Qwen-2.5-Math-7B-Max-v4-accuracy / train_results.json
chenggong
Model save
8dd73c2 verified
raw
history blame contribute delete
200 Bytes
{
"total_flos": 0.0,
"train_loss": 0.07567721740583536,
"train_runtime": 24349.6607,
"train_samples": 7500,
"train_samples_per_second": 0.924,
"train_steps_per_second": 0.007
}