training_log.json

{
  "base_model_name": "Yi-34B-GPTQ",
  "base_model_class": "YiForCausalLM",
  "base_loaded_in_4bit": false,
  "base_loaded_in_8bit": false,
  "projections": "gate, down, up, q, k, v, o",
  "loss": 1.1146,
  "learning_rate": 0.0002730941704035874,
  "epoch": 0.28,
  "current_steps": 5503,
  "train_runtime": 18904.2861,
  "train_samples_per_second": 3.03,
  "train_steps_per_second": 0.024,
  "total_flos": 8039607892967424.0,
  "train_loss": 1.2526764671007793
}