pytorch2keras

pytorch2keras copied to clipboard

From looking at the commit history, it looks like the optional argument was renamed from "names" to "name_policy," and the README still showed the optional arugment as "names."

balisujohn

Batch Normalization Layers not present in the converted Keras model

1

**Describe the bug** No batch normalization layers present in the converted keras model. **To Reproduce** __Snippet of your code__ **Expected behavior** A clear and concise description of what you expected...

rahman-mdatiqur

Does it enable me to convert PyTorch model (NCHW) to TFlite (NHWC)?

8

.

MaratZakirov

Add support for Conv1D

2

**Feature request** It would be great if you can add support for conv1D as well!

syomantak

ValueError: Operands could not be broadcast together with shapes , when converting with change_ordering

1

**Describe the bug** When loading a model and trying to convert it from pytorch to keras with change_ordering=True, I get an error. is there a way to resolve or figure...

shairoz-deci

**Describe the bug** I have a pth file for the SSD model which I wish to convert in Keras format, but I need to change the ordering (NCHW -> NHWC)....

yogeesh-alphaics

'Tensor' object has no attribute '_keras_shape'

4

``` # create pytorch model from pytorch2keras import pytorch_to_keras from torch.autograd import Variable from models.mobilenet import PoseEstimationWithMobileNet model = PoseEstimationWithMobileNet(num_class=10) # create input tensor import numpy as np import torch...

Guocode

"image_data_format": "channels_first",

1

Is this setting required?and is it also required when the transformed Keras model is loaded?

Gs928

Converted keras model has different parameters

1

The original PyTorch model has no bias (only weights) in certain Convo layers but the converted Keras model does have bias respectively. And the Keras model shows no BatchNorm2D layers...

mike-zyz

PS1TD