Qwen2.5-7B-Instruct-method2 / train_results.json
luowenyang's picture
Upload folder using huggingface_hub
d30b2d7 verified
raw
history blame contribute delete
204 Bytes
{
"epoch": 3.0,
"total_flos": 178716929818624.0,
"train_loss": 0.46517832586259555,
"train_runtime": 7796.3988,
"train_samples_per_second": 1.352,
"train_steps_per_second": 0.169
}