Ck773/llama-cpp

1https://github.com/ggerganov/llama.cpp/blob/master/examples/server/README.md2 3https://llama-cpp-python.readthedocs.io/en/latest/server/4 5```6curl --request GET \7    --url https://ck773-llama-cpp-lab.hf.space/v1/health8 9curl --request POST \10    --url https://ck773-llama-cpp-lab.hf.space/completion \11    --header "Content-Type: application/json" \12    --data '{"prompt": "Hello","n_predict": 128}'13```14 15 16python3 -m llama_cpp.server --config_file <config_file>17 18 19```json20{21    "host": "0.0.0.0",22    "port": 8080,23    "models": [24        {25            "model": "models/OpenHermes-2.5-Mistral-7B-GGUF/openhermes-2.5-mistral-7b.Q4_K_M.gguf",26            "model_alias": "gpt-3.5-turbo",27            "chat_format": "chatml",28            "n_gpu_layers": -1,29            "offload_kqv": true,30            "n_threads": 12,31            "n_batch": 512,32            "n_ctx": 204833        },34        {35            "model": "models/OpenHermes-2.5-Mistral-7B-GGUF/openhermes-2.5-mistral-7b.Q4_K_M.gguf",36            "model_alias": "gpt-4",37            "chat_format": "chatml",38            "n_gpu_layers": -1,39            "offload_kqv": true,40            "n_threads": 12,41            "n_batch": 512,42            "n_ctx": 204843        },44        {45            "model": "models/ggml_llava-v1.5-7b/ggml-model-q4_k.gguf",46            "model_alias": "gpt-4-vision-preview",47            "chat_format": "llava-1-5",48            "clip_model_path": "models/ggml_llava-v1.5-7b/mmproj-model-f16.gguf",49            "n_gpu_layers": -1,50            "offload_kqv": true,51            "n_threads": 12,52            "n_batch": 512,53            "n_ctx": 204854        },55        {56            "model": "models/mistral-7b-v0.1-GGUF/ggml-model-Q4_K.gguf",57            "model_alias": "text-davinci-003",58            "n_gpu_layers": -1,59            "offload_kqv": true,60            "n_threads": 12,61            "n_batch": 512,62            "n_ctx": 204863        },64        {65            "model": "models/replit-code-v1_5-3b-GGUF/replit-code-v1_5-3b.Q4_0.gguf",66            "model_alias": "copilot-codex",67            "n_gpu_layers": -1,68            "offload_kqv": true,69            "n_threads": 12,70            "n_batch": 1024,71            "n_ctx": 921672        }73    ]74}75```