Qwen3-4B-Base-Instruct / train_results.json
ShourenWSR's picture
Upload folder using huggingface_hub
ab3ca84 verified
{
"epoch": 3.0,
"total_flos": 1.0710730446253916e+18,
"train_loss": 0.6314072675106341,
"train_runtime": 9371.9509,
"train_samples_per_second": 16.64,
"train_steps_per_second": 0.693
}