Qwen-2.5-1.5B-Instruct-Datamix / eval_results.json
andrewsiah's picture
End of training
e02586f verified
raw
history blame
166 Bytes
{
"eval_loss": 1.040108561515808,
"eval_runtime": 203.3075,
"eval_samples": 700,
"eval_samples_per_second": 3.443,
"eval_steps_per_second": 0.02
}