{
  "arguments_by_text_source": {
    "train_chaski_r2.py": {
      "bf16": true,
      "gradient_accumulation_steps": 4,
      "learning_rate": 0.0002,
      "logging_steps": 1,
      "lr_scheduler_type": "constant_with_warmup",
      "num_train_epochs": 3,
      "optim": "adamw_8bit",
      "output_dir": "outputs",
      "per_device_train_batch_size": 1,
      "push_to_hub": true,
      "report_to": "none",
      "warmup_steps": 6,
      "weight_decay": 0.01
    }
  },
  "migration_parent_revision": "96db8a2a2c27e84d79392b27dee3ea0d5ecae813",
  "purpose": "non-executable record replacing serialized trainer arguments",
  "replaces": {
    "deserialized": false,
    "introduced_by_revision": "b83b5a552957a410cc74c9b68f288d2acff0ec5f",
    "path": "training_args.bin",
    "serialization": "pytorch-zip-pickle",
    "sha256": "ecab4c36d396567a4733156c93437f64806673972a95a97565a49815b53f5e02",
    "size_bytes": 5713
  },
  "repository": "SZLHOLDINGS/chaski-r2",
  "schema": "szl.training-arguments/v1",
  "scope": {
    "argument_completeness": "script-explicit-subset",
    "generated_by_deserializing_pickle": false,
    "inference_required": false,
    "resume_checkpoint": false
  },
  "sources": [
    {
      "path": "README.md",
      "revision": "b83b5a552957a410cc74c9b68f288d2acff0ec5f",
      "sha256": "eb3f974aeb3f296745762d690585dea2afa112497a03f2ae8bb5b380fccc6ab4"
    },
    {
      "path": "train_chaski_r2.py",
      "revision": "b83b5a552957a410cc74c9b68f288d2acff0ec5f",
      "sha256": "f8b06195536c75dc4d17d68b2cacdf791f9c9fe29b8afea28f95d26d07b70984"
    }
  ],
  "training": {
    "framework_versions": {
      "datasets": "4.3.0",
      "pytorch": "2.11.0",
      "tokenizers": "0.22.2",
      "transformers": "5.5.0",
      "trl": "0.24.0"
    },
    "method": "SFT"
  }
}