| { | |
| "bias": "none", | |
| "peft_type": "MIXLORA", | |
| "r": 16, | |
| "lora_alpha": 32, | |
| "lora_dropout": 0.05, | |
| "target_modules": [ | |
| "q_proj", | |
| "k_proj", | |
| "v_proj", | |
| "o_proj", | |
| "w1_proj", | |
| "w2_proj", | |
| "w3_proj" | |
| ], | |
| "routing_strategy": "mixtral", | |
| "num_experts": 8, | |
| "act_fn": "silu", | |
| "top_k": 2, | |
| "base_model_name_or_path": "/host_data/Llama-2-7b-hf", | |
| "task_type": "CAUSAL_LM" | |
| } |