删除 --cache-type-k 和 --cache-type-v 参数，让 llama.cpp 使用默认的 f16 格式

Try Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q5_K_P Dockfile --n-gpu-layers 99

Try Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q5_K_P Dockfile

Try Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q5_K_P

Update pinned: true

Update README.md

Update removal of append_history=False

Update Waking-up-takes-about-9min-Sleep-after-15min-of-inactivity

Update overwrite libbitsandbytes_cpu.so with libbitsandbytes_cuda118.so lsb_release -a bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update overwrite libbitsandbytes_cpu.so with libbitsandbytes_cuda118.so lsb_release -a bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update overwrite libbitsandbytes_cpu.so with libbitsandbytes_cuda118.so lsb_release -a bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update lsb_release -a bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update lsb_release -a bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update lsb_release -a bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update lsb_release -a bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update lsb_release -a bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update bitsandbytes device_map={'': 0} rich, check nvidia-smi with subproces.run

Update bitsandbytes-cuda113 device_map={'': 0} rich, check nvidia-smi with subproces.run

Update device_map={'': 0} rich, check nvidia-smi with subproces.run

Update device_map={'': 0} rich, check nvidia-smi with subproces.run

Update device_map={'': 0} rich, check nvidia-smi with subproces.run

Update rich, check nvidia-smi with subproces.run

Update check nvidia-smi with subproces.run

Update cudatoolkit in requirements.txt

Update transformers requirements.txt

Fix torch==2.0.1 in requirements.txt

Fix bitsandbytes==0.39.0 in requirements.txt

Update git+https://github.com/TimDettmers/bitsandbytes.git in requirements.txt

Update requirements.txt https://github.com/TimDettmers/bitsandbytes

Remove packages.txt

Update packages.txt with cuda、cuda-tookit

Update bitsandbytes-cuda116

Back to Qwen/Qwen-7B-Chat

Update tangger/Qwen-7B-Chat since Qwen/Qwen-7B-Chat is gone

Update

Update

Update

Update

Update pundkey='#' {poundkey} in info

Update API ready, TODO: fix info

Update logger.trace disabled

Update yield from bot_stream in bot_stream_state

Update turn trace on

Update info on top_k=0

Update user_clear for submit.click buttn

Update model.generation_config.update

Update single model

Update model replaced by stats.value.llm