hact-net

hact-net copied to clipboard

Preparing transaction: done Verifying transaction: done Executing transaction: done Installing pip dependencies: / Ran pip subprocess with arguments: ['/usr/local/envs/hactnet/bin/python', '-m', 'pip', 'install', '-U', '-r', '/content/drive/MyDrive/hact-net-onbracs/hact-net/condaenv.4lthjbvu.requirements.txt', '--exists-action=b'] Pip subprocess output: Collecting...

asortubay

hi, how to get the row-normalized 7-class confusion matrix for HACT-Net in the paper? thx

StevenHaojc

Val loss nan keeps nan

1

Thank you for your excellent work. I tried to apply your work to our own data to achieve binary classification, but the val loss calculation kept nan all the time,...

ZYICE111

Pretrained weights do not exist

3

Hello, I am trying to run inference with the Hact-net model, but it appears that the download link for the pretrained weights does not work.

ritvikg4

failed during cell graph generation; failed during tissue graph generation; failed during assignment matrix generation

1

Hi! nuclei_map, nuclei_centroids = self.nuclei_detector._process(image) nuclei_map, nuclei_centroids = self.nuclei_detector._process(image, tissue_mask) I'm not sure if "tissue_mask" is a required parameter for graph generation 

ZhangRan24

I have a question on the Graph Generated. In the paper it said it is a undirected Hierarchical Graph. But DGL supports directed graph it seems like. How is the...

PeterrHH

Hello, may I ask where can I download pre-trained models?

huihhui

cell graph train loss turn into nan after several epoches

2

Thanks for your excellent work，but when i train the cell graph model，the trian loss become nan after several epoches，could you give me some hints？

WxpHB

tissue detection

2

How to get the tissue detection visualization from SLIC algorithm like the tissue detection shown in the entity detection part?

Rukhmini

1. in dataloader.py: function _load_cg and function _load_tg: As shown in the figure，the 'cg_fnames'(line 69) and the 'tg_fnames'(line 84) has included the full path, so the code in line 73...

mantras-cloud