qywu / TorchFly

A PyTorch Deep Learning Kit
MIT License
12 stars 5 forks source link

clone apex #5

Open littletree-Y opened 5 months ago

littletree-Y commented 5 months ago

i notice that this project needs "git clone https://github.com/qywu/apex", but apex may not be public now. is that right? Could you provide some help? many thanks!

qywu commented 5 months ago

You can clone this repo instead: https://github.com/NVIDIA/apex. There might be errors during compilation. You can try to fix it by changing the lines in https://github.com/NVIDIA/apex/blob/master/setup.py#L32.

littletree-Y commented 5 months ago

You can clone this repo instead: https://github.com/NVIDIA/apex. There might be errors during compilation. You can try to fix it by changing the lines in https://github.com/NVIDIA/apex/blob/master/setup.py#L32.

Thank you for your reply. I noticed that Torchfly uses Apex's amp and other modules, but NVIDIA's amp in Apex is outdated. Dealing with these dependencies is a tricky task. If your Apex repository is still available, could you please provide it? It will help me a lot. Thank you very much

qywu commented 5 months ago

I modified the code at L32 in this file https://github.com/NVIDIA/apex/blob/master/setup.py#L32. https://github.com/qywu/apex. you can check if it works.

On Mon, Jan 22, 2024 at 11:49 AM littletree @.***> wrote:

You can clone this repo instead: https://github.com/NVIDIA/apex. There might be errors during compilation. You can try to fix it by changing the lines in https://github.com/NVIDIA/apex/blob/master/setup.py#L32.

Thank you for your reply. I noticed that Torchfly uses Apex's amp and other modules, but NVIDIA's amp in Apex is outdated. Dealing with these dependencies is a tricky task. If your Apex repository is still available, could you please provide it? It will help me a lot. Thank you very much

— Reply to this email directly, view it on GitHub https://github.com/qywu/TorchFly/issues/5#issuecomment-1904405098, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEK2IFSL57OKAG2M53TU5JTYP2KB7AVCNFSM6AAAAABCFQNPUSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBUGQYDKMBZHA . You are receiving this because you commented.Message ID: @.***>

littletree-Y commented 5 months ago

I modified the code at L32 in this file https://github.com/NVIDIA/apex/blob/master/setup.py#L32. https://github.com/qywu/apex. you can check if it works. On Mon, Jan 22, 2024 at 11:49 AM littletree @.> wrote: You can clone this repo instead: https://github.com/NVIDIA/apex. There might be errors during compilation. You can try to fix it by changing the lines in https://github.com/NVIDIA/apex/blob/master/setup.py#L32. Thank you for your reply. I noticed that Torchfly uses Apex's amp and other modules, but NVIDIA's amp in Apex is outdated. Dealing with these dependencies is a tricky task. If your Apex repository is still available, could you please provide it? It will help me a lot. Thank you very much — Reply to this email directly, view it on GitHub <#5 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEK2IFSL57OKAG2M53TU5JTYP2KB7AVCNFSM6AAAAABCFQNPUSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBUGQYDKMBZHA . You are receiving this because you commented.Message ID: @.>

thanks a lot! i installed apex successfully