EfficientNet-PyTorch

EfficientNet-PyTorch copied to clipboard

Use It as a backbone for different purposes ?

8

I would like to use efficientnet as a backbone for object detection. However non of my attempts yield results. What do you suggest ? (In resnet implementation this code snippet...

ertugrulsmz

More issues with Gradcam

1

I've found this issue https://github.com/lukemelas/EfficientNet-PyTorch/issues/184 but it's about a checkboard effect. And it's not my case. I'm using GradCAM and GradCAM++ to EfficientNet-B0 layer model._blocks[-1]._project_conv And the results are weird,...

virilo

Is "Training is not implemented" out of date now?

1

Hi, there's this comment on top of the example: _Evaluate on ImageNet. Note that at the moment, training is not implemented (I am working on it). that being said, evaluation...

morawi

Can not use torch.jit.script to export model

10

torch.jit.frontend.NotSupportedError: Compiled functions can't take variable number of arguments or use keyword-only arguments with defaults: at /usr/local/lib/python3.7/dist-packages/torch/autograd/function.py:26:25 def mark_dirty(self, *args): ~~~~~

mathfinder

Cannot resolve operator 'Shape' with opsets: ai.onnx v9

1

I got this Error when I run the exported onnx model on the browser. `opset.ts:48 Uncaught (in promise) TypeError: cannot resolve operator 'Shape' with opsets: ai.onnx v9 ` **onnx :**...

amirdy

In examples/imagenet/main.py, the optimizer of training is set as SGD. But in the paper, the optimizer of training is set as RMSProp. I want to train EfficientNet but do not...

shenwenxiao

[Feature request] Add the Noisy Students pretrained weights

17

Hello @lukemelas , Thank's for your amazing work. Do you have any plans to incorporate the pretrained weights using [noisy students scheme](https://arxiv.org/abs/1911.04252)? Weights are available for the [TPU / Tensorflow](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet)...

Cyril9227

How could I construct a neck structure similar to unet by the output of some feature extraction layer (conv2D)

2

How could I construct a neck structure similar to unet by the output of some feature extraction layer (conv2D) my code :from efficientnet_pytorch import EfficientNet model = EfficientNet.from_name('efficientnet-b0') list(model.children())[2] I...

zzc-ai

How to get intermediate output not only the last features

2

#coding:utf8 import torch from efficientnet_pytorch import EfficientNet import torch.nn as nn class myModel(): def __init__(self): base_model = EfficientNet.from_pretrained(model_name="efficientnet-b5", weights_path="efficientnet-b5-b6417697.pth") #self.model = nn.Sequential(*list(base_model.children())[:2], *list(base_model.children())[2][:38]) self.model = nn.Sequential(*list(base_model.children())[:6]) self.model.eval() #print(self.model) def forword(self,...

liutianling

Start training from scratch on a new data set, without calling pre-training weights.

1

Hello, thank you very much for your code. I want to start from scratch on the new data set without calling the pre-training weights. May I ask what operation is...

lwzhaojun