qwen2.5_0.5b_1M_stack_16kcw_2ep / train_results.json
ahmedheakl's picture
End of training
7e5216c verified
raw
history blame contribute delete
227 Bytes
{
"epoch": 1.9999935863106213,
"total_flos": 1.0432719577686213e+19,
"train_loss": 0.004895635656955134,
"train_runtime": 187864.5289,
"train_samples_per_second": 13.279,
"train_steps_per_second": 1.66
}