qwen2.5-0.5b-expo-DPO-ES-1000 / train_results.json
hZzy's picture
Model save
2573912 verified
raw
history blame
232 Bytes
{
"epoch": 3.122342938119981,
"total_flos": 0.0,
"train_loss": 1018.9660795454546,
"train_runtime": 6860.7986,
"train_samples": 50802,
"train_samples_per_second": 37.023,
"train_steps_per_second": 0.257
}