PyTorch
llama
alignment-handbook
Generated from Trainer
MambaInLlama_0_50 / train_results.json
Junxiong Wang
add models
a981c20
raw
history blame
219 Bytes
{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 0.4717116741438516,
"train_runtime": 20989.4384,
"train_samples": 133368,
"train_samples_per_second": 6.354,
"train_steps_per_second": 0.199
}