Closed dengdeng-cat closed 3 months ago
我尝试你了你命令是可以正常生成的。花屏基本是由于权重没有正确加载导致的,我觉得有两个可能:
命令如果换行,需要加 \
python scripts/inference.py configs/opensora-v1-2/inference/sample.py \
--num-frames 2s --resolution 360p \
--layernorm-kernel False --flash-attn False \
--prompt "a beautiful waterfall"
请问你有修改 config 吗?因为我们的权重应该都会自动下载。你运行的过程中有看到权重在下载的进度条吗?
我尝试你了你命令是可以正常生成的。花屏基本是由于权重没有正确加载导致的,我觉得有两个可能:
- 命令如果换行,需要加 \
python scripts/inference.py configs/opensora-v1-2/inference/sample.py \ --num-frames 2s --resolution 360p \ --layernorm-kernel False --flash-attn False \ --prompt "a beautiful waterfall"
- 请问你有修改 config 吗?因为我们的权重应该都会自动下载。你运行的过程中有看到权重在下载的进度条吗?
非常感谢您的回复!
对了,补充一点,我还将配置文件中的“df16”改成了“fp16”,因为我的机器不支持,这一点在1.1中验证是没有问题的。
对了,补充一点,我还将配置文件中的“df16”改成了“fp16”,因为我的机器不支持,这一点在1.1中验证是没有问题的。
同样的问题,我使用的也是fp16,怀疑是fp16的问题
We confirm this is a bug and is working on it.
The problem is fixed by PR #523
The problem is fixed by PR #523
Thanks.
I got a similar collapsed video using sample.sh
from stand vbench prompts. I checked
load_pretrained
for gradio/app.py and inference.py I got a similar collapsed video using
sample.sh
from stand vbench prompts. I checked
- the ckpt path, both hpcai/STDiT and my fine-tuned checkpoint by printing it during initialization
- using gradio/app.py to generate video with two ckpt which is all right. Do you have any ideas about this? I think the backbone of load ckpt are
load_pretrained
for gradio/app.py and inference.py
Sorry, this issue has been solved. That is because the num_frame in sample.sh should be passed 2s instead of 16.
Sorry, this issue has been solved. That is because the num_frame in sample.sh should be passed 2s instead of 16.
Hello, I'm also trying to use sample.sh provided by Open-Sora. Yet I'm having trouble in getting the ckpt path right.
I'm not sure if I need to download the pre-trained weights from huggingface beforehand.
Would you mind sharing your thoughts on it?
It would help a lot. Thanks in advance!
! 请问1.2的生成结果为花屏是什么原因,是哪一个库版本不对么? python scripts/inference.py configs/opensora-v1-2/inference/sample.py \ --num-frames 2s --resolution 360p \ --layernorm-kernel False --flash-attn False \ --prompt "a beautiful waterfall"