Qwen2.5-7B-gen-dpo-2k / train_results.json
AmberYifan's picture
Model save
cdf3864 verified
{
"epoch": 0.992,
"total_flos": 0.0,
"train_loss": 0.7647512930054818,
"train_runtime": 925.9529,
"train_samples": 1992,
"train_samples_per_second": 2.151,
"train_steps_per_second": 0.067
}