# Instruction SFT - distilgpt2
Artefatti del training **Instruction SFT**.

## Parametri di Training
```json
{
  "model_name": "distilgpt2",
  "dataset_name": "/home/raniero/test_instr.jsonl",
  "output_dir": "/app/instruction_output",
  "seed": 42,
  "num_train_epochs": 1,
  "per_device_train_batch_size": 2,
  "learning_rate": 2e-05,
  "warmup_ratio": 0.03,
  "gradient_accumulation_steps": 1,
  "bf16": false,
  "repo_id": "raniero/test_instr",
  "hf_token": ""
}
```

## Metriche
```json
{
  "train_runtime": 0.531,
  "train_samples_per_second": 5.649,
  "train_steps_per_second": 3.766,
  "total_flos": 32151748608.0,
  "train_loss": 6.437064170837402,
  "epoch": 1.0
}
```

## File & SHA256
```json
{
  "generation_config.json": "3da00cdf443a1b5fc8234e028411b32f81f919e3b9154b14aba6d9c1441684b8",
  "train_instruction.log": "f3566e7c733e4a6549930d2b7b2e02e1d37a48a51933dffc374b38b8847588da",
  "special_tokens_map.json": "8bcf4e58d0970bbcb3b111a4c1e297c082fbf61253b14b3c0c8f2972ccb61ba4",
  "tokenizer_config.json": "f7cd5e1168079212fdf834ed8113826afe1384b42c39024753c09d3eac9335c0",
  "README.md": "7031f98b79de51938e7febb7134419f24bd3b60e0265dd11ab4f922f847c776c",
  "upload_meta.json": "cb01cdcb18a8c107b17b5fe64c3d1a1c33a6282f7ae4c9b2f07e590aa0bd7d70",
  "metrics.json": "74d200af150bad486402e9170ad1da8cc1c51fa1429e4e3873f8ece6ee4dadf3",
  "model.safetensors": "11b7c2c699d17b6142e9774cffb8411617d64e652d30b7b6d5b4d79d4cbe9e07",
  "config.json": "7ff605d8dfddfa0eb05bfeb09b458ffe79552291bc1c78a7b99c763cf4cc705a",
  "tokenizer.json": "5b1d2a4396fc21ff3a5814684a1f90fce358c2666bddd51ffad78cf22e7a1fee",
  "tokenizer.model": "dadfd56d766715c61d2ef780a525ab43b8e6da4de6865bda3d95fdef5e134055",
  "model.safetensors.index.json": "f0cb2f153f1cc69af0264f8e24a6152638231bf06c84c68763d6d70ae153ed51",
  "vocab.json": "3ba3c3109ff33976c4bd966589c11ee14fcaa1f4c9e5e154c2ed7f99d80709e7",
  "merges.txt": "1ce1664773c50f3e0cc8842619a93edc4624525b728b188a9e0be33b7726adc5"
}
```