zephyr-7b-dpo-lora / train_results.json
sambar's picture
Model save
d214789
raw
history blame contribute delete
195 Bytes
{
"epoch": 3.0,
"train_loss": 0.6037390495627379,
"train_runtime": 33041.7964,
"train_samples": 61966,
"train_samples_per_second": 5.626,
"train_steps_per_second": 0.022
}