Qwen2.5-7B-Instruct-method1 / train_results.json
luowenyang's picture
Upload folder using huggingface_hub
734265b verified
raw
history blame contribute delete
201 Bytes
{
"epoch": 3.0,
"total_flos": 148330505830400.0,
"train_loss": 0.33294462082964,
"train_runtime": 5380.5167,
"train_samples_per_second": 1.959,
"train_steps_per_second": 0.245
}