hzzone / pytorch-openpose Goto Github PK

I'm especially interested in the 3D real-time single-person keypoint detection using 3D triangulation which is available in the original OpenPose Caffe. I read in the README that you used caffemodel2pytorch for converting this code from Caffe. Could you maybe give some more details on this? So I could also convert the 3D triangulation code from the original repo?

Thanks in advance!

File "demo_camera.py", line 15, in
print(f"Torch device: {torch.cuda.get_device_name()}")
TypeError: get_device_name() missing 1 required positional argument: 'device'

RuntimeError: Given groups=1, weight of size [256, 128, 3, 3], expected input[1, 3, 184, 488] to have 128 channels, but got 3 channels instead

hello, I'm wondering if this code can run in real-time when detecting both body and hand keypoints? processing one single image just costed around 9s on GPU when I try the demo.py, that's really time consuming~

Thank you guys for providing the PyTorch model.

I am trying to get it to work with Kinect using libfreenect2(Openkinect) on my desktop, but I get the error below.

Do you know of a way to resolve it? or
How do I change the settings to connect to Kinect?

(pytorch-openpose) User@computer:~/software/pytorch-openpose$ python demo_camera.py
Torch device: GeForce RTX 2080 Ti
[ WARN:0] global /io/opencv/modules/videoio/src/cap_v4l.cpp (887) open VIDEOIO(V4L2:/dev/video0): can't open camera by index
Traceback (most recent call last):
File "demo_camera.py", line 22, in
candidate, subset = body_estimation(oriImg)
File "/home/femifapo/software/pytorch-openpose/src/body.py", line 31, in call
multiplier = [x * boxsize / oriImg.shape[0] for x in scale_search]
File "/home/femifapo/software/pytorch-openpose/src/body.py", line 31, in
multiplier = [x * boxsize / oriImg.shape[0] for x in scale_search]
AttributeError: 'NoneType' object has no attribute 'shape'

The description seems to suggest that the hand pose detection is conditioned on the body pose detection. However, in the demo image, it shows a standalone detected hand. I guess my question is: given an image of only a hand, can the model detect it? Because the original OpenPose does not.

Hi, I am using demo.py to test on my image, then I got a candidate whose size is (21,4) , I just confused. How can I get the 18 body keypoint position? Thx.

Hello, I want to get a hand keypoint detector without body, is this paper---Hand keypoint detection in single images using multiview bootstrapping not open source? I want to know and learn how it is trained and view the model of the network, the openpose project seems to provide only the code to detect the body.
can you give me some guidance?Thany you very much.

I only got 18 points after run this demo, is there any method to get 25 points of each body?

Any idea what's happening?

SED-ML-0148% python demo1.py
Traceback (most recent call last):
File "demo1.py", line 17, in
candidate, subset = body_estimation(oriImg)
File "python/body.py", line 46, in call
Mconv7_stage6_L1, Mconv7_stage6_L2 = self.model(data)
File "/Users/ehartz01/Desktop/SignPhonologizer/nicey/lib/python2.7/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "python/model.py", line 108, in forward
out1 = self.model0(x)
File "/Users/ehartz01/Desktop/SignPhonologizer/nicey/lib/python2.7/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "/Users/ehartz01/Desktop/SignPhonologizer/nicey/lib/python2.7/site-packages/torch/nn/modules/container.py", line 92, in forward
input = module(input)
File "/Users/ehartz01/Desktop/SignPhonologizer/nicey/lib/python2.7/site-packages/torch/nn/modules/module.py", line 489, in call
result = self.forward(*input, **kwargs)
File "/Users/ehartz01/Desktop/SignPhonologizer/nicey/lib/python2.7/site-packages/torch/nn/modules/conv.py", line 320, in forward
self.padding, self.dilation, self.groups)
RuntimeError: Given groups=1, weight of size [256, 512, 3, 3], expected input[1, 3, 46, 32] to have 512 channels, but got 3 channels instead

Hi, Appreciations for such a nice work!!

I was looking this project and noticed that you didn't add any license to this project. I want to know whether this can be used for commercial projects or not.

I searched in official Openpose and found out that they don't offer commercial license for free.

Since your project is a derivative of their work, so it's owned by CMU based on the license of openpose. Could you look this issue from OpenPose repo for more detail?

Also, we can't use their pretrained model which follows the same license as their code, this is the issue, I'm referring to. Since you converted their Caffe model to PyTorch model, it's kind of derivative of their work only and as per their license, your work is CMU's work if used commercially and CMU's Openpose license apply by default.

So, this or any other implementation of Openpose can't be used commercially!

What do you think about this?

can you provide the model parameters of 'body_pose_model.pth'? Thank you so much.

Hello. Thanks for your excellent job. I can successfully in running the demo. But when I was running with another image, I found a ZeroDivisionError

Traceback (most recent call last):
File "/home/hkuit155/Documents/bottom_up/pytorch-openpose/pose_track_demo.py",
line 18, in a
candidate, subset = body_estimation(oriImg)
File "/home/hkuit155/Documents/bottom_up/pytorch-openpose/src/body.py",
line 131, in call
0.5 * oriImg.shape[0] / norm - 1, 0) ZeroDivisionError: float division by zero

If multiple bodies are detected, the subset seems to increase, which is normal, but the ID for each body part changes. For example, the left eye has the ID of 14 but for three skeletons instead of IDs being multiples of 14 they are put next to each other in the subset so they become 14,15,16. So, now the right eye does not correspond to its ID anymore. Is there a way around this problem, to make all the ID corresponds to the skeleton point even if multiple people are detected.

Traceback (most recent call last):
File "/home/moziaijason/pycode/detection/pytorch-openpose-master/src/hand.py", line 89, in
canvas = util.draw_handpose(oriImg, peaks, True)
File "/home/moziaijason/pycode/detection/pytorch-openpose-master/src/util.py", line 94, in draw_handpose
if np.sum(np.all(peaks[e], axis=1)==0)==0:
File "<array_function internals>", line 6, in all
File "/home/moziaijason/.local/lib/python3.6/site-packages/numpy/core/fromnumeric.py", line 2398, in all
return _wrapreduction(a, np.logical_and, 'all', axis, None, out, keepdims=keepdims)
File "/home/moziaijason/.local/lib/python3.6/site-packages/numpy/core/fromnumeric.py", line 90, in _wrapreduction
return ufunc.reduce(obj, axis, dtype, out, **passkwargs)
numpy.AxisError: axis 1 is out of bounds for array of dimension 1

util.py

for peaks in all_hand_peaks:
for ie, e in enumerate(edges):
if np.sum(np.all(peaks[e], axis=1)==0)==0:
print(peaks[e])
x1, y1 = peaks[e[0]]
x2, y2 = peaks[e[1]]
ax.plot([x1, x2], [y1, y2], color=matplotlib.colors.hsv_to_rgb([ie/float(len(edges)), 1.0, 1.0]))

当我运行hand.py的时候，出现上面的错误

当运行 python demo_video.py images/1.mp4 时， ffprobe(args.file)，这句会返回FileNotFoundError: [WinError 2] 系统找不到指定的文件，请问怎么解决

Hello，can you consider support foot joint point detection？ths

Are you planning on adding functionality for face keypoints as well? Or how might I go about adding this in?

Could I know the pytorch version of this project?

Thanks for your great work!

When I run the demo.py with a change of test_image = 'images/090932040.jpg', there is an error :

body.py:123: RuntimeWarning: invalid value encountered in true_divide
  vec = np.divide(vec, norm)
Traceback (most recent call last):
  File "demo.py", line 16, in <module>
    candidate, subset = body_estimation(oriImg)
  File "src\body.py", line 135, in __call__
    0.5 * oriImg.shape[0] / norm - 1, 0)
ZeroDivisionError: float division by zero

This is the image I used. Could you help me with the error?

First of all: thanks for your amazing work! I would like to use OpenPose, but can't run Caffe in my current setup/environment. This repo is a life-saver.

Now to the point: I got the following error when running the demo.py:
TypeError: can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.

I fixed this by changing the line in python/hand.py:
output = self.model(data).numpy()

to:
output = self.model(data).cpu().numpy()

As numpy can't work on stuff that is still on the GPU.. I'm not sure if this needs to be changed in the repo and if this is the best solution for this error? Thanks.

My environment specs are the following:
OS: CentOS Linux release 7.5.1804
CUDA version: 8.0
cuDNN version: 7.0.5
Torch version: 0.4.1
Python version: Python 3.6.6
GPU's: 4x GeForce RTX 2070

The full error stack:

Traceback (most recent call last):
  File "demo.py", line 31, in <module>
    peaks = hand_estimation(oriImg[y:y+w, x:x+w, :])
  File "python/hand.py", line 46, in __call__
    output = self.model(data).numpy()
TypeError: can't convert CUDA tensor to numpy. Use Tensor.cpu() to copy the tensor to host memory first.

Why not recognize gestures？

hands_list = util.handDetect(candidate, subset, oriImg)
print(hands_list)

Result
[]
[]
[]
How to solve？
Must be a complete human body structure (including the hand) to detect the hand?

I'm working on something that requires an OpenPose model that was pre-trained on MPII, was the converted model in this repo the MPII or COCO model?

If it's not, any recommendation for getting the MPII model up & running? I tried using caffemodel2pytorch on the trained OP model from this repo, but have been running into errors/issues

There is an inconsistency between your semantics and openpose semantics in the body pose estimation model could you please provide the semantic of each point?

Another inconsistency if a point is absent from the predictions openpose returns (0,0), you don't.

First of all, great work! Second of all, I have the following question:
I am running the demo script to do inference for the image that you have provided in the repo, which is called demo.png. While I can successfully get the result of pose estimation for the demo.png, I see that the inference takes about 0.7 sec using GPU.
Is that expected, or I am missing something here?
Thanks

Feel depressed to see the non-realtime-performance

RuntimeWarning: invalid value encountered in true_divide
vec = np.divide(vec, norm)
Traceback (most recent call last):
File "demo.py", line 16, in
candidate, subset = body_estimation(oriImg)
File "src\body.py", line 135, in call
0.5 * oriImg.shape[0] / norm - 1, 0)
ZeroDivisionError: float division by zero

I have run your code and compare with the official OpenPose and find some difference on hand estimation.
What will cause this?

This is really a great project. Now I have trouble in extracting bone image. How can I extract bone image

for example.I just want to get the coordinates of No.3 and No.4 point?What should I do?

Thanks for such a great work.
I have a question (may be off-topic), I'm trying to save the output video in mp4 format but I'm getting error
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x56294b5339a0]
moov atom not found pipe:: Invalid data found when processing input

mov format is working but mp4 is not working. Kindly suggest some solution.

Hi, guy, i see from task manager when i am runing demo_camera.py that the CUDA usage is only about 10%, another repo is 85%, so i think we can spped up the inference by making more usage of CUDA, but how ? Can you give me some inspiration ?

检测只会返回图片么？不会返回身体坐标数据？
https://www.paddlepaddle.org.cn/hubdetail?name=openpose_body_estimation&en_category=KeyPointDetection

Hello, I need 25 keypoints model, but I can not successfully run the conversion repo. So I hope you can upload the 25 keypoints model, thank you so much!

hello,
Thank you very much for sharing the code.
When I save the video processing results, there is a bug. The program （demo_camera.py）ends but the video is not saved. The output size will be 6KB or 8KB or other. At first, I thought it was a video coding format error, but it's not.
Finally, I found that the problem was that the size of the frame was changed after the hand model calculation. The input size changed from 720 × 1280 to 400 × 711, which led to the failure of cv2. Videowriter() to save the video.
Please verify and fix this bug.

Hello, I am trying to check wheather a finger keypoint is above the shoulder keypoint, how can I map a particual shoulder keypoint to the exact finger keypoint for this?

Your work is awesome. I want to get the candidate and subset values on the images, could you tell me how to get them? Thanks!

i run your code and get pose_25 result, i wonder how to get 18 keypoints?

Hey i am running into a very weird issue where i used the model to estimate poses from a video stream it works on the first frame but fails after that with this error..

label_img, label_numbers = label(binary, return_num=True, connectivity = binary.ndim)
TypeError: 'str' object is not callable

Any guidance you can provide on the most probable cause of the error would be very helpful thanks!

Hello bro.
I have used this repo, it's REALLY cool!
Now I wanna get the json files the Openpose output, could you please help me if you know how to realize?

Hello，
convert tool ：https://github.com/vadimkantorov/caffemodel2pytorch
caffemodel：(from https://github.com/CMU-Perceptual-Computing-Lab/openpose_train/tree/master/experimental_models)
http://posefs1.perception.cs.cmu.edu/OpenPose/models/pose/1_25BBkg/body_25b/pose_iter_XXXXXX.caffemodel

when I convert caffemodel ，what happen as follow:
Skipping layer [prelu4_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [prelu4_3_CPM, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [prelu4_4_CPM, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage0_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage0_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage0_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv1_stage0_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage0_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage0_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage0_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv2_stage0_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage0_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage0_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage0_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv3_stage0_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage0_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage0_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage0_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv4_stage0_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage0_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage0_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage0_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv5_stage0_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu6_stage0_L2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [concat_stage1_L2, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage1_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage1_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage1_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv1_stage1_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage1_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage1_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage1_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv2_stage1_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage1_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage1_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage1_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv3_stage1_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage1_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage1_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage1_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv4_stage1_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage1_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage1_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage1_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv5_stage1_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu6_stage1_L2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [concat_stage2_L2, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage2_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage2_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage2_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv1_stage2_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage2_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage2_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage2_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv2_stage2_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage2_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage2_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage2_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv3_stage2_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage2_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage2_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu4_stage2_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv4_stage2_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage2_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage2_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu5_stage2_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv5_stage2_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu6_stage2_L2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [concat_stage3_L2, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage3_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage3_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu1_stage3_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv1_stage3_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage3_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage3_L2_1, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu2_stage3_L2_2, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mconv2_stage3_L2_concat, Concat, Concat]: not found in caffemodel2pytorch.modules dict
Skipping layer [Mprelu3_stage3_L2_0, PReLU, PReLU]: not found in caffemodel2pytorch.modules dict

can you help me？thx!

This code is very nice. Thank you for your contribution. I just used your demo to test on coco2017 dataset, and I got the following results:
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets= 20 ] = 0.291
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets= 20 ] = 0.505
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets= 20 ] = 0.282
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.143
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.498
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 20 ] = 0.344
Average Recall (AR) @[ IoU=0.50 | area= all | maxDets= 20 ] = 0.530
Average Recall (AR) @[ IoU=0.75 | area= all | maxDets= 20 ] = 0.348
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets= 20 ] = 0.155
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets= 20 ] = 0.604
It seems so low. Could you tell me the mAP you test, or some test details? Thanks a lot.

hi， when run the body.py ,meet follow error
body.py:121: RuntimeWarning: invalid value encountered in true_divide
vec = np.divide(vec, norm)
Traceback (most recent call last):
File "body.py", line 217, in
candidate, subset = body_estimation(oriImg)
File "body.py", line 134, in call
0.5 * oriImg.shape[0] / norm - 1, 0)
ZeroDivisionError: float division by zero
I use python3.6 ,pytorch 0.4 .

also when i use pyton2.7 pytorch 0.40
Traceback (most recent call last):
File "body.py", line 218, in
candidate, subset = body_estimation(oriImg)
File "body.py", line 49, in call
Mconv7_stage6_L1, Mconv7_stage6_L2 = self.model(data)
File "/root/anaconda3/envs/caffe/lib/python2.7/site-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/root/dance/AIdance/pytorch-openpose/python/model.py", line 591, in forward
out1 = self.model0(x)
File "/root/anaconda3/envs/caffe/lib/python2.7/site-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/root/anaconda3/envs/caffe/lib/python2.7/site-packages/torch/nn/modules/container.py", line 91, in forward
input = module(input)
File "/root/anaconda3/envs/caffe/lib/python2.7/site-packages/torch/nn/modules/module.py", line 491, in call
result = self.forward(*input, **kwargs)
File "/root/anaconda3/envs/caffe/lib/python2.7/site-packages/torch/nn/modules/conv.py", line 301, in forward
self.padding, self.dilation, self.groups)
RuntimeError: Given groups=1, weight[256, 512, 3, 3], so expected input[1, 3, 46, 50] to have 512 channels, but got 3 channels instead

do you have any advice , thanks！

Now that you use the anaconda to manage the package.So what's the meaning that you you use the pip to get the package????

you only install the packages by pip into the system,not the virtual envs in the anaconda..

Hi, maybe you could help me with something or have found a solution already.

First of all, I'm using a different approach for actually detecting the hands and hand-regions in the image (that works faster than the OpenPose one), and I'm also using Keras+TF, not Pytorch, but that is irrelevant.

I've noticed, that the hand-keypoint detector works much better (i.e. gives much more accurate keypoints) for one hand (the right one, I think), but works far worse for the other one. Example below:

[image]

I assume that you're supposed to know beforehand (pardon the pun), which hand you're processing, and then flip the image horizontally, when necessary? Or is there some other trick to it?

Dose anyone got the same error as i do?I cant figure out how to solve it ,can anyone help me

Is there anyway to train the hand pose?

Is the performance too bad? It takes 1 - 2 seconds to predict a image on GTX 1070 8G. Is there a plan to optimize performance and increase the speed of predicit?