NuminaMath-7B-TIR / train_results.json
lewtun's picture
lewtun HF staff
Add AI-MO/deepseek-math-7b-sft-aimo_v51.2 checkpoint
167ee6b verified
raw
history blame contribute delete
No virus
232 Bytes
{
"epoch": 4.0,
"total_flos": 340525408321536.0,
"train_loss": 0.3252074327502578,
"train_runtime": 5343.4474,
"train_samples": 73061,
"train_samples_per_second": 41.505,
"train_steps_per_second": 1.297
}