xincheng Ju

Results 9 comments of xincheng Ju

I also have the same problem. I used one 4090 and it needs about 3s for only one sample

![image](https://github.com/PKU-YuanGroup/Video-LLaVA/assets/38723604/78a03042-003d-4a2f-b324-b3343eeb3960)

After finetuing_lora.sh, I get some file in checkpoint ![image](https://github.com/PKU-YuanGroup/Video-LLaVA/assets/38723604/c5c6c5ce-1dbd-4216-b8f2-4a7b14c7ff32) How can I use this checkpoint to infer or eval ? I want to use this new finetuing model to infer...

![image](https://github.com/PKU-YuanGroup/Video-LLaVA/assets/38723604/eb666b74-3ff6-4be3-90a2-bc21924d8a33)

数据集可以选择huggingface中的某一组,valley的可以只下一部分,然后打开用于测试

![image](https://github.com/PKU-YuanGroup/Video-LLaVA/assets/38723604/69aef432-6443-41ca-8255-b7de4fb00d83) ![image](https://github.com/PKU-YuanGroup/Video-LLaVA/assets/38723604/4578826f-b33a-45ac-a422-e3e0ea8ddf0c) 你会下载这两个文件吧。然后解压下valley.json文件。 然后你写个代码把有视频的挑出来。

你这个是视频文件地方。 json也有个下载的地方的。 ![image](https://github.com/PKU-YuanGroup/Video-LLaVA/assets/38723604/4341ad3e-4d2a-4684-9069-af3c0d9ebee4)

> @cm-xcju 兄弟,可以麻烦您给一个环境的各种包的版本吗,我在安装deepspeed的时候,发现他和python=3.10和torch 2.0.1有很多冲突,想参考一下你的环境。感谢感谢🙏,期待回复 有个environment文件,你按照这个就行

你好,可以问下,在加载模型时遇到的特征大小不匹配得情况是如何解决的 Uploading 1.png… 你看下使用的模块是否对应。 还有就是multimodal_encoder中的build文件貌似要修改,我记得曾经改过。 ![image](https://github.com/PKU-YuanGroup/Video-LLaVA/assets/38723604/dc0b2f76-b527-47c7-9a31-f6a0bb9811d6)