+    {
+      "epoch": 0.8611410118406889,
+      "grad_norm": 5.1129703933838755e-05,
+      "learning_rate": 2.434102379786553e-05,
+      "loss": 0.1049,
+      "num_input_tokens_seen": 1468224,
+      "step": 400,
+      "train_runtime": 517.7561,
+      "train_tokens_per_second": 2835.744
+    },
+    {
+      "epoch": 0.8826695371367062,
+      "grad_norm": 5.446456270874478e-05,
+      "learning_rate": 2.4074367249174864e-05,
+      "loss": 0.1024,
+      "num_input_tokens_seen": 1505088,
+      "step": 410,
+      "train_runtime": 530.4568,
+      "train_tokens_per_second": 2837.343
+    },
+    {
+      "epoch": 0.9041980624327234,
+      "grad_norm": 8.551870996598154e-05,
+      "learning_rate": 2.380310867174219e-05,
+      "loss": 0.1155,
+      "num_input_tokens_seen": 1542048,
+      "step": 420,
+      "train_runtime": 543.2418,
+      "train_tokens_per_second": 2838.603
+    },
+    {
+      "epoch": 0.9257265877287406,
+      "grad_norm": 2.7210369808017276e-05,
+      "learning_rate": 2.3527385633260724e-05,
+      "loss": 0.117,
+      "num_input_tokens_seen": 1578496,
+      "step": 430,
+      "train_runtime": 555.7658,
+      "train_tokens_per_second": 2840.218
+    },
+    {
+      "epoch": 0.9472551130247578,
+      "grad_norm": 3.511784234433435e-05,
+      "learning_rate": 2.3247337965557097e-05,
+      "loss": 0.1023,
+      "num_input_tokens_seen": 1613440,
+      "step": 440,
+      "train_runtime": 568.0483,
+      "train_tokens_per_second": 2840.322
+    },
+    {
+      "epoch": 0.9687836383207751,
+      "grad_norm": 2.4012464564293623e-05,
+      "learning_rate": 2.2963107693676192e-05,
+      "loss": 0.1046,
+      "num_input_tokens_seen": 1649760,
+      "step": 450,
+      "train_runtime": 580.7665,
+      "train_tokens_per_second": 2840.66
+    },
+    {
+      "epoch": 0.9903121636167922,
+      "grad_norm": 4.6500768803525716e-05,
+      "learning_rate": 2.2674838963853738e-05,
+      "loss": 0.1106,
+      "num_input_tokens_seen": 1686016,
+      "step": 460,
+      "train_runtime": 593.3342,
+      "train_tokens_per_second": 2841.596
+    },
+    {
+      "epoch": 1.0107642626480087,
+      "grad_norm": 0.00010867212404264137,
+      "learning_rate": 2.2382677970413097e-05,
+      "loss": 0.0859,
+      "num_input_tokens_seen": 1721152,
+      "step": 470,
+      "train_runtime": 605.4653,
+      "train_tokens_per_second": 2842.693
+    },
+    {
+      "epoch": 1.0322927879440258,
+      "grad_norm": 3.351278428453952e-05,
+      "learning_rate": 2.2086772881623455e-05,
+      "loss": 0.1089,
+      "num_input_tokens_seen": 1757568,
+      "step": 480,
+      "train_runtime": 618.0558,
+      "train_tokens_per_second": 2843.705
+    },
+    {
+      "epoch": 1.0538213132400431,
+      "grad_norm": 4.3039726733695716e-05,
+      "learning_rate": 2.1787273764556827e-05,
+      "loss": 0.0852,
+      "num_input_tokens_seen": 1794848,
+      "step": 490,
+      "train_runtime": 630.8577,
+      "train_tokens_per_second": 2845.092
+    },
+    {
+      "epoch": 1.0753498385360603,
+      "grad_norm": 4.8467773012816906e-05,
+      "learning_rate": 2.1484332508982167e-05,
+      "loss": 0.0988,
+      "num_input_tokens_seen": 1832256,
+      "step": 500,
+      "train_runtime": 643.6632,
+      "train_tokens_per_second": 2846.607
+    },
+    {
+      "epoch": 1.0968783638320776,
+      "grad_norm": 3.706678398884833e-05,
+      "learning_rate": 2.11781027503351e-05,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 1868928,
+      "step": 510,
+      "train_runtime": 656.3448,
+      "train_tokens_per_second": 2847.479
+    },
+    {
+      "epoch": 1.1184068891280947,
+      "grad_norm": 4.0160564822144806e-05,
+      "learning_rate": 2.0868739791802312e-05,
+      "loss": 0.1107,
+      "num_input_tokens_seen": 1905824,
+      "step": 520,
+      "train_runtime": 669.0663,
+      "train_tokens_per_second": 2848.483
+    },
+    {
+      "epoch": 1.1399354144241118,
+      "grad_norm": 7.405655196635053e-05,
+      "learning_rate": 2.0556400525560168e-05,
+      "loss": 0.0894,
+      "num_input_tokens_seen": 1941760,
+      "step": 530,
+      "train_runtime": 681.5699,
+      "train_tokens_per_second": 2848.952
+    },
+    {
+      "epoch": 1.1614639397201292,
+      "grad_norm": 4.077348785358481e-05,
+      "learning_rate": 2.0241243353207484e-05,
+      "loss": 0.1016,
+      "num_input_tokens_seen": 1978208,
+      "step": 540,
+      "train_runtime": 694.297,
+      "train_tokens_per_second": 2849.224
+    },
+    {
+      "epoch": 1.1829924650161463,
+      "grad_norm": 4.100809746887535e-05,
+      "learning_rate": 1.9923428105432796e-05,
+      "loss": 0.0932,
+      "num_input_tokens_seen": 2015904,
+      "step": 550,
+      "train_runtime": 707.2242,
+      "train_tokens_per_second": 2850.445
+    },
+    {
+      "epoch": 1.2045209903121636,
+      "grad_norm": 8.349807467311621e-05,
+      "learning_rate": 1.9603115960956852e-05,
+      "loss": 0.1056,
+      "num_input_tokens_seen": 2052672,
+      "step": 560,
+      "train_runtime": 719.9875,
+      "train_tokens_per_second": 2850.983
+    },
+    {
+      "epoch": 1.2260495156081808,
+      "grad_norm": 2.9793711291858926e-05,
+      "learning_rate": 1.9280469364791505e-05,
+      "loss": 0.1033,
+      "num_input_tokens_seen": 2088736,
+      "step": 570,
+      "train_runtime": 732.5942,
+      "train_tokens_per_second": 2851.15
+    },
+    {
+      "epoch": 1.247578040904198,
+      "grad_norm": 6.373477663146332e-05,
+      "learning_rate": 1.8955651945856354e-05,
+      "loss": 0.1035,
+      "num_input_tokens_seen": 2125600,
+      "step": 580,
+      "train_runtime": 745.4583,
+      "train_tokens_per_second": 2851.4
+    },
+    {
+      "epoch": 1.2691065662002152,
+      "grad_norm": 3.1842911994317546e-05,
+      "learning_rate": 1.8628828433995013e-05,
+      "loss": 0.0811,
+      "num_input_tokens_seen": 2161792,
+      "step": 590,
+      "train_runtime": 758.1079,
+      "train_tokens_per_second": 2851.562
+    },
+    {
+      "epoch": 1.2906350914962326,
+      "grad_norm": 4.18422459915746e-05,
+      "learning_rate": 1.8300164576433035e-05,
+      "loss": 0.0888,
+      "num_input_tokens_seen": 2197920,
+      "step": 600,
+      "train_runtime": 770.7916,
+      "train_tokens_per_second": 2851.51
+    },
+    {
+      "epoch": 1.3121636167922497,
+      "grad_norm": 4.866390736424364e-05,
+      "learning_rate": 1.7969827053719862e-05,
+      "loss": 0.0968,
+      "num_input_tokens_seen": 2235168,
+      "step": 610,
+      "train_runtime": 783.669,
+      "train_tokens_per_second": 2852.184
+    },
+    {
+      "epoch": 1.333692142088267,
+      "grad_norm": 4.293969686841592e-05,
+      "learning_rate": 1.7637983395197465e-05,
+      "loss": 0.1311,
+      "num_input_tokens_seen": 2270816,
+      "step": 620,
+      "train_runtime": 796.2193,
+      "train_tokens_per_second": 2851.998
+    },
+    {
+      "epoch": 1.3552206673842842,
+      "grad_norm": 6.0659429436782375e-05,
+      "learning_rate": 1.7304801894038522e-05,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 2308128,
+      "step": 630,
+      "train_runtime": 809.1462,
+      "train_tokens_per_second": 2852.547
+    },
+    {
+      "epoch": 1.3767491926803013,
+      "grad_norm": 3.0738956411369145e-05,
+      "learning_rate": 1.6970451521897198e-05,
+      "loss": 0.0887,
+      "num_input_tokens_seen": 2344992,
+      "step": 640,
+      "train_runtime": 822.0044,
+      "train_tokens_per_second": 2852.773
+    },
+    {
+      "epoch": 1.3982777179763186,
+      "grad_norm": 8.285656076623127e-05,
+      "learning_rate": 1.663510184321587e-05,
+      "loss": 0.1041,
+      "num_input_tokens_seen": 2381056,
+      "step": 650,
+      "train_runtime": 834.6005,
+      "train_tokens_per_second": 2852.929
+    },
+    {
+      "epoch": 1.419806243272336,
+      "grad_norm": 5.0590566388564184e-05,
+      "learning_rate": 1.6298922929231194e-05,
+      "loss": 0.1022,
+      "num_input_tokens_seen": 2417152,
+      "step": 660,
+      "train_runtime": 847.302,
+      "train_tokens_per_second": 2852.763
+    },
+    {
+      "epoch": 1.441334768568353,
+      "grad_norm": 7.397153240162879e-05,
+      "learning_rate": 1.5962085271723158e-05,
+      "loss": 0.0831,
+      "num_input_tokens_seen": 2454048,
+      "step": 670,
+      "train_runtime": 860.0949,
+      "train_tokens_per_second": 2853.229
+    },
+    {
+      "epoch": 1.4628632938643702,
+      "grad_norm": 4.698023258242756e-05,
+      "learning_rate": 1.5624759696550874e-05,
+      "loss": 0.1055,
+      "num_input_tokens_seen": 2491552,
+      "step": 680,
+      "train_runtime": 873.1207,
+      "train_tokens_per_second": 2853.617
+    },
+    {
+      "epoch": 1.4843918191603875,
+      "grad_norm": 4.3750838813139126e-05,
+      "learning_rate": 1.528711727701893e-05,
+      "loss": 0.0812,
+      "num_input_tokens_seen": 2528128,
+      "step": 690,
+      "train_runtime": 885.9021,
+      "train_tokens_per_second": 2853.733
+    },
+    {
+      "epoch": 1.5059203444564049,
+      "grad_norm": 6.37572884443216e-05,
+      "learning_rate": 1.4949329247118258e-05,
+      "loss": 0.1016,
+      "num_input_tokens_seen": 2564448,
+      "step": 700,
+      "train_runtime": 898.5373,
+      "train_tokens_per_second": 2854.025
+    },
+    {
+      "epoch": 1.527448869752422,
+      "grad_norm": 4.437049574335106e-05,
+      "learning_rate": 1.4611566914685489e-05,
+      "loss": 0.0956,
+      "num_input_tokens_seen": 2601056,
+      "step": 710,
+      "train_runtime": 911.3255,
+      "train_tokens_per_second": 2854.146
+    },
+    {
+      "epoch": 1.5489773950484391,
+      "grad_norm": 5.1571758376667276e-05,
+      "learning_rate": 1.4274001574524894e-05,
+      "loss": 0.0936,
+      "num_input_tokens_seen": 2637888,
+      "step": 720,
+      "train_runtime": 924.1173,
+      "train_tokens_per_second": 2854.495
+    },
+    {
+      "epoch": 1.5705059203444565,
+      "grad_norm": 6.383092113537714e-05,
+      "learning_rate": 1.3936804421536906e-05,
+      "loss": 0.0728,
+      "num_input_tokens_seen": 2673312,
+      "step": 730,
+      "train_runtime": 937.0636,
+      "train_tokens_per_second": 2852.861
+    },
+    {
+      "epoch": 1.5920344456404736,
+      "grad_norm": 2.989832137245685e-05,
+      "learning_rate": 1.3600146463897327e-05,
+      "loss": 0.0839,
+      "num_input_tokens_seen": 2709696,
+      "step": 740,
+      "train_runtime": 949.7913,
+      "train_tokens_per_second": 2852.938
+    },
+    {
+      "epoch": 1.6135629709364907,
+      "grad_norm": 0.00010528949496801943,
+      "learning_rate": 1.3264198436331193e-05,
+      "loss": 0.1308,
+      "num_input_tokens_seen": 2746528,
+      "step": 750,
+      "train_runtime": 962.4815,
+      "train_tokens_per_second": 2853.591
+    },
+    {
+      "epoch": 1.635091496232508,
+      "grad_norm": 3.306402868474834e-05,
+      "learning_rate": 1.2929130713525396e-05,
+      "loss": 0.0927,
+      "num_input_tokens_seen": 2782656,
+      "step": 760,
+      "train_runtime": 975.0888,
+      "train_tokens_per_second": 2853.746
+    },
+    {
+      "epoch": 1.6566200215285254,
+      "grad_norm": 3.532276605255902e-05,
+      "learning_rate": 1.2595113223723788e-05,
+      "loss": 0.1135,
+      "num_input_tokens_seen": 2819744,
+      "step": 770,
+      "train_runtime": 987.9762,
+      "train_tokens_per_second": 2854.061
+    },
+    {
+      "epoch": 1.6781485468245425,
+      "grad_norm": 6.426728941733018e-05,
+      "learning_rate": 1.2262315362548825e-05,
+      "loss": 0.0813,
+      "num_input_tokens_seen": 2857216,
+      "step": 780,
+      "train_runtime": 1001.0086,
+      "train_tokens_per_second": 2854.337
+    },
+    {
+      "epoch": 1.6996770721205596,
+      "grad_norm": 8.145067113218829e-05,
+      "learning_rate": 1.1930905907093196e-05,
+      "loss": 0.0933,
+      "num_input_tokens_seen": 2892000,
+      "step": 790,
+      "train_runtime": 1013.4209,
+      "train_tokens_per_second": 2853.701
+    },
+    {
+      "epoch": 1.721205597416577,
+      "grad_norm": 0.00011963866563746706,
+      "learning_rate": 1.1601052930325236e-05,
+      "loss": 0.0881,
+      "num_input_tokens_seen": 2928832,
+      "step": 800,
+      "train_runtime": 1026.3231,
+      "train_tokens_per_second": 2853.714
+    },
+    {
+      "epoch": 1.7427341227125943,
+      "grad_norm": 5.128612610860728e-05,
+      "learning_rate": 1.127292371585141e-05,
+      "loss": 0.079,
+      "num_input_tokens_seen": 2965632,
+      "step": 810,
+      "train_runtime": 1039.1265,
+      "train_tokens_per_second": 2853.966
+    },
+    {
+      "epoch": 1.7642626480086114,
+      "grad_norm": 7.751452358206734e-05,
+      "learning_rate": 1.0946684673079074e-05,
+      "loss": 0.1132,
+      "num_input_tokens_seen": 3002496,
+      "step": 820,
+      "train_runtime": 1051.9839,
+      "train_tokens_per_second": 2854.127
+    },
+    {
+      "epoch": 1.7857911733046286,
+      "grad_norm": 4.0213129977928475e-05,
+      "learning_rate": 1.0622501252822628e-05,
+      "loss": 0.0793,
+      "num_input_tokens_seen": 3038944,
+      "step": 830,
+      "train_runtime": 1064.6802,
+      "train_tokens_per_second": 2854.326
+    },
+    {
+      "epoch": 1.807319698600646,
+      "grad_norm": 4.586371505865827e-05,
+      "learning_rate": 1.0300537863395823e-05,
+      "loss": 0.0918,
+      "num_input_tokens_seen": 3076032,
+      "step": 840,
+      "train_runtime": 1077.5859,
+      "train_tokens_per_second": 2854.559
+    },
+    {
+      "epoch": 1.8288482238966632,
+      "grad_norm": 3.2480071240570396e-05,
+      "learning_rate": 9.98095778723271e-06,
+      "loss": 0.093,
+      "num_input_tokens_seen": 3112448,
+      "step": 850,
+      "train_runtime": 1090.3056,
+      "train_tokens_per_second": 2854.656
+    },
+    {
+      "epoch": 1.8503767491926801,
+      "grad_norm": 5.541834980249405e-05,
+      "learning_rate": 9.663923098079656e-06,
+      "loss": 0.0973,
+      "num_input_tokens_seen": 3149376,
+      "step": 860,
+      "train_runtime": 1103.1339,
+      "train_tokens_per_second": 2854.935
+    },
+    {
+      "epoch": 1.8719052744886975,
+      "grad_norm": 3.055919296457432e-05,
+      "learning_rate": 9.349594578800259e-06,
+      "loss": 0.0995,
+      "num_input_tokens_seen": 3186304,
+      "step": 870,
+      "train_runtime": 1116.0618,
+      "train_tokens_per_second": 2854.953
+    },
+    {
+      "epoch": 1.8934337997847148,
+      "grad_norm": 4.6285113057820126e-05,
+      "learning_rate": 9.038131639834977e-06,
+      "loss": 0.1238,
+      "num_input_tokens_seen": 3222752,
+      "step": 880,
+      "train_runtime": 1128.7812,
+      "train_tokens_per_second": 2855.072
+    },
+    {
+      "epoch": 1.914962325080732,
+      "grad_norm": 4.483849261305295e-05,
+      "learning_rate": 8.7296922383568e-06,
+      "loss": 0.0912,
+      "num_input_tokens_seen": 3258976,
+      "step": 890,
+      "train_runtime": 1141.5159,
+      "train_tokens_per_second": 2854.954
+    },
+    {
+      "epoch": 1.936490850376749,
+      "grad_norm": 5.5885881010908633e-05,
+      "learning_rate": 8.424432798163838e-06,
+      "loss": 0.0935,
+      "num_input_tokens_seen": 3295040,
+      "step": 900,
+      "train_runtime": 1154.1132,
+      "train_tokens_per_second": 2855.041
+    },
+    {
+      "epoch": 1.9580193756727664,
+      "grad_norm": 3.8836213207105175e-05,
+      "learning_rate": 8.122508130349711e-06,
+      "loss": 0.0722,
+      "num_input_tokens_seen": 3331392,
+      "step": 910,
+      "train_runtime": 1166.7365,
+      "train_tokens_per_second": 2855.308
+    },
+    {
+      "epoch": 1.9795479009687837,
+      "grad_norm": 6.445301551138982e-05,
+      "learning_rate": 7.82407135479164e-06,
+      "loss": 0.1001,
+      "num_input_tokens_seen": 3368032,
+      "step": 920,
+      "train_runtime": 1179.5494,
+      "train_tokens_per_second": 2855.355
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 1.651374805078376e-05,
+      "learning_rate": 7.529273822496379e-06,
+      "loss": 0.0647,
+      "num_input_tokens_seen": 3403584,
+      "step": 930,
+      "train_runtime": 1191.8451,
+      "train_tokens_per_second": 2855.727
+    },
+    {
+      "epoch": 2.0215285252960173,
+      "grad_norm": 9.509055962553248e-05,
+      "learning_rate": 7.238265038843143e-06,
+      "loss": 0.0881,
+      "num_input_tokens_seen": 3440000,
+      "step": 940,
+      "train_runtime": 1204.5181,
+      "train_tokens_per_second": 2855.914
+    },
+    {
+      "epoch": 2.0430570505920342,
+      "grad_norm": 5.2039082220289856e-05,
+      "learning_rate": 6.951192587762555e-06,
+      "loss": 0.0573,
+      "num_input_tokens_seen": 3476640,
+      "step": 950,
+      "train_runtime": 1217.2855,
+      "train_tokens_per_second": 2856.06
+    },
+    {
+      "epoch": 2.0645855758880516,
+      "grad_norm": 3.832239599432796e-05,
+      "learning_rate": 6.668202056890052e-06,
+      "loss": 0.0646,
+      "num_input_tokens_seen": 3512992,
+      "step": 960,
+      "train_runtime": 1229.9665,
+      "train_tokens_per_second": 2856.169
+    },
+    {
+      "epoch": 2.086114101184069,
+      "grad_norm": 5.945537122897804e-05,
+      "learning_rate": 6.389436963731707e-06,
+      "loss": 0.0682,
+      "num_input_tokens_seen": 3549504,
+      "step": 970,
+      "train_runtime": 1242.6763,
+      "train_tokens_per_second": 2856.338
+    },
+    {
+      "epoch": 2.1076426264800863,
+      "grad_norm": 6.592880527023226e-05,
+      "learning_rate": 6.115038682879904e-06,
+      "loss": 0.0931,
+      "num_input_tokens_seen": 3586592,
+      "step": 980,
+      "train_runtime": 1255.5495,
+      "train_tokens_per_second": 2856.591
+    },
+    {
+      "epoch": 2.129171151776103,
+      "grad_norm": 2.9424012609524652e-05,
+      "learning_rate": 5.845146374315781e-06,
+      "loss": 0.0686,
+      "num_input_tokens_seen": 3622944,
+      "step": 990,
+      "train_runtime": 1268.2309,
+      "train_tokens_per_second": 2856.691
+    },
+    {
+      "epoch": 2.1506996770721205,
+      "grad_norm": 5.797382618766278e-05,
+      "learning_rate": 5.579896912834807e-06,
+      "loss": 0.0917,
+      "num_input_tokens_seen": 3660096,
+      "step": 1000,
+      "train_runtime": 1281.1218,
+      "train_tokens_per_second": 2856.946
+    },
+    {
+      "epoch": 2.172228202368138,
+      "grad_norm": 5.529261034098454e-05,
+      "learning_rate": 5.319424818631293e-06,
+      "loss": 0.0931,
+      "num_input_tokens_seen": 3697024,
+      "step": 1010,
+      "train_runtime": 1294.9799,
+      "train_tokens_per_second": 2854.889
+    },
+    {
+      "epoch": 2.193756727664155,
+      "grad_norm": 5.912356573389843e-05,
+      "learning_rate": 5.063862189076988e-06,
+      "loss": 0.0855,
+      "num_input_tokens_seen": 3734240,
+      "step": 1020,
+      "train_runtime": 1307.7689,
+      "train_tokens_per_second": 2855.428
+    },
+    {
+      "epoch": 2.215285252960172,
+      "grad_norm": 5.3946208936395124e-05,
+      "learning_rate": 4.813338631728471e-06,
+      "loss": 0.0909,
+      "num_input_tokens_seen": 3771232,
+      "step": 1030,
+      "train_runtime": 1320.5384,
+      "train_tokens_per_second": 2855.829
+    },
+    {
+      "epoch": 2.2368137782561894,
+      "grad_norm": 6.292902253335342e-05,
+      "learning_rate": 4.567981198597168e-06,
+      "loss": 0.0817,
+      "num_input_tokens_seen": 3808416,
+      "step": 1040,
+      "train_runtime": 1333.3555,
+      "train_tokens_per_second": 2856.264
+    },
+    {
+      "epoch": 2.2583423035522068,
+      "grad_norm": 4.050394636578858e-05,
+      "learning_rate": 4.327914321715459e-06,
+      "loss": 0.0599,
+      "num_input_tokens_seen": 3844768,
+      "step": 1050,
+      "train_runtime": 1345.9495,
+      "train_tokens_per_second": 2856.547
+    },
+    {
+      "epoch": 2.2798708288482237,
+      "grad_norm": 5.740273991250433e-05,
+      "learning_rate": 4.09325975003146e-06,
+      "loss": 0.0711,
+      "num_input_tokens_seen": 3881088,
+      "step": 1060,
+      "train_runtime": 1358.6041,
+      "train_tokens_per_second": 2856.673
+    },
+    {
+      "epoch": 2.301399354144241,
+      "grad_norm": 7.280943100340664e-05,
+      "learning_rate": 3.864136487664538e-06,
+      "loss": 0.0723,
+      "num_input_tokens_seen": 3916928,
+      "step": 1070,
+      "train_runtime": 1371.1849,
+      "train_tokens_per_second": 2856.601
+    },
+    {
+      "epoch": 2.3229278794402584,
+      "grad_norm": 4.93960760650225e-05,
+      "learning_rate": 3.640660733552863e-06,
+      "loss": 0.0524,
+      "num_input_tokens_seen": 3952928,
+      "step": 1080,
+      "train_runtime": 1383.7869,
+      "train_tokens_per_second": 2856.602
+    },
+    {
+      "epoch": 2.3444564047362757,
+      "grad_norm": 7.801376341376454e-05,
+      "learning_rate": 3.4229458225235696e-06,
+      "loss": 0.0513,
+      "num_input_tokens_seen": 3990720,
+      "step": 1090,
+      "train_runtime": 1396.9725,
+      "train_tokens_per_second": 2856.692
+    },
+    {
+      "epoch": 2.3659849300322926,
+      "grad_norm": 4.576309584081173e-05,
+      "learning_rate": 3.211102167815469e-06,
+      "loss": 0.0773,
+      "num_input_tokens_seen": 4026656,
+      "step": 1100,
+      "train_runtime": 1409.5674,
+      "train_tokens_per_second": 2856.661
+    },
+    {
+      "epoch": 2.38751345532831,
+      "grad_norm": 0.0001336225977865979,
+      "learning_rate": 3.0052372050834025e-06,
+      "loss": 0.0896,
+      "num_input_tokens_seen": 4064192,
+      "step": 1110,
+      "train_runtime": 1422.5111,
+      "train_tokens_per_second": 2857.055
+    },
+    {
+      "epoch": 2.4090419806243273,
+      "grad_norm": 8.807591075310484e-05,
+      "learning_rate": 2.805455337912722e-06,
+      "loss": 0.0719,
+      "num_input_tokens_seen": 4101632,
+      "step": 1120,
+      "train_runtime": 1435.3867,
+      "train_tokens_per_second": 2857.51
+    },
+    {
+      "epoch": 2.4305705059203446,
+      "grad_norm": 3.673565879580565e-05,
+      "learning_rate": 2.611857884871379e-06,
+      "loss": 0.0801,
+      "num_input_tokens_seen": 4138208,
+      "step": 1130,
+      "train_runtime": 1448.0115,
+      "train_tokens_per_second": 2857.856
+    },
+    {
+      "epoch": 2.4520990312163615,
+      "grad_norm": 2.5486542654107325e-05,
+      "learning_rate": 2.424543028126661e-06,
+      "loss": 0.0821,
+      "num_input_tokens_seen": 4174592,
+      "step": 1140,
+      "train_runtime": 1460.6997,
+      "train_tokens_per_second": 2857.94
+    },
+    {
+      "epoch": 2.473627556512379,
+      "grad_norm": 6.899920845171437e-05,
+      "learning_rate": 2.243605763652489e-06,
+      "loss": 0.0654,
+      "num_input_tokens_seen": 4211616,
+      "step": 1150,
+      "train_runtime": 1473.5308,
+      "train_tokens_per_second": 2858.18
+    },
+    {
+      "epoch": 2.495156081808396,
+      "grad_norm": 4.8038589739007875e-05,
+      "learning_rate": 2.0691378530525933e-06,
+      "loss": 0.0767,
+      "num_input_tokens_seen": 4247744,
+      "step": 1160,
+      "train_runtime": 1486.1222,
+      "train_tokens_per_second": 2858.274
+    },
+    {
+      "epoch": 2.5166846071044136,
+      "grad_norm": 7.49904866097495e-05,
+      "learning_rate": 1.9012277770239829e-06,
+      "loss": 0.0655,
+      "num_input_tokens_seen": 4284224,
+      "step": 1170,
+      "train_runtime": 1498.696,
+      "train_tokens_per_second": 2858.634
+    },
+    {
+      "epoch": 2.5382131324004304,
+      "grad_norm": 6.3829000282567e-05,
+      "learning_rate": 1.7399606904843224e-06,
+      "loss": 0.0584,
+      "num_input_tokens_seen": 4320672,
+      "step": 1180,
+      "train_runtime": 1511.4477,
+      "train_tokens_per_second": 2858.631
+    },
+    {
+      "epoch": 2.559741657696448,
+      "grad_norm": 5.6509288697270676e-05,
+      "learning_rate": 1.5854183793859577e-06,
+      "loss": 0.0601,
+      "num_input_tokens_seen": 4358080,
+      "step": 1190,
+      "train_runtime": 1524.3591,
+      "train_tokens_per_second": 2858.959
+    },
+    {
+      "epoch": 2.581270182992465,
+      "grad_norm": 4.297928171581589e-05,
+      "learning_rate": 1.437679219238498e-06,
+      "loss": 0.0665,
+      "num_input_tokens_seen": 4394752,
+      "step": 1200,
+      "train_runtime": 1537.0429,
+      "train_tokens_per_second": 2859.225
+    },
+    {
+      "epoch": 2.602798708288482,
+      "grad_norm": 8.681658800924197e-05,
+      "learning_rate": 1.2968181353609854e-06,
+      "loss": 0.0658,
+      "num_input_tokens_seen": 4431808,
+      "step": 1210,
+      "train_runtime": 1549.8548,
+      "train_tokens_per_second": 2859.499
+    },
+    {
+      "epoch": 2.6243272335844994,
+      "grad_norm": 3.503639891277999e-05,
+      "learning_rate": 1.1629065648838495e-06,
+      "loss": 0.0898,
+      "num_input_tokens_seen": 4467328,
+      "step": 1220,
+      "train_runtime": 1562.3415,
+      "train_tokens_per_second": 2859.38
+    },
+    {
+      "epoch": 2.6458557588805167,
+      "grad_norm": 3.0625513318227604e-05,
+      "learning_rate": 1.0360124205198102e-06,
+      "loss": 0.0619,
+      "num_input_tokens_seen": 4503808,
+      "step": 1230,
+      "train_runtime": 1575.0206,
+      "train_tokens_per_second": 2859.523
+    },
+    {
+      "epoch": 2.667384284176534,
+      "grad_norm": 6.452269735746086e-05,
+      "learning_rate": 9.162000561222672e-07,
+      "loss": 0.0763,
+      "num_input_tokens_seen": 4541536,
+      "step": 1240,
+      "train_runtime": 1588.024,
+      "train_tokens_per_second": 2859.866
+    },
+    {
+      "epoch": 2.6889128094725514,
+      "grad_norm": 2.9151902708690614e-05,
+      "learning_rate": 8.035302340484646e-07,
+      "loss": 0.0909,
+      "num_input_tokens_seen": 4577664,
+      "step": 1250,
+      "train_runtime": 1600.6009,
+      "train_tokens_per_second": 2859.966
+    },
+    {
+      "epoch": 2.7104413347685683,
+      "grad_norm": 6.271711754379794e-05,
+      "learning_rate": 6.980600943441273e-07,
+      "loss": 0.0683,
+      "num_input_tokens_seen": 4614176,
+      "step": 1260,
+      "train_runtime": 1613.2327,
+      "train_tokens_per_second": 2860.205
+    },
+    {
+      "epoch": 2.7319698600645856,
+      "grad_norm": 5.843062899657525e-05,
+      "learning_rate": 5.998431257651132e-07,
+      "loss": 0.0867,
+      "num_input_tokens_seen": 4650848,
+      "step": 1270,
+      "train_runtime": 1625.9328,
+      "train_tokens_per_second": 2860.418
+    },
+    {
+      "epoch": 2.7534983853606025,
+      "grad_norm": 5.628829603665508e-05,
+      "learning_rate": 5.089291386508055e-07,
+      "loss": 0.0648,
+      "num_input_tokens_seen": 4687456,
+      "step": 1280,
+      "train_runtime": 1638.6467,
+      "train_tokens_per_second": 2860.565
+    },
+    {
+      "epoch": 2.77502691065662,
+      "grad_norm": 4.340551095083356e-05,
+      "learning_rate": 4.2536423966299256e-07,
+      "loss": 0.0672,
+      "num_input_tokens_seen": 4723840,
+      "step": 1290,
+      "train_runtime": 1651.315,
+      "train_tokens_per_second": 2860.653
+    },
+    {
+      "epoch": 2.7965554359526372,
+      "grad_norm": 7.965038093971089e-05,
+      "learning_rate": 3.491908084030776e-07,
+      "loss": 0.089,
+      "num_input_tokens_seen": 4761056,
+      "step": 1300,
+      "train_runtime": 1664.1238,
+      "train_tokens_per_second": 2860.999
+    },
+    {
+      "epoch": 2.8180839612486546,
+      "grad_norm": 5.358497583074495e-05,
+      "learning_rate": 2.804474759194176e-07,
+      "loss": 0.07,
+      "num_input_tokens_seen": 4798240,
+      "step": 1310,
+      "train_runtime": 1676.8916,
+      "train_tokens_per_second": 2861.39
+    },
+    {
+      "epoch": 2.839612486544672,
+      "grad_norm": 3.2774096325738356e-05,
+      "learning_rate": 2.1916910511575627e-07,
+      "loss": 0.0807,
+      "num_input_tokens_seen": 4834560,
+      "step": 1320,
+      "train_runtime": 1689.486,
+      "train_tokens_per_second": 2861.557
+    },
+    {
+      "epoch": 2.861141011840689,
+      "grad_norm": 6.476557609857991e-05,
+      "learning_rate": 1.6538677307062933e-07,
+      "loss": 0.0669,
+      "num_input_tokens_seen": 4871328,
+      "step": 1330,
+      "train_runtime": 1702.2714,
+      "train_tokens_per_second": 2861.664
+    },
+    {
+      "epoch": 2.882669537136706,
+      "grad_norm": 8.176828123396263e-05,
+      "learning_rate": 1.1912775527676278e-07,
+      "loss": 0.0552,
+      "num_input_tokens_seen": 4907744,
+      "step": 1340,
+      "train_runtime": 1714.8671,
+      "train_tokens_per_second": 2861.88
+    },
+    {
+      "epoch": 2.9041980624327235,
+      "grad_norm": 7.269597699632868e-05,
+      "learning_rate": 8.04155118084099e-08,
+      "loss": 0.0917,
+      "num_input_tokens_seen": 4943712,
+      "step": 1350,
+      "train_runtime": 1727.4271,
+      "train_tokens_per_second": 2861.893
+    },
+    {
+      "epoch": 2.9257265877287404,
+      "grad_norm": 3.572047353372909e-05,
+      "learning_rate": 4.9269675423671556e-08,
+      "loss": 0.074,
+      "num_input_tokens_seen": 4980480,
+      "step": 1360,
+      "train_runtime": 1740.2346,
+      "train_tokens_per_second": 2861.959
+    },
+    {
+      "epoch": 2.9472551130247577,
+      "grad_norm": 8.610687655163929e-05,
+      "learning_rate": 2.570604160781753e-08,
+      "loss": 0.0599,
+      "num_input_tokens_seen": 5016800,
+      "step": 1370,
+      "train_runtime": 1752.9074,
+      "train_tokens_per_second": 2861.988
+    },
+    {
+      "epoch": 2.968783638320775,
+      "grad_norm": 5.934851287747733e-05,
+      "learning_rate": 9.736560562675912e-09,
+      "loss": 0.0593,
+      "num_input_tokens_seen": 5053440,
+      "step": 1380,
+      "train_runtime": 1765.6959,
+      "train_tokens_per_second": 2862.01
+    },
+    {
+      "epoch": 2.9903121636167924,
+      "grad_norm": 3.7563720979960635e-05,
+      "learning_rate": 1.3693311461326908e-09,
+      "loss": 0.0763,
+      "num_input_tokens_seen": 5090592,
+      "step": 1390,
+      "train_runtime": 1778.5239,
+      "train_tokens_per_second": 2862.257
+    },
+    {
+      "epoch": 3.0,
+      "num_input_tokens_seen": 5107072,
+      "step": 1395,
+      "total_flos": 2.186269295986606e+17,
+      "train_loss": 0.10808860245144068,
+      "train_runtime": 1785.3385,
+      "train_samples_per_second": 6.244,
+      "train_steps_per_second": 0.781
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 1395,
+  "num_input_tokens_seen": 5107072,
+  "num_train_epochs": 3,
+  "save_steps": 1000,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 2.186269295986606e+17,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0da07f323cd8cadc582879cf09f450ad877537a29fbf8dd1bdb4c30cac53e814
+size 6161

training_args.yaml

@@ -0,0 +1,38 @@
+cutoff_len: 2048
+dataset: treino_pt_rde
+dataset_dir: data
+ddp_timeout: 180000000
+do_train: true
+double_quantization: true
+enable_thinking: true
+finetuning_type: lora
+flash_attn: auto
+fp16: true
+gradient_accumulation_steps: 2
+include_num_input_tokens_seen: true
+learning_rate: 3.0e-05
+logging_steps: 10
+lora_alpha: 16
+lora_dropout: 0
+lora_rank: 8
+lora_target: all
+lr_scheduler_type: cosine
+max_grad_norm: 1.0
+max_samples: 3716
+model_name_or_path: mistralai/Mistral-7B-Instruct-v0.3
+num_train_epochs: 3.0
+optim: adamw_torch
+output_dir: saves/Mistral-7B-Instruct-v0.3/lora/mistral-treinado
+packing: false
+per_device_train_batch_size: 4
+plot_loss: true
+preprocessing_num_workers: 16
+quantization_bit: 4
+quantization_method: bnb
+report_to: none
+save_steps: 1000
+stage: sft
+template: alpaca
+trust_remote_code: true
+use_unsloth: true
+warmup_steps: 0