general_reasoner-step_rft_fixed / train_results.json
Renjie-Ranger's picture
Upload folder using huggingface_hub
1fe9217 verified
{
"epoch": 2.0,
"total_flos": 8.119097669378376e+17,
"train_loss": 0.558736775154458,
"train_runtime": 23106.9266,
"train_samples_per_second": 11.774,
"train_steps_per_second": 0.023
}