Qwen-2.5-3B-Simple-RL / train_results.json
jingluo's picture
Model save
24cd100 verified
raw
history blame contribute delete
200 Bytes
{
"total_flos": 0.0,
"train_loss": 0.005782647762212089,
"train_runtime": 41644.7787,
"train_samples": 7500,
"train_samples_per_second": 0.54,
"train_steps_per_second": 0.004
}