Yang Sun

求助 A6000 InternVL1.5 Chat 多卡推理出错

3

与官方安装不同点： 1. 由于服务器无法链接git，flash_attn使用了轮子flash_attn-2.3.6+cu118torch2.0cxx11abiFALSE-cp39-cp39-linux_x86_64.whl进行安装 2. 由于transformers-4.36.2 报错无法找到“InternLM2Tokenizer” 安装了 transformers-4.37.0 代码内容： ``` import torch from PIL import Image from transformers import AutoModel, CLIPImageProcessor from transformers import AutoTokenizer path = "/home/yangsun/checkpoint/InternVL-Chat-V1-5" # from...

**Describe** For my personal research, I would like to have the pre-training parameters for the BEiT3-base-indomain version of the decoder. Is there any place where I can download it?

Download Weights byModelScope

1

When I follow the link below, I get a request to download clip-vit-large-patch14-336. Should I download it separately offline? And how do I use it after downloading it offline? https://github.com/Q-Future/Q-Instruct/tree/main/model_zoo#modelscope...

The output has a value range of 0.0 to 255.0 and a similar shape. Input may be 576*768, 576*1024 and other variable shapes. The shortest edge length is 576 Thank...