| { |
| "config": { |
| "model": { |
| "family": "qwen2.5", |
| "checkpoint": "Qwen/Qwen2.5-7B", |
| "dtype": "bfloat16", |
| "compute_dtype": null |
| }, |
| "data": [ |
| { |
| "name": "json", |
| "config_name": null, |
| "data_files": "/workspace/sweep-b/mix/software_docs_qa_d01.jsonl", |
| "split": "train", |
| "format": "text", |
| "text_column": "text", |
| "text_template": null, |
| "seq_len": 128, |
| "num_proc": 16, |
| "append_eos": false |
| } |
| ], |
| "training": { |
| "epochs": 3, |
| "batch_size": 8, |
| "gradient_accumulation_steps": 16, |
| "log_every": 100, |
| "save_every_steps": 0, |
| "output_dir": "/workspace/sweep-b/out/software_docs_qa_d01_s3", |
| "seed": 3, |
| "compile": true, |
| "max_steps": -1, |
| "save_resumable": false, |
| "gradient_checkpointing": false |
| }, |
| "optim": { |
| "lr": 7e-05, |
| "betas": [ |
| 0.9, |
| 0.95 |
| ], |
| "weight_decay": 0.1, |
| "scheduler": "cosine", |
| "warmup_steps": 0, |
| "warmup_ratio": 0.1, |
| "grad_clip": 1.0, |
| "min_lr_ratio": 0.0 |
| }, |
| "profiling": { |
| "peak_flops": "auto" |
| } |
| }, |
| "resolved": { |
| "warmup_steps": 296, |
| "total_steps": 2958, |
| "steps_per_epoch": 986, |
| "num_train_examples": 126250, |
| "effective_batch_size": 128, |
| "peak_flops_bf16": 2250000000000000.0, |
| "dtype_loaded": "torch.bfloat16" |
| }, |
| "env": { |
| "torch_version": "2.11.0+cu130", |
| "cuda_version": "13.0", |
| "device_name": "NVIDIA B200" |
| }, |
| "code": { |
| "nanoft_version": null, |
| "git_commit": "46ac50b419fa13ee25c5de923369b616d65d4a48", |
| "git_dirty": false |
| } |
| } |
