| { |
| "base_model": "Qwen/Qwen2.5-1.5B-Instruct", |
| "dataset_repo": "whosouravsharma/paper-qa-qasper-sft", |
| "lora": { |
| "r": 16, |
| "alpha": 32, |
| "dropout": 0.05, |
| "target_modules": [ |
| "q_proj", |
| "k_proj", |
| "v_proj", |
| "o_proj" |
| ] |
| }, |
| "training": { |
| "learning_rate": 0.0002, |
| "epochs": 3.0, |
| "per_device_batch_size": 4, |
| "gradient_accumulation_steps": 4, |
| "max_seq_length": 2048 |
| }, |
| "quantization": "4bit-nf4", |
| "trainable_params": 4358144, |
| "total_params": 892974592, |
| "trainable_pct": 0.488, |
| "final_eval_metrics": { |
| "eval_loss": 1.285805106163025, |
| "eval_runtime": 207.0525, |
| "eval_samples_per_second": 8.283, |
| "eval_steps_per_second": 2.072, |
| "eval_entropy": 1.2657356866589793, |
| "eval_num_tokens": 3310809.0, |
| "eval_mean_token_accuracy": 0.7257560479057419, |
| "epoch": 3.0 |
| }, |
| "trained_at": "2026-09-04T11:54:00.127450+00:00" |
| } |
