Comments (34)
Hi guys. I encountered several errors and did workarounds and eventually got bvh file!
execute the following code at the top of "Process the video" section.
# 1. removes the duplicated 'log_dir' registration
!sed -i '59d' ../hmr/src/config.py
# 2. add None as the first auguments of tf.name_scope for these three files
!sed -i '40d' ../hmr/src/models.py
!sed -i '40i \ with tf.name_scope(None, "Encoder_resnet", [x]):' ../hmr/src/models.py
!sed -i '21d' ../hmr/src/tf_smpl/batch_lbs.py
!sed -i '21i \ with tf.name_scope(None, "batch_skew", [vec]):' ../hmr/src/tf_smpl/batch_lbs.py
!sed -i '107d' ../hmr/src/tf_smpl/batch_lbs.py
!sed -i '107i \ with tf.name_scope(None, "batch_forward_kinematics", [Rs, Js]):' ../hmr/src/tf_smpl/batch_lbs.py
# 3. cast the sorting key to string
!sed -i '194d' ../hmr/demo.py
!sed -i '194i \ all_files.sort(key=lambda x: str(x.split("/")[-1].split("/")[0]))' ../hmr/demo.py
Apparently from version 1.14 tensorflow comes with absl-py and it has the 'log_dir' flag already and collides with /hmr/src/config.py. Also after version 1.2 tf.name_scope started having different arguments and it throws error. This workaround works at least for now but not sure in the future.
Before processing actual target video, I recommend to try out with very short video with fps 1 and make sure all codes run through.
from video_to_bvh.
Provide more information, please.
from video_to_bvh.
I have installed the inital imports and uploaded my mp4 video file to 100%. I set the fps for 24 and i get an error for the video processing that says (FileNotFoundError: [Errno 2] No such file or directory: '2d_pose_estimation.py'). and for BVH download i get a message that says ( FileNotFoundError: Cannot find file: hmr/output/bvh_animation/estimated_animation.bvh). I dont know why im receiving the errors and if this is why i have not received the BVH file.
from video_to_bvh.
I missed a few errors after installing the initial imports. ERROR: Failed building wheel for opendr AND ERROR: Command "/usr/bin/python2 -u -c 'import setuptools, tokenize;file='"'"'/tmp/pip-install-FHu_g5/opendr/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(file);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' install --record /tmp/pip-record-weAs9R/install-record.txt --single-version-externally-managed --compile" failed with error code 1 in /tmp/pip-install-FHu_g5/opendr
from video_to_bvh.
There was a bug with opendr
version. That's already fixed. Are you sure you're using the last version?
from video_to_bvh.
I think so, Is there another version that you can show me?
from video_to_bvh.
Just follow the 1st step of usage from README
. Also follow other steps too. And check if your video is actually uploaded.
from video_to_bvh.
getting ERROR: Failed building wheel for opendr and ERROR: Command "/usr/bin/python2 -u -c 'import setuptools, tokenize;file='"'"'/tmp/pip-install-IM6ZRU/opendr/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(file);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code, file, '"'"'exec'"'"'))' install --record /tmp/pip-record-Xc5U23/install-record.txt --single-version-externally-managed --compile" failed with error code 1 in /tmp/pip-install-IM6ZRU/opendr/ ON THE FIRST STEP
from video_to_bvh.
Ah, that's not the case of your issue. That is the requirement that I forgot to remove, just ignore this error. There is another install of opendr required version - opendr-0.77
in the end of first cell - !pip2 install opendr==0.77
The last lines of 1st cell output should be:
Installing collected packages: opendr
Successfully installed opendr-0.77
So double check the video that you're uploading and correctness of following steps described in README
from video_to_bvh.
the upload of the video is 100% and i processed the video but it says sample_videos/NEYO: No such file or directory. where should i find the download BVH file once everything has worked properly? I dont think the video is converting to video images.
from video_to_bvh.
I've checked again - everything works fine. You're doing something wrong. Reset your Google Colab runtime and try again. Result .bvh should be in hmr/output/bvh_animation/
What video you're trying to process? Maybe the issue is in it's extension or something.
from video_to_bvh.
what is the maximum length or frames that can be uploaded .
from video_to_bvh.
I'm not sure about restriction in that case. Google Colab has 300+ GB of space for you.
from video_to_bvh.
I have made the video shorter and followed all the steps. I'm not sure why its not working.
from video_to_bvh.
its an mp4. once i finish the BVH step I see no BVH made no file nothing.
from video_to_bvh.
Maybe 2d pose estimation or hmr can't detect any human in frame. If you want, share the video so I could test it myself. Or upload another video with one person that should work for 100%.
from video_to_bvh.
ok, how can i share it with you?
from video_to_bvh.
https://wetransfer.com/ upload it here and share the link here
from video_to_bvh.
from video_to_bvh.
The problem is that you have space in your file name. Rename it to Mad_2820.mp4
and try again. Also reset your Colab runtime just in case.
from video_to_bvh.
THANKS! it worked and has started to process the video but now it says "GPU memory usage is close to the limit" I have even made the video shorter but its saying the same. sigh
from video_to_bvh.
Make the screenshot please.
from video_to_bvh.
from video_to_bvh.
Im not sure how to fix that but thats what it says. i have cut the video down to 10 seconds and its still saying memory is close to the limit.
from video_to_bvh.
Click IGNORE
, that's not an issue.
from video_to_bvh.
i tried to ignore it but it wont allow it.
from video_to_bvh.
Collab was running something around 3+ hours and I had no problems at all. Here is the estimated .bvh
: https://we.tl/t-XrVWPeBT7W That's not the full video though.
from video_to_bvh.
not sure why im having this much trouble.
from video_to_bvh.
Hi
@Dene33
Thanks for the awesome code
I have a similar problem @flip321
it worked for me but now it doesn't work
can i ask you to help me?
Thank you
from video_to_bvh.
Please read carefully through the thread and try to solve it. Or provide more information. Thanks.
from video_to_bvh.
@Dene33 Thank you for answering :)
I'm surprised that I had already run this code successfully!
But it's not currently running with the same video
I run the code with different videos but I did not succeed
I think these messages are not correct when the Process the video cell is running, that's why the BVH file is not downloaded
I hope you will help me :)
from video_to_bvh.
@Dene33 Thank you for answering :)
I'm surprised that I had already run this code successfully!
But it's not currently running with the same video
I run the code with different videos but I did not succeed
I think these messages are not correct when the Process the video cell is running, that's why the BVH file is not downloaded
I hope you will help me :)
I have the same question!Do you resolve it?
from video_to_bvh.
Hi guys. I encountered several errors and did workarounds and eventually got bvh file!
execute the following code at the top of "Process the video" section.
# 1. removes the duplicated 'log_dir' registration !sed -i '59d' ../hmr/src/config.py # 2. add None as the first auguments of tf.name_scope for these three files !sed -i '40d' ../hmr/src/models.py !sed -i '40i \ with tf.name_scope(None, "Encoder_resnet", [x]):' ../hmr/src/models.py !sed -i '21d' ../hmr/src/tf_smpl/batch_lbs.py !sed -i '21i \ with tf.name_scope(None, "batch_skew", [vec]):' ../hmr/src/tf_smpl/batch_lbs.py !sed -i '107d' ../hmr/src/tf_smpl/batch_lbs.py !sed -i '107i \ with tf.name_scope(None, "batch_forward_kinematics", [Rs, Js]):' ../hmr/src/tf_smpl/batch_lbs.py # 3. cast the sorting key to string !sed -i '194d' ../hmr/demo.py !sed -i '194i \ all_files.sort(key=lambda x: str(x.split("/")[-1].split("/")[0]))' ../hmr/demo.py
Apparently from version 1.14 tensorflow comes with absl-py and it has the 'log_dir' flag already and collides with /hmr/src/config.py. Also after version 1.2 tf.name_scope started having different arguments and it throws error. This workaround works at least for now but not sure in the future.
Before processing actual target video, I recommend to try out with very short video with fps 1 and make sure all codes run through.
Hi guys. I encountered several errors and did workarounds and eventually got bvh file!
execute the following code at the top of "Process the video" section.
# 1. removes the duplicated 'log_dir' registration !sed -i '59d' ../hmr/src/config.py # 2. add None as the first auguments of tf.name_scope for these three files !sed -i '40d' ../hmr/src/models.py !sed -i '40i \ with tf.name_scope(None, "Encoder_resnet", [x]):' ../hmr/src/models.py !sed -i '21d' ../hmr/src/tf_smpl/batch_lbs.py !sed -i '21i \ with tf.name_scope(None, "batch_skew", [vec]):' ../hmr/src/tf_smpl/batch_lbs.py !sed -i '107d' ../hmr/src/tf_smpl/batch_lbs.py !sed -i '107i \ with tf.name_scope(None, "batch_forward_kinematics", [Rs, Js]):' ../hmr/src/tf_smpl/batch_lbs.py # 3. cast the sorting key to string !sed -i '194d' ../hmr/demo.py !sed -i '194i \ all_files.sort(key=lambda x: str(x.split("/")[-1].split("/")[0]))' ../hmr/demo.py
Apparently from version 1.14 tensorflow comes with absl-py and it has the 'log_dir' flag already and collides with /hmr/src/config.py. Also after version 1.2 tf.name_scope started having different arguments and it throws error. This workaround works at least for now but not sure in the future.
Before processing actual target video, I recommend to try out with very short video with fps 1 and make sure all codes run through.
not work, already(
from video_to_bvh.
Ah, that's not the case of your issue. That is the requirement that I forgot to remove, just ignore this error. There is another install of opendr required version -
opendr-0.77
in the end of first cell -!pip2 install opendr==0.77
The last lines of 1st cell output should be:
Installing collected packages: opendr
Successfully installed opendr-0.77
So double check the video that you're uploading and correctness of following steps described in
README
cant install opendr
error: subprocess-exited-with-error
× python setup.py bdist_wheel did not run successfully.
│ exit code: 1
╰─> See above for output.
note: This error originates from a subprocess, and is likely not a problem with pip.
Building wheel for opendr (setup.py) ... error
ERROR: Failed building wheel for opendr
Running setup.py clean for opendr
Failed to build opendr
ERROR: Could not build wheels for opendr, which is required to install pyproject.toml-based projects
from video_to_bvh.
Related Issues (20)
- "Process the video" error HOT 1
- 3D mash model
- tensorflow.contrib.slim error in "Process the video" HOT 2
- in 2d_posse_estimation.py->model.predict : name 'model' is not defined HOT 1
- 'Object' object has no attribute 'select'
- How can I use blender to get motion of .bvh ?
- csv_joined key pionts type HOT 2
- FileNotFoundError Due to [Choose Files] Button the Upload Widget is not Available Even the Session is Rerun HOT 2
- ImportError: No module named contrib.slim HOT 4
- please delete it i am dumb
- Distortions in feet in bvh.
- specify version in requirements.txt HOT 1
- Process the video HOT 1
- fixed errors for beginners HOT 1
- hmr directory problem HOT 3
- How did you get the rotation of the joints using XYZ?
- cannot import batch normalization HOT 2
- No module tensorflow
- Can u change HMR old model to some SOTA model like PARE or 3DCrowdNet?
- Run error HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from video_to_bvh.