qwen2.5-0.5b-zhihu-sft / train_results.json
yinmingzhang's picture
Model save
09f143b verified
raw
history blame contribute delete
226 Bytes
{
"epoch": 5.0,
"total_flos": 18203285913600.0,
"train_loss": 3.477997573216756,
"train_runtime": 36.1944,
"train_samples": 568,
"train_samples_per_second": 12.295,
"train_steps_per_second": 1.658
}