Qwen2.5-PP-14B-Instruct / all_results.json
root
init
af30826
raw
history blame contribute delete
202 Bytes
{
"epoch": 2.976,
"total_flos": 17507936501760.0,
"train_loss": 0.28009381858251425,
"train_runtime": 885.0055,
"train_samples_per_second": 6.78,
"train_steps_per_second": 0.21
}