tttx
/

PEFT
Safetensors
qwen2
alignment-handbook
trl
sft
Generated from Trainer
qwen-32b-sft-best / train_results.json
aadityap's picture
Initial commit from sft_r1_32b revision beac96e4091b2661eb90814b64cbdd09f0deb04d
42cafdb verified
raw
history blame contribute delete
237 Bytes
{
"epoch": 4.0,
"total_flos": 1.922993650610995e+16,
"train_loss": 0.38463607464530847,
"train_runtime": 27029.4974,
"train_samples": 10466,
"train_samples_per_second": 1.549,
"train_steps_per_second": 0.049
}