TextZoom

TextZoom copied to clipboard

Cannot load trained checkpoints

4

Pytorch 1.3 used for training (Ok) and for inference. I had a problem with Dataparallel, which I solved by copying multi-gpu model load code in into 1 gpu case: ```...

hcl14

hello，can you tell me the weight of the two loss, 20 and 1e-4 in proviede code?thank you !

xcc19970423

dataset

2

Hello, I'm very interested in your research. Could you please provide the textzoom dataset in picture format and the label information in TXT format for the experiment in the paper?

xcc19970423

Hi, thanks a lot for your great work! i would like to use your labels ground truth (sr_raw.json) but they do not macth the original images of the SRRAW dataset....

ishaygoldinpersonal

Where are the images saved?

2

After I run the demo, where are the images of the super resolution texts saved? Like after the image has been processed before the text reading process.

3BBUAE

what's the difference between train1 and train2?

xcc19970423

make my dataset like textzoom

1

Hello ～，thank you for your excellent work first, can you tell me how to make the mdb dataset, I have images and tags, how to make my own dataset?

happy20200

Hello, what is the meaning of transpose in line 94(tsrn.py)? Does anyone know? Thanks in advance. 

yuinnLIU

Could you please give a detailed training environment configuration?

1

When I tried to train the network with cuda 10.1, cudnn 8.0.2 and torch 1.4.1, it always reported a "core dump" error. And when I turned off cudnn the error...

Imalne

Dear JasonBoy1, would it be possible for you to add a license file, so as to make clear what usage you want to permit? E.g. [Apache 2.0-License](https://www.apache.org/licenses/LICENSE-2.0) Thank you and...

Theophrastos