any use for arm64 MacOS Monterey on a M1 MAX Macbook? #294
Assignees
CalvinBackup (Calvin Backup)
Clear assignees
CalvinBackup (Calvin Backup)
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: CalvinBackup/neuralchen-SimSwap#294
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
gettin some error on:
python3 train.py --name simswap224_test --batchSize 8 --gpu_ids 0 --dataset /var/deep/DSC_0015.JPG Traceback (most recent call last): File "train.py", line 17, in <module> import numpy as np ModuleNotFoundError: No module named 'numpy'fixed on my system above and following errors with:
python3 -m pip install modulbut ehhh I got none Cuda gpu
script tells me:
Traceback (most recent call last): File "/Users/Smiril/deep/SimSwap/train.py", line 139, in <module> model.initialize(opt) File "/Users/Smiril/deep/SimSwap/models/projected_model.py", line 44, in initialize self.netG.cuda() File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 689, in cuda return self._apply(lambda t: t.cuda(device)) File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 579, in _apply module._apply(fn) File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 579, in _apply module._apply(fn) File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 602, in _apply param_applied = fn(param) File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 689, in <lambda> return self._apply(lambda t: t.cuda(device)) File "/opt/homebrew/lib/python3.9/site-packages/torch/cuda/__init__.py", line 211, in _lazy_init raise AssertionError("Torch not compiled with CUDA enabled") AssertionError: Torch not compiled with CUDA enabledanyone can help?
Try following the instructions here.
I'm not sure about onnxruntime and torch compatibility on a Mac, but you're going to have to do a bit of research for anything built for CUDA, and find the M1 builds in place of it.
thank you it helps a lot ..
the test on the instruction
work with 👍
conda create -n torch-gpu python=3.9 conda activate torch-gpu conda install pytorch torchvision torchaudio -c pytorch-nightly conda install -c conda-forge jupyter jupyterlab python3 -m pip install torch`import torch
import math
print(torch.backends.mps.is_available())
print(torch.backends.mps.is_built())
dtype = torch.float
device = torch.device("mps")
x = torch.linspace(-math.pi, math.pi, 2000, device=device, dtype=dtype)
y = torch.sin(x)
a = torch.randn((), device=device, dtype=dtype)
b = torch.randn((), device=device, dtype=dtype)
c = torch.randn((), device=device, dtype=dtype)
d = torch.randn((), device=device, dtype=dtype)
learning_rate = 1e-6
for t in range(2000):
y_pred = a + b * x + c * x ** 2 + d * x ** 3
print(f'Result: y = {a.item()} + {b.item()} x + {c.item()} x^2 + {d.item()} x^3')`
give following output ...
True True 99 569.908935546875 199 399.8172302246094 299 281.5032958984375 399 199.13519287109375 499 141.7447052001953 599 101.72549438476562 699 73.79811096191406 799 54.29474639892578 899 40.664642333984375 999 31.1326904296875 1099 24.46240234375 1199 19.79174041748047 1299 16.51934051513672 1399 14.225279808044434 1499 12.616241455078125 1599 11.487083435058594 1699 10.694307327270508 1799 10.137439727783203 1899 9.746109962463379 1999 9.471000671386719 Result: y = 0.026000667363405228 + 0.8636452555656433 x + -0.004485549405217171 x^2 + -0.09431249648332596 x^3Great! If you haven't already, reinstall Simswap using these instructions, but replace the
conda install pytorch...lines with your Mac instructions, and let us know if it works.replaced and work ... I'll made a pull request of two scripts bash and python3
the cuda error is every time popup
------------ Options ------------- Arc_path: arcface_model/arcface_checkpoint.tar Gdeep: False batchSize: 8000000000000 beta1: 0.0 checkpoints_dir: ./checkpoints continue_train: False dataset: crop_224/6.jpg gpu_ids: 0 isTrain: True lambda_feat: 10.0 lambda_id: 30.0 lambda_rec: 10.0 load_pretrain: ./checkpoints/simswap224_test log_frep: 200 lr: 0.0004 model_freq: 10000 name: simswap224_test niter: 10000 niter_decay: 10000 phase: train sample_freq: 1000 total_step: 1000000 use_tensorboard: False which_epoch: 10000 -------------- End ---------------- GPU used : 0 Traceback (most recent call last): File "/Users/pandora/Documents/deep/SimSwap/train.py", line 139, in <module> model.initialize(opt) File "/Users/pandora/Documents/deep/SimSwap/models/projected_model.py", line 44, in initialize self.netG.cuda() File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 689, in cuda return self._apply(lambda t: t.cuda(device)) File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 579, in _apply module._apply(fn) File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 579, in _apply module._apply(fn) File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 602, in _apply param_applied = fn(param) File "/opt/homebrew/lib/python3.9/site-packages/torch/nn/modules/module.py", line 689, in <lambda> return self._apply(lambda t: t.cuda(device)) File "/opt/homebrew/lib/python3.9/site-packages/torch/cuda/__init__.py", line 211, in _lazy_init raise AssertionError("Torch not compiled with CUDA enabled") AssertionError: Torch not compiled with CUDA enabledhello I use TensorFlow GPU script from my git fork m1_tf_setup.sh
what calculate with m1_tf_test.py
well work fine !!!!
thank you :)
Layer (type) Output Shape Param #
flatten (Flatten) (None, 784) 0
dense (Dense) (None, 128) 100480
dropout (Dropout) (None, 128) 0
dense_1 (Dense) (None, 10) 1290
dropout_1 (Dropout) (None, 10) 0
dense_2 (Dense) (None, 2) 22
=================================================================
Total params: 101,792
Trainable params: 101,792
Non-trainable params: 0
take a closer look on this:
2022-07-10 16:41:42.188712: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:305] Could not identify NUMA node of platform GPU ID 0, defaulting to 0. Your kernel may not have been built with NUMA support.
2022-07-10 16:41:42.188823: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:271] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 0 MB memory) -> physical PluggableDevice (device: 0, name: METAL, pci bus id: )
2022-07-10 16:41:42.315135: W tensorflow/core/platform/profile_utils/cpu_utils.cc:128] Failed to get CPU frequency: 0 Hz
Epoch 1/10
2022-07-10 16:41:42.480309: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.
1875/1875 [==============================] - 11s 5ms/step - loss: 0.1113 - accuracy: 0.1557
Epoch 2/10
1875/1875 [==============================] - 10s 5ms/step - loss: 0.0886 - accuracy: 0.1642
Epoch 3/10
1875/1875 [==============================] - 10s 6ms/step - loss: 0.0828 - accuracy: 0.1690
Epoch 4/10
1875/1875 [==============================] - 10s 5ms/step - loss: 0.0776 - accuracy: 0.1713
Epoch 5/10
1875/1875 [==============================] - 10s 5ms/step - loss: 0.0757 - accuracy: 0.1713
Epoch 6/10
1875/1875 [==============================] - 10s 5ms/step - loss: 0.0751 - accuracy: 0.1716
Epoch 7/10
1875/1875 [==============================] - 10s 6ms/step - loss: 0.0749 - accuracy: 0.1715
Epoch 8/10
1875/1875 [==============================] - 10s 5ms/step - loss: 0.0742 - accuracy: 0.1724
Epoch 9/10
1875/1875 [==============================] - 10s 5ms/step - loss: 0.0743 - accuracy: 0.1724
Epoch 10/10
1875/1875 [==============================] - 10s 5ms/step - loss: 0.0744 - accuracy: 0.1717
Model: "sequential"
Layer (type) Output Shape Param #
flatten (Flatten) (None, 784) 0
dense (Dense) (None, 128) 100480
dropout (Dropout) (None, 128) 0
dense_1 (Dense) (None, 10) 1290
dropout_1 (Dropout) (None, 10) 0
dense_2 (Dense) (None, 2) 22
dropout_2 (Dropout) (None, 2) 0
=================================================================
Total params: 101,792
Trainable params: 101,792
Non-trainable params: 0
unfortunly ... no luck ... but a error log
python3 test_one_image.py --name people --Arc_path arcface_model/arcface_checkpoint.tar --pic_a_path crop_224/3.jpg --pic_b_path crop_224/1.jpg --output_path ./ Traceback (most recent call last): File "/Users/Smiril/Documents/deep/SimSwap/test_one_image.py", line 30, in <module> opt = TestOptions().parse() File "/Users/Smiril/Documents/deep/SimSwap/options/base_options.py", line 84, in parse torch.cuda.set_device(self.opt.gpu_ids[0]) File "/Users/Smiril/opt/miniconda3/envs/torch-m1/lib/python3.10/site-packages/torch/cuda/__init__.py", line 314, in set_device torch._C._cuda_setDevice(device) AttributeError: module 'torch._C' has no attribute '_cuda_setDevice'should stand "mps" NOT "cuda" --->>> device = torch.device("mps")
possible to find with :
print(torch.backends.mps.is_available())Hi @Smiril, thanks for raising the issue and for the feedback. I am under M1 arm64 and I have the following error that is raised for the command line:
pip install insightface==0.2.1 onnxruntime moviepylikely due toonnxas @ExponentialML said too.UPDATE
I finally solved my problem and gave a solution for working with amr64 by #180
Did you find a way to run the command? Also is SimSwap executable with a CPU? Thanks a lot!
updated pull request
adding following:
# set gpu ids
if not torch.backends.mps.is_available():
if not torch.backends.mps.is_built():
if len(self.opt.gpu_ids) > 0:
torch.cuda.set_device(self.opt.gpu_ids[0])
else:
torch.device("mps")
else:
torch.device("mps")
else:
torch.device("mps")
updated pull request
running:
pip3 install opencv-python
pip3 install torch torchvision torchaudio onnx insightface==0.2.1 onnxruntime moviepy imageio imageio-ffmpeg
by adding :pull request
is the following error on:
python3 test_video_swapsingle.py --crop_size 224 --use_mask --name people --Arc_path arcface_model/arcface_checkpoint.tar --pic_a_path demo_file/Iron_man.jpg --video_path demo_file/multi_people_1080p.mp4 --output_path output/multi_test_swapsingle.mp4 --temp_path temp_resultsTraceback (most recent call last):
File "/Users/Smiril/Documents/deep/SimSwap/test_video_swapsingle.py", line 18, in
from options.test_options import TestOptions
File "/Users/Smiril/Documents/deep/SimSwap/options/test_options.py", line 11
class TestOptions(BaseOptions):
IndentationError: unexpected indent
want to run on "mps" the M1 GPU
@Smiril to run with
mps/cpuwith M1 chip, use the following command to install PyTorch:conda install pytorch torchvision -c pytorch-nightly, see here. After doing it, running with thecpuseems to have same (or better) performance thanmps(https://github.com/pytorch/pytorch/issues/77799) for the moment.