qwen2.5-0.5b-expo-DPO-ES-0.1 / train_results.json
hZzy's picture
Model save
88a4513 verified
raw
history blame
234 Bytes
{
"epoch": 1.5588096362777515,
"total_flos": 0.0,
"train_loss": 0.4483427975394509,
"train_runtime": 15087.4424,
"train_samples": 50802,
"train_samples_per_second": 16.836,
"train_steps_per_second": 0.117
}