```You must select at least one handle point and target point.`` Should I read the paper first? about draggan

[Zeqiang-Lai]

Ok, I see.

Try this to use GPU 1

export CUDA_VISIBLE_DEVICES=1
python gradio_app.py

[kanseaveg]

still get error.

[Zeqiang-Lai]

It is not necessary. You have to click two points at least. Blue: handle point, Red: target point, like this.

Then click drag it, the model would drag blue point towards red point.

BTW: Do you have any suggestion on some better error message ? Does handle point confuse you ?

[kanseaveg]

File "/home/amax/euan/code/draggan/drag_gan.py", line 173, in drag_gan
F0 = F.detach().clone()
RuntimeError: CUDA out of memory. Tried to allocate 512.00 MiB (GPU 0; 23.69 GiB total capacity; 4.47 GiB already allocated; 150.94 MiB free; 4.99 GiB reserved in total by PyTorch)

I think maybe it's out of memory. lol~

how can i deploy it on my server. I have four 3090 gpus.

[Zeqiang-Lai]

It can be deployed with about 9 GB GPU memory.

[kanseaveg]

still got error. lol.

i think it's out of memory now

Traceback (most recent call last):
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/gradio/routes.py", line 421, in run_predict
event_data=event_data,
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/gradio/blocks.py", line 1321, in process_api
fn_index, inputs, iterator, request, event_id, event_data
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/gradio/blocks.py", line 1064, in call_function
prediction = await utils.async_iteration(iterator)
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/gradio/utils.py", line 514, in async_iteration
return await iterator.anext()
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/gradio/utils.py", line 508, in anext
run_sync_iterator_async, self.iterator, limiter=self.limiter
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/anyio/to_thread.py", line 32, in run_sync
func, *args, cancellable=cancellable, limiter=limiter
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/anyio/_backends/_asyncio.py", line 937, in run_sync_in_worker_thread return await future
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/anyio/_backends/_asyncio.py", line 867, in run
result = context.run(func, *args)
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/gradio/utils.py", line 490, in run_sync_iterator_async
return next(iterator)
File "gradio_app.py", line 103, in on_drag
max_iters=max_iters):
File "/home/amax/euan/code/draggan/drag_gan.py", line 182, in drag_gan
sample2, F2 = g_ema.generate(latent, noise)
File "/home/amax/euan/code/draggan/drag_gan.py", line 107, in generate
out = conv1(out, latent[:, i], noise=noise1)
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/amax/euan/code/draggan/stylegan2/model.py", line 358, in forward
out = self.conv(input, style)
File "/home/amax/miniconda3/envs/cyy/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "/home/amax/euan/code/draggan/stylegan2/model.py", line 275, in forward
input, weight, padding=0, stride=2, groups=batch
File "/home/amax/euan/code/draggan/stylegan2/op/conv2d_gradfix.py", line 64, in conv_transpose2d
).apply(input, weight, bias)
File "/home/amax/euan/code/draggan/stylegan2/op/conv2d_gradfix.py", line 146, in forward
**common_kwargs,
RuntimeError: CUDA out of memory. Tried to allocate 130.00 MiB (GPU 0; 23.69 GiB total capacity; 4.85 GiB already allocated; 120.94 MiB free; 5.02 GiB reserved in total by PyTorch)

[Zeqiang-Lai]

Could you show the output of nvidia-smi ?

I think it might be lack of enough memory of your server

[kanseaveg]

if __name__ == '__main__':
    demo = main()
    demo = demo.queue(concurrency_count=1, max_size=20).launch(share=True, server_name='10.xx.xx.239', port=6666)

i just modify the last line to deploy it on my own server which has 4x3090 gpus.

But it got into touble.

nvidia-smi

(cyy) amax@admin:~/euan/code/draggan$ nvidia-smi
Sun May 21 20:33:53 2023
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.105.17 Driver Version: 525.105.17 CUDA Version: 12.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 NVIDIA GeForce ... On | 00000000:18:00.0 Off | N/A |
| 39% 28C P8 19W / 350W | 17580MiB / 24576MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
| 1 NVIDIA GeForce ... On | 00000000:3B:00.0 Off | N/A |
| 39% 29C P8 26W / 350W | 2MiB / 24576MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
| 2 NVIDIA GeForce ... On | 00000000:86:00.0 Off | N/A |
| 42% 27C P8 23W / 350W | 2MiB / 24576MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
| 3 NVIDIA GeForce ... On | 00000000:AF:00.0 Off | N/A |
| 30% 28C P8 15W / 350W | 2MiB / 24576MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 4410 C python 8758MiB |
| 0 N/A N/A 18275 C python 8820MiB |
+-----------------------------------------------------------------------------+

[kanseaveg]

I would like to ask, is this Gradio service running on a remote server or on a local PyTorch

[kanseaveg]

I would like to ask, is this Gradio service running on a remote server or on a local PyTorch

Or to put it another way, Gradio service runs on the remote interface provided by Gradio, but the graphics card used is local

[Zeqiang-Lai]

Well, I guess you want a shareable link? If it is, once you have launch the service via python gradio_app.py. You will get a link similar as

Running on local URL:  http://127.0.0.1:7860
Running on public URL: https://bf5e8576f09a6582f7.gradio.live

This share link expires in 72 hours. For free permanent hosting and GPU upgrades (NEW!), check out Spaces: https://huggingface.co/spaces

https://bf5e8576f09a6582f7.gradio.live can be accessed anywhere and use your local GPU

[kanseaveg]

it works now.!!!!!

Thank you!!!

My dear friends.

[Zeqiang-Lai]

Cool, you are welcome

[kanseaveg]

it 's now smile. LOL. Thank you my friend. I forget to switch my gpu.