qwen2.5-0.5b-expo-DPO-ES-TRY3 / train_results.json
hZzy's picture
Model save
dc3c59b verified
raw
history blame contribute delete
235 Bytes
{
"epoch": 2.9976381672177608,
"total_flos": 0.0,
"train_loss": 0.022545777056648425,
"train_runtime": 4386.5835,
"train_samples": 50802,
"train_samples_per_second": 34.744,
"train_steps_per_second": 0.241
}