hnsywangxin

Results 41 comments of hnsywangxin

test result on Apollo depth dataset

hello, @passion3394 , the apollo dataset only has depth map,converting it to a disparity map requires the camera's focal length and baseline, but I didn't find these values.　can you...

test result on Apollo depth dataset

@passion3394 I want to know how is　the model of you trained generalization ability

why does the testing process need more GPU memory than the training process

hello, @JiaRenChang my computure will be reboot when I run your program,I was so confused,could you give me some advise? thank you

[Bug] TTS Python no longer working

same error

There is no offsets.txt file

the video can not too short，large than 100 frames

微调后的模型怎么放入web_demo中进行使用啊？

> > 这里：https://github.com/THUDM/VisualGLM-6B/blob/main/model/infer_util.py#L21 > > 同时还要在web_demo.py的main函数之前`from finetune_visualglm import FineTuneVisualGLMModel` > > > 改了下还是不行 > > ``` > For torch.distributed users or loading model parallel models, set environment variables RANK, WORLD_SIZE...

After redis reconnects, the worker can not consume

> I would request you to use the latest stable release of celery==5.3.1 atleast I have tried v5.3.1, but the same error.

After redis reconnects, the worker can not consume

> which redis-py version you are using? I only have redis which installed by pip, and version is 4.6.0

After redis reconnects, the worker can not consume

> Celery [v5.4.0rc1](https://github.com/celery/celery/releases/tag/v5.4.0rc1) is ready for testing! Please report if this issue is resolved! 🙏 ok，thank you

大佬们，请问一下，为什么我按照官方教程步骤训练，我用的CMRL数据集，训练syncnet模型、训练wav2lip模型，最后生成的模型唇形一点都对不上

> > > 听说CMLR数据集，视频和音频对不上，得自己清洗一下，或者过滤一下 > > > > > > wav2lip对中文太不友好了，只能重新训练 > > 我加载了它给的英文训练好的 sync_lip 权重，然后在 CMLR 上训练，loss 降的很慢，6500 step 的 loss 才到 0.7 附近，现在拿来训练 wav2lip 试试看，效果不好的话，我就不加载预训练好的权重，直接从零在 CMLR 上训练 sycn_lip...