Rafikg

apply_parallel takes forever

1

## General - **Operating System**: Ubuntu with 48 cpus and 350G of memory - **Python version**:3.8.10 - **Pandas version**:1.4.0 - **Pandarallel version**:1.5.7 ## Acknowledgement - [ ] My issue is...

Understanding the Metadata

1

How to much the metadata with train/test dataset ?

@yunjey I am running your example and I could show the tensorboard but the graph does not saved?

Hi @aladdinpersson having num_layers =2 > 1, break this line and generate **RuntimeError: Expected hidden[0] size (2, 1, 1024), got [1, 1, 1024]** https://github.com/aladdinpersson/Machine-Learning-Collection/blob/ac5dcd03a40a08a8af7e1a67ade37f28cf88db43/ML/Pytorch/more_advanced/Seq2Seq_attention/utils.py#L37

keypoints out off the crop

2

Hi @michalfaber https://github.com/michalfaber/keras_Realtime_Multi-Person_Pose_Estimation/blob/b595cfbb35dffe5abe2cb4cf6a1bde4d0986d125/training/augmentors.py#L207-L210 Here I think that you have to check if the keypoints are inside or outside the crop. I see that you checked the negative coordinates inside the...

Why you take only one person

1

Hi @michalfaber https://github.com/michalfaber/keras_Realtime_Multi-Person_Pose_Estimation/blob/b595cfbb35dffe5abe2cb4cf6a1bde4d0986d125/training/dataflow.py#L245-L250 I am not sure here that I understand what you were doing. You process all the persons in each image and after some filters (areas, number of...

What is this value coming from

2

https://github.com/michalfaber/keras_Realtime_Multi-Person_Pose_Estimation/blob/b595cfbb35dffe5abe2cb4cf6a1bde4d0986d125/training/train_pose.py#L27 ?

Masked the outputs

2

Thanks @michalfaber, I am wondering if still make sense to mask the outputs if we have already masks the input image before. If we remove the outputs mask, we don't...

Frame rate for this implementation

2

@michalfaber What about the frame rate of this implementation?

run_webcam

1

@salihkaragoz Is it possible to provide a script to do the inference using a webcam ?