| { | |
| "name": "minicpm 2b 16k", | |
| "load_params": { | |
| "n_ctx": 16384, | |
| "n_batch": 1024, | |
| "rope_freq_base": 10000, | |
| "rope_freq_scale": 1, | |
| "n_gpu_layers": 1, | |
| "use_mlock": true, | |
| "main_gpu": 0, | |
| "tensor_split": [ | |
| 0 | |
| ], | |
| "seed": -1, | |
| "f16_kv": true, | |
| "use_mmap": true, | |
| "num_experts_used": 2 | |
| }, | |
| "inference_params": { | |
| "n_threads": 8, | |
| "n_predict": -1, | |
| "top_k": 1, | |
| "min_p": 0, | |
| "top_p": 0.1, | |
| "temp": 0.1, | |
| "repeat_penalty": 1.1, | |
| "input_prefix": "<用户>", | |
| "input_suffix": "<AI>", | |
| "antiprompt": [ | |
| "</s>", | |
| "<AI>" | |
| ], | |
| "pre_prompt": "", | |
| "pre_prompt_suffix": "", | |
| "pre_prompt_prefix": "", | |
| "seed": -1, | |
| "tfs_z": 1, | |
| "typical_p": 1, | |
| "repeat_last_n": 64, | |
| "frequency_penalty": 0, | |
| "presence_penalty": 0, | |
| "n_keep": 0, | |
| "logit_bias": {}, | |
| "mirostat": 0, | |
| "mirostat_tau": 5, | |
| "mirostat_eta": 0.1, | |
| "memory_f16": true, | |
| "multiline_input": false, | |
| "penalize_nl": true | |
| } | |
| } | |