qwen2.5-0.5b-expo-DPO-ES2-0.1 / train_results.json
hZzy's picture
Model save
6a901b2 verified
raw
history blame contribute delete
233 Bytes
{
"epoch": 0.9919697685403873,
"total_flos": 0.0,
"train_loss": 0.5958910301753453,
"train_runtime": 9674.8125,
"train_samples": 50802,
"train_samples_per_second": 26.255,
"train_steps_per_second": 0.182
}