运行helixfold_single_inference.py 报错

[xiong3134]

AttributeError: module 'paddle.fluid.core_avx.eager.ops' has no attribute 'fused_gate_attention

[LihangLiu]

Hi, which version of PaddlePaddle are you using? To run HelixFold-Single, you need to install PaddlePaddle dev as stated here.

[xiong3134]

i have installed dev 0.0.0. CPU version ,but is also "AttributeError: module 'paddle.fluid.core_avx.eager.ops' has no attribute 'fused_gate_attention", it must install 0.0.0 GPU version?

i have installed dev 0.0.0. CPU version ,but is also "AttributeError: module 'paddle.fluid.core_avx.eager.ops' has no attribute 'fused_gate_attention", it must install 0.0.0 GPU version?

[LihangLiu]

Yes, fused_gate_attention only support GPU version so far. As for the CPU version, we will contact PaddlePaddle team about this issue.

[sokrypton]

Any way to speed up this command?

pip install paddlepaddle-gpu==0.0.0.post112 -f https://www.paddlepaddle.org.cn/whl/linux/gpu/develop.html

I'm interested in developing a Google Colab notebook for helixfold-single and it appears a dev version installation is required for fused_gate_attention.

Unfortunately, it took over 30 minutes... :(

[LihangLiu]

Any way to speed up this command?

pip install paddlepaddle-gpu==0.0.0.post112 -f https://www.paddlepaddle.org.cn/whl/linux/gpu/develop.html

I'm interested in developing a Google Colab notebook for helixfold-single and it appears a dev version installation is required for fused_gate_attention.

Unfortunately, it took over 30 minutes... :(

Sorry, this seems to be an issue of PaddlePaddle Installation page... An alternative is to directly visit this link and download the xxx.whl file to the Colab (choose the correct xxx.whl according to your environment), and then use the following command to install:

pip install xxx.whl

[milkboylyf]

Yes, fused_gate_attention only support GPU version so far. As for the CPU version, we will contact PaddlePaddle team about this issue.

这个问题解决了吗？现在cpu版本能使用吗fused_gate_attention这个函数吗？

[LihangLiu]

Yes, fused_gate_attention only support GPU version so far. As for the CPU version, we will contact PaddlePaddle team about this issue.

这个问题解决了吗？现在cpu版本能使用吗fused_gate_attention这个函数吗？

你好，框架方面还没给出排期，我们也在跟进中，感谢您持续关注。