DotWang

Hope to support mixed-precision training

2

The function `rcca.ca_forward_cuda` seems doesn't support mixed-precision training and raise error: RuntimeError: expected scalar type Float but found Half (data at /data/.conda/envs/pytorch/lib/python3.5/site-packages/torch/include/ATen/core/TensorMethods.h:1821) when the **opt_level** of NVIDIA APEX is set...

多尺度处理完训练验证集和测试集后分别有68325和71888张 不过我在查阅相关issue时，发现68325并不等于单独处理训练集和验证集的加和（后者80000+） 请问这是为什么？

Question about seed dynamic updating

4

It's a very interesting work. We are conducting a similar segmentation task without initial cues file, and I hope to generate seeds in the first epoch and update them in...

新手求救！！从没见过的错误：filename，directory name，or volume label syntax incorrect.

3

``` C:\Users\ww\Desktop\QA\TopSup-master>python GetQuestionTessAndroid.py 采用方式 1 获取截图 Traceback (most recent call last): File "GetQuestionTessAndroid.py", line 20, in question, choices = ocr.ocr_img(img) File "C:\Users\ww\Desktop\QA\TopSup-master\common\ocr.py", line 76, in ocr_img question = pytesseract.image_to_string(question_im, lang='chi_sim',...

Thank you.