Aaron Chung

萌新求问，只有llama.cpp才能量化吗？

5

如题，我想用hf transformers来推理，但是看到[手动模型合并与转换](https://github.com/ymcui/Chinese-LLaMA-Alpaca/wiki/%E6%89%8B%E5%8A%A8%E6%A8%A1%E5%9E%8B%E5%90%88%E5%B9%B6%E4%B8%8E%E8%BD%AC%E6%8D%A2)这里所言的步骤只有： 1. 转换原版llama 2. 合并lora 并没有量化步骤。请问量化在哪里做呢？ 另外有一点不懂，就是llama和alpaca这两个是一样的吗...，意思就是Chinese llama、Chinese alpaca都可以和原版llama合并lora权重？

[Usage]: Dev instructions for implementing new LoRA features on VLLM

1

What is the easiest way, when implementing new LoRA features on VLLM? For example, I want to modify the forward pass of LoRA model, ref to the forward pass in...

How to save the custom module into adapter_model.safetensrs when integrating new peft method

1

Just don't know where to save and load the module, or something can mark which module need to be saved. For example, we want a moe of lora, where multi-lora...

FEAT WANT: Dynamic quantization config for bitsandbytes

3

### Feature request Will transformers support dynamic quantization config for bitsandbytes? Currently transformers support hqq dynamic quantization, via ```python q4_config = {"nbits": 4, "group_size": 64} q8_config = {"nbits": 8, "group_size":...

问题：在显示屏没开启hdr的情况下，播放hdr杜比发灰是正常的吗？

16

用win11原生的媒体播放器播放的时候色彩比较艳丽，但是使用mpv就有点发灰，猜测可能是hdr模式，不知是否是这样，以及能否转到sdr模式呢？ 下图左侧为MPV，右侧为win11媒体播放器