SSL4MIS

SSL4MIS copied to clipboard

train own 2D pictures

2

hello，thanks for your revision very much。 I still want to ask a question。 We can see that the h5 file includes both image and label data， and when we train，for...

cebain

train on my dataset

1

Great job! I want to know how to train on my own dataset? For example, if I want to perform segmentation on the abdominal multi organ dataset BTCV, what modifications...

sharonlee12

Thank you for your project

2

"Thank you for your project. When can we expect it to continue updating? I'm looking forward to your continued updates on the overview."

CharlieFengCN

KeyError: '136'

4

Hello, there was a key error when I was running the train_mean_teacher_2D.py. I tried many methods but failed to solve it. May I ask what the problem is? If possible,...

demondemoliu

Hello. Apologies for the confusion caused by the keyboard issue earlier. Thank you for your excellent work. In the implementation of the TwoStreamBatchSampler(Sampler) class, you mentioned, "An 'epoch' is one...

Yuyu-Liu11037

Hi, thanks for your excellent work! In my experimental settings, the EMA update in the BN layer is also necessary. Why does not it need here?

ZhouCX117

for

3

hello，thanks for your revision very much。 I still want to ask a question。 We can see that the h5 file includes both image and label data， and when we train，for...

cebain

conf_thresh in fixmatch

1

发现train_fixmatch_standard_augs.py文件存在一个有趣的问题：  如果通过命令行指令修改--conf_thresh 的值，会报错没有这个参数 我在尝试寻找问题时将该参数设置为了必需，再次通过命令行指令修改--conf_thresh 的值 ` parser.add_argument("--conf_thresh", type=float, required=True, help="confidence threshold for using pseudo-labels") ` 我发现：  当未指定conf_thresh时，会报错conf_thresh参数是必需的，指定conf_thresh又会导致报错：不能识别conf_thresh 这可太有意思了 我目前还在寻找这种情况出现的原因。

Duckyee728

could u let me know where does this code that i have provided below come from?

2

yvonneng996

supgy