moabitcoin ig65m-pytorch issues

moabitcoin / ig65m-pytorch

PyTorch 3D video classification models pre-trained on 65 million Instagram videos

MIT License

265 stars 30 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

other pretrained models?

#44 dragen1860 opened 2 years ago
3
Update project to recent stable dependencies: pytorch, cuda, pip deps

#43 daniel-j-h opened 3 years ago
1
can I download IG65M dataset?

#42 pritamqu closed 3 years ago
2
about extract features

#41 kimsimple closed 3 years ago
1
Citation information

#40 Farabi-shafkat closed 3 years ago
3
r2plus1d_34_8_ig65m with 16 frames input

#39 jaypatravali closed 2 years ago
5
fine tune the pre-trained model on UCF101

#38 Yueeeeee-1 opened 4 years ago
4
How can I fine-tune the whole model using the torch-hub ?

#37 shamanez closed 4 years ago
2
Something wrong when run code : torch.hub.list("moabitcoin/ig65m-pytorch")

#36 huamgmin closed 4 years ago
4
are there any preprocessing to the input video clips?

#35 FesianXu closed 4 years ago
10
Need help with prediction codes

#34 wuvei closed 4 years ago
1
using R(2+1)D features extraction instead of C3D

#33 arash9900 closed 4 years ago
4
Script to convert caffe to pytorch

#32 sanchit2843 closed 4 years ago
2
Channel mismatched with README.md

#31 YunseokJANG closed 4 years ago
3
Match the num_class in PyTorch Hub with Models

#30 YunseokJANG closed 4 years ago
1
Adds video GAN framework

#29 daniel-j-h opened 5 years ago
4
The models 'r2plus1d_34_clip32_ft_kinetics_from_ig65m' fine-tuned on UCF101 can't get good performance, just 78%, why?

#28 fwcheng closed 5 years ago
14
Adds approximate nearest neighbor similarity searching tools

#27 daniel-j-h closed 5 years ago
2
Adds flake8 integration

#26 daniel-j-h closed 5 years ago
0
Disables grads in extract tool

#25 daniel-j-h closed 5 years ago
0
Upgrades pytorch and torchvision to stable release

#24 daniel-j-h closed 5 years ago
0
Adds hubconf.py for pytorch hub integration, resolves #4

#23 daniel-j-h closed 5 years ago
0
Thanks!

#22 albanie closed 5 years ago
1
Builds and publishes Docker images from Travis, resolves #14

#21 daniel-j-h closed 5 years ago
2
Run integration tests on Travis

#20 daniel-j-h closed 5 years ago
2
Do video models dream of electrical sheeps in motion

#19 daniel-j-h closed 5 years ago
11
Visualize features for all clips in a video "semantic bar codes"

#18 daniel-j-h closed 5 years ago
8
Adds easy to use and adapt feature extraction tool, resolves #6

#17 daniel-j-h closed 5 years ago
0
Adds Travis CI integration, resolves #13

#16 daniel-j-h closed 5 years ago
1
Refactors into proper project with cli tools

#15 daniel-j-h closed 5 years ago
1
Let CI build docker images

#14 daniel-j-h closed 5 years ago
0
Add Travis CI integration

#13 daniel-j-h closed 5 years ago
1
Fix released weight asset filenames and broken pb file

#12 daniel-j-h closed 5 years ago
3
Adds GPU Dockerfile, convenience Makefile, resolves #1

#11 daniel-j-h closed 5 years ago
0
Support r2plus1d_34_clip32_ig65m_from_scratch model

#10 sandhawalia closed 5 years ago
0
Provide CSN models and weights

#9 daniel-j-h opened 5 years ago
2
Provide r(2+1)d 152-layer models and weights

#8 daniel-j-h opened 5 years ago
1
Provide convenient fine-tuning tool for users

#7 daniel-j-h opened 5 years ago
2
Provide convenient and efficient feature extraction tool for users

#6 daniel-j-h closed 5 years ago
0
Upstream our changes to PyTorch r(2+1)d architecture

#5 daniel-j-h closed 5 years ago
3
Look into PyTorch Hub integration

#4 daniel-j-h closed 5 years ago
0
Port missing r(2+1)d 34-layer IG65-M model

#3 daniel-j-h closed 5 years ago
3
Run on full Kinetics-400 dataset to verify accuracy claims

#2 daniel-j-h opened 5 years ago
7
Provide nvidia-docker based Dockerfile

#1 daniel-j-h closed 5 years ago
0