| { | |
| "base_model_name": "Undi95_Meta-Llama-3-8B-hf", | |
| "base_model_class": "LlamaForCausalLM", | |
| "base_loaded_in_4bit": true, | |
| "base_loaded_in_8bit": false, | |
| "projections": "q, v", | |
| "loss": 2.2972, | |
| "grad_norm": 0.7305300831794739, | |
| "learning_rate": 0.0, | |
| "epoch": 4.82, | |
| "current_steps": 387, | |
| "train_runtime": 3885.0034, | |
| "train_samples_per_second": 0.425, | |
| "train_steps_per_second": 0.006, | |
| "total_flos": 1.478817225941975e+17, | |
| "train_loss": 2.4749738693237306 | |
| } |