-
您好,我用了你的ResNeXt-101数据集,50g左右,相比于grid-feats-vqa里的数据集小了很多,我能拿到您对应的ResNeXt-152数据集么
-
# Aggregated Residual Transformations for Deep Neural Networks #
- Author: Saining Xie, Ross Girshick, Piotr Dollár, Zhuowen Tu, Kaiming He
- Origin: https://arxiv.org/abs/1611.05431
- Related:
…
-
Hi.
I want to fine-tune resnext-101 on hmdb51_split1 with lr=0.001 and weight_decay=1e-5,but after 200 epoch, the eval only get 10% accuracy. I have checked train.log and val.log, the acc of train.lo…
-
Hello!
I debug the code and run the online_test.py on egogesture with your pretrained model successfully.
However the final result is poor.
**report/egogesture_resnext_101_Depth_32.pth, resnext,…
-
python 3dtrain.py --model pretrain_data/resnext-101-kinetics.pth --mode score --model_depth 101 --structure ResNeXt101 --batch_size 32 --model_name resnext --resnet_shortcut B --cropCsize 256 224 192 …
-
I run following command:
`python main.py --input ./input --video_root ./videos --output ./output.json --model_name resnext --model_depth 101 --model ./models/resnext-101-kinetics.pth --mode score`
…
-
@kenshohara, thanks for sharing this repo!
In your paper, you say that the `N` (number of blocks per layer) for `ResNext-101` is `[3, 24, 36, 3]`:
![Screenshot 2019-06-11 at 2 54 45 PM](https://us…
-
https://github.com/facebookresearch/ResNeXt
-
## 🚀 Feature
Support for various backbones like VGG, ResNeXT, etc.
## Motivation
Currently, only ResNet and FPN backbones are supported. It would be useful if other backbones like VGG Net or R…
-
what is "other_utils.resnet other_utils.resnext"? How can i get it?