Files
qwen3-moe-neuron_structure_…/train_results.json

8 lines
211 B
JSON
Raw Permalink Normal View History

{
"total_flos": 63845040947200.0,
"train_loss": 0.868978128804789,
"train_runtime": 18382.0346,
"train_samples": 1000,
"train_samples_per_second": 0.333,
"train_steps_per_second": 0.021
}