search

med-air / 3DSAM-adapter

Holistic Adaptation of SAM from 2D to 3D for Promptable Medical Image Segmentation
170 stars 12 forks source link

The evaluation results using the provided pretrained checkpoint are not consistent with that in the paper. #14

Open ryylcc opened 1 year ago

ryylcc commented 1 year ago 

[14:46:55.321]  Case ./data/Task10_Colon/imagesTr/colon_008.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:46:55,321 -  Case ./data/Task10_Colon/imagesTr/colon_008.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:01.455]  Case ./data/Task10_Colon/imagesTr/colon_009.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:01,455 -  Case ./data/Task10_Colon/imagesTr/colon_009.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:07.196]  Case ./data/Task10_Colon/imagesTr/colon_026.nii.gz - Dice 0.015407 | NSD 0.065295
2023-08-15 14:47:07,196 -  Case ./data/Task10_Colon/imagesTr/colon_026.nii.gz - Dice 0.015407 | NSD 0.065295
[14:47:12.848]  Case ./data/Task10_Colon/imagesTr/colon_030.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:12,848 -  Case ./data/Task10_Colon/imagesTr/colon_030.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:20.227]  Case ./data/Task10_Colon/imagesTr/colon_040.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:20,227 -  Case ./data/Task10_Colon/imagesTr/colon_040.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:30.019]  Case ./data/Task10_Colon/imagesTr/colon_046.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:30,019 -  Case ./data/Task10_Colon/imagesTr/colon_046.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:39.798]  Case ./data/Task10_Colon/imagesTr/colon_091.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:39,798 -  Case ./data/Task10_Colon/imagesTr/colon_091.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:46.366]  Case ./data/Task10_Colon/imagesTr/colon_095.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:46,366 -  Case ./data/Task10_Colon/imagesTr/colon_095.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:53.232]  Case ./data/Task10_Colon/imagesTr/colon_098.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:53,232 -  Case ./data/Task10_Colon/imagesTr/colon_098.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:59.353]  Case ./data/Task10_Colon/imagesTr/colon_104.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:59,353 -  Case ./data/Task10_Colon/imagesTr/colon_104.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:07.321]  Case ./data/Task10_Colon/imagesTr/colon_129.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:07,321 -  Case ./data/Task10_Colon/imagesTr/colon_129.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:15.462]  Case ./data/Task10_Colon/imagesTr/colon_133.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:15,462 -  Case ./data/Task10_Colon/imagesTr/colon_133.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:25.344]  Case ./data/Task10_Colon/imagesTr/colon_136.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:25,344 -  Case ./data/Task10_Colon/imagesTr/colon_136.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:32.039]  Case ./data/Task10_Colon/imagesTr/colon_141.nii.gz - Dice 0.000000 | NSD 0.017542
2023-08-15 14:48:32,039 -  Case ./data/Task10_Colon/imagesTr/colon_141.nii.gz - Dice 0.000000 | NSD 0.017542
[14:48:38.911]  Case ./data/Task10_Colon/imagesTr/colon_142.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:38,911 -  Case ./data/Task10_Colon/imagesTr/colon_142.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:44.344]  Case ./data/Task10_Colon/imagesTr/colon_145.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:44,344 -  Case ./data/Task10_Colon/imagesTr/colon_145.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:53.632]  Case ./data/Task10_Colon/imagesTr/colon_163.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:53,632 -  Case ./data/Task10_Colon/imagesTr/colon_163.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:00.167]  Case ./data/Task10_Colon/imagesTr/colon_164.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:00,167 -  Case ./data/Task10_Colon/imagesTr/colon_164.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:06.922]  Case ./data/Task10_Colon/imagesTr/colon_181.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:06,922 -  Case ./data/Task10_Colon/imagesTr/colon_181.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:14.982]  Case ./data/Task10_Colon/imagesTr/colon_194.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:14,982 -  Case ./data/Task10_Colon/imagesTr/colon_194.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:21.982]  Case ./data/Task10_Colon/imagesTr/colon_202.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:21,982 -  Case ./data/Task10_Colon/imagesTr/colon_202.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:26.959]  Case ./data/Task10_Colon/imagesTr/colon_205.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:26,959 -  Case ./data/Task10_Colon/imagesTr/colon_205.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:33.510]  Case ./data/Task10_Colon/imagesTr/colon_207.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:33,510 -  Case ./data/Task10_Colon/imagesTr/colon_207.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:41.245]  Case ./data/Task10_Colon/imagesTr/colon_208.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:41,245 -  Case ./data/Task10_Colon/imagesTr/colon_208.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:50.006]  Case ./data/Task10_Colon/imagesTr/colon_215.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:50,006 -  Case ./data/Task10_Colon/imagesTr/colon_215.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:57.969]  Case ./data/Task10_Colon/imagesTr/colon_218.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:57,969 -  Case ./data/Task10_Colon/imagesTr/colon_218.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:58,035 - - Test metrics Dice: 0.00059259165
2023-08-15 14:49:58,035 - - Test metrics NSD: 0.0031860285794686172
peterant330 commented 1 year ago 

[14:46:55.321]  Case ./data/Task10_Colon/imagesTr/colon_008.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:46:55,321 -  Case ./data/Task10_Colon/imagesTr/colon_008.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:01.455]  Case ./data/Task10_Colon/imagesTr/colon_009.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:01,455 -  Case ./data/Task10_Colon/imagesTr/colon_009.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:07.196]  Case ./data/Task10_Colon/imagesTr/colon_026.nii.gz - Dice 0.015407 | NSD 0.065295
2023-08-15 14:47:07,196 -  Case ./data/Task10_Colon/imagesTr/colon_026.nii.gz - Dice 0.015407 | NSD 0.065295
[14:47:12.848]  Case ./data/Task10_Colon/imagesTr/colon_030.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:12,848 -  Case ./data/Task10_Colon/imagesTr/colon_030.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:20.227]  Case ./data/Task10_Colon/imagesTr/colon_040.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:20,227 -  Case ./data/Task10_Colon/imagesTr/colon_040.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:30.019]  Case ./data/Task10_Colon/imagesTr/colon_046.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:30,019 -  Case ./data/Task10_Colon/imagesTr/colon_046.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:39.798]  Case ./data/Task10_Colon/imagesTr/colon_091.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:39,798 -  Case ./data/Task10_Colon/imagesTr/colon_091.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:46.366]  Case ./data/Task10_Colon/imagesTr/colon_095.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:46,366 -  Case ./data/Task10_Colon/imagesTr/colon_095.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:53.232]  Case ./data/Task10_Colon/imagesTr/colon_098.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:53,232 -  Case ./data/Task10_Colon/imagesTr/colon_098.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:59.353]  Case ./data/Task10_Colon/imagesTr/colon_104.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:59,353 -  Case ./data/Task10_Colon/imagesTr/colon_104.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:07.321]  Case ./data/Task10_Colon/imagesTr/colon_129.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:07,321 -  Case ./data/Task10_Colon/imagesTr/colon_129.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:15.462]  Case ./data/Task10_Colon/imagesTr/colon_133.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:15,462 -  Case ./data/Task10_Colon/imagesTr/colon_133.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:25.344]  Case ./data/Task10_Colon/imagesTr/colon_136.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:25,344 -  Case ./data/Task10_Colon/imagesTr/colon_136.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:32.039]  Case ./data/Task10_Colon/imagesTr/colon_141.nii.gz - Dice 0.000000 | NSD 0.017542
2023-08-15 14:48:32,039 -  Case ./data/Task10_Colon/imagesTr/colon_141.nii.gz - Dice 0.000000 | NSD 0.017542
[14:48:38.911]  Case ./data/Task10_Colon/imagesTr/colon_142.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:38,911 -  Case ./data/Task10_Colon/imagesTr/colon_142.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:44.344]  Case ./data/Task10_Colon/imagesTr/colon_145.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:44,344 -  Case ./data/Task10_Colon/imagesTr/colon_145.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:53.632]  Case ./data/Task10_Colon/imagesTr/colon_163.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:53,632 -  Case ./data/Task10_Colon/imagesTr/colon_163.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:00.167]  Case ./data/Task10_Colon/imagesTr/colon_164.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:00,167 -  Case ./data/Task10_Colon/imagesTr/colon_164.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:06.922]  Case ./data/Task10_Colon/imagesTr/colon_181.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:06,922 -  Case ./data/Task10_Colon/imagesTr/colon_181.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:14.982]  Case ./data/Task10_Colon/imagesTr/colon_194.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:14,982 -  Case ./data/Task10_Colon/imagesTr/colon_194.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:21.982]  Case ./data/Task10_Colon/imagesTr/colon_202.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:21,982 -  Case ./data/Task10_Colon/imagesTr/colon_202.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:26.959]  Case ./data/Task10_Colon/imagesTr/colon_205.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:26,959 -  Case ./data/Task10_Colon/imagesTr/colon_205.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:33.510]  Case ./data/Task10_Colon/imagesTr/colon_207.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:33,510 -  Case ./data/Task10_Colon/imagesTr/colon_207.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:41.245]  Case ./data/Task10_Colon/imagesTr/colon_208.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:41,245 -  Case ./data/Task10_Colon/imagesTr/colon_208.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:50.006]  Case ./data/Task10_Colon/imagesTr/colon_215.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:50,006 -  Case ./data/Task10_Colon/imagesTr/colon_215.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:57.969]  Case ./data/Task10_Colon/imagesTr/colon_218.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:57,969 -  Case ./data/Task10_Colon/imagesTr/colon_218.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:58,035 - - Test metrics Dice: 0.00059259165
2023-08-15 14:49:58,035 - - Test metrics NSD: 0.0031860285794686172

Sorry we just did some modification to the training and testing code, but the checkpoint is not updated yet. I will update the checkpoint within this week.

CunminZhao commented 1 year ago 

[14:46:55.321]  Case ./data/Task10_Colon/imagesTr/colon_008.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:46:55,321 -  Case ./data/Task10_Colon/imagesTr/colon_008.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:01.455]  Case ./data/Task10_Colon/imagesTr/colon_009.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:01,455 -  Case ./data/Task10_Colon/imagesTr/colon_009.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:07.196]  Case ./data/Task10_Colon/imagesTr/colon_026.nii.gz - Dice 0.015407 | NSD 0.065295
2023-08-15 14:47:07,196 -  Case ./data/Task10_Colon/imagesTr/colon_026.nii.gz - Dice 0.015407 | NSD 0.065295
[14:47:12.848]  Case ./data/Task10_Colon/imagesTr/colon_030.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:12,848 -  Case ./data/Task10_Colon/imagesTr/colon_030.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:20.227]  Case ./data/Task10_Colon/imagesTr/colon_040.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:20,227 -  Case ./data/Task10_Colon/imagesTr/colon_040.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:30.019]  Case ./data/Task10_Colon/imagesTr/colon_046.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:30,019 -  Case ./data/Task10_Colon/imagesTr/colon_046.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:39.798]  Case ./data/Task10_Colon/imagesTr/colon_091.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:39,798 -  Case ./data/Task10_Colon/imagesTr/colon_091.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:46.366]  Case ./data/Task10_Colon/imagesTr/colon_095.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:46,366 -  Case ./data/Task10_Colon/imagesTr/colon_095.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:53.232]  Case ./data/Task10_Colon/imagesTr/colon_098.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:53,232 -  Case ./data/Task10_Colon/imagesTr/colon_098.nii.gz - Dice 0.000000 | NSD 0.000000
[14:47:59.353]  Case ./data/Task10_Colon/imagesTr/colon_104.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:47:59,353 -  Case ./data/Task10_Colon/imagesTr/colon_104.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:07.321]  Case ./data/Task10_Colon/imagesTr/colon_129.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:07,321 -  Case ./data/Task10_Colon/imagesTr/colon_129.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:15.462]  Case ./data/Task10_Colon/imagesTr/colon_133.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:15,462 -  Case ./data/Task10_Colon/imagesTr/colon_133.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:25.344]  Case ./data/Task10_Colon/imagesTr/colon_136.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:25,344 -  Case ./data/Task10_Colon/imagesTr/colon_136.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:32.039]  Case ./data/Task10_Colon/imagesTr/colon_141.nii.gz - Dice 0.000000 | NSD 0.017542
2023-08-15 14:48:32,039 -  Case ./data/Task10_Colon/imagesTr/colon_141.nii.gz - Dice 0.000000 | NSD 0.017542
[14:48:38.911]  Case ./data/Task10_Colon/imagesTr/colon_142.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:38,911 -  Case ./data/Task10_Colon/imagesTr/colon_142.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:44.344]  Case ./data/Task10_Colon/imagesTr/colon_145.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:44,344 -  Case ./data/Task10_Colon/imagesTr/colon_145.nii.gz - Dice 0.000000 | NSD 0.000000
[14:48:53.632]  Case ./data/Task10_Colon/imagesTr/colon_163.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:48:53,632 -  Case ./data/Task10_Colon/imagesTr/colon_163.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:00.167]  Case ./data/Task10_Colon/imagesTr/colon_164.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:00,167 -  Case ./data/Task10_Colon/imagesTr/colon_164.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:06.922]  Case ./data/Task10_Colon/imagesTr/colon_181.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:06,922 -  Case ./data/Task10_Colon/imagesTr/colon_181.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:14.982]  Case ./data/Task10_Colon/imagesTr/colon_194.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:14,982 -  Case ./data/Task10_Colon/imagesTr/colon_194.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:21.982]  Case ./data/Task10_Colon/imagesTr/colon_202.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:21,982 -  Case ./data/Task10_Colon/imagesTr/colon_202.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:26.959]  Case ./data/Task10_Colon/imagesTr/colon_205.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:26,959 -  Case ./data/Task10_Colon/imagesTr/colon_205.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:33.510]  Case ./data/Task10_Colon/imagesTr/colon_207.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:33,510 -  Case ./data/Task10_Colon/imagesTr/colon_207.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:41.245]  Case ./data/Task10_Colon/imagesTr/colon_208.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:41,245 -  Case ./data/Task10_Colon/imagesTr/colon_208.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:50.006]  Case ./data/Task10_Colon/imagesTr/colon_215.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:50,006 -  Case ./data/Task10_Colon/imagesTr/colon_215.nii.gz - Dice 0.000000 | NSD 0.000000
[14:49:57.969]  Case ./data/Task10_Colon/imagesTr/colon_218.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:57,969 -  Case ./data/Task10_Colon/imagesTr/colon_218.nii.gz - Dice 0.000000 | NSD 0.000000
2023-08-15 14:49:58,035 - - Test metrics Dice: 0.00059259165
2023-08-15 14:49:58,035 - - Test metrics NSD: 0.0031860285794686172

Sorry we just did some modification to the training and testing code, but the checkpoint is not updated yet. I will update the checkpoint within this week.

So, is the training and test code provided on GitHub the version before or after your update?

peterant330 commented 1 year ago

ing and test

The code is after the updating. But the checkpoint is before the updating (where the architecture is actually one of the ablation). Sorry very much for the inconvenience.

CunminZhao commented 1 year ago

ing and test

The code is after the updating. But the checkpoint is before the updating (where the architecture is actually one of the ablation). Sorry very much for the inconvenience.

thanks for your reply

bingcheg commented 1 year ago

Dear author

Have you updated the latest checkpoint so that I can reproduce your results?

peterant330 commented 1 year ago

Dear author

Have you updated the latest checkpoint so that I can reproduce your results?

Hi,

I just uploaded checkpoint for MSD-pancreas, lits and MSD-Colon. The training of Kits takes longer and will be uploaded soon.

Best

amparobetancourt commented 1 year ago

Dear author, thanks for your contribution! I downloaded the latest lits checkpoint to reproduce your results on the test data, however I can't load the weights I keep getting an error RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory

Which doesnt happen with the MSD-pancreas weights and I have the recommended versions and packages , do you have any advice on this ? Thanks in advance!

peterant330 commented 1 year ago

Dear author, thanks for your contribution! I downloaded the latest lits checkpoint to reproduce your results on the test data, however I can't load the weights I keep getting an error RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory

Which doesnt happen with the MSD-pancreas weights and I have the recommended versions and packages , do you have any advice on this ? Thanks in advance!

Hi, I think this is the error due to the broken file. I am unsure if your downloading process breaks the file or if I uploaded the broken file. I just uploaded it again and maybe you can have a try.

amparobetancourt commented 1 year ago

Thank you! It works now!

ghost commented 1 year ago

Dear author, I also meet the similar problem when using the pretrained checkpoint for the MSD-colon, it says "RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory" Can you please make sure the uploaded checkpoint for the MSD-colon is correct?

Thank you so much

peterant330 commented 1 year ago

Dear author, I also meet the similar problem when using the pretrained checkpoint for the MSD-colon, it says "RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory" Can you please make sure the uploaded checkpoint for the MSD-colon is correct?

Thank you so much

Hi, we have uploaded a new one. Can you try again?

yun263214678 commented 1 year ago

2023-10-14 19:53:25,603 - - Test metrics Dice: 0.58578235 2023-10-14 19:53:25,603 - - Test metrics NSD: 0.7426014035058957

Hi, thanks a lot for your work, I tested colon with --num_prompts 1 , but in the test.py, can we use rand_crop_size in test data? which means that a volume has more than one point, and should the testing phase be better at the original size? (no rand_crop) . Or comparative experiments with other models, for example, does nnU-Net also use random crop size when testing?

peterant330 commented 1 year ago

2023-10-14 19:53:25,603 - - Test metrics Dice: 0.58578235 2023-10-14 19:53:25,603 - - Test metrics NSD: 0.7426014035058957

Hi, thanks a lot for your work, I tested colon with --num_prompts 1 , but in the test.py, can we use rand_crop_size in test data? which means that a volume has more than one point, and should the testing phase be better at the original size? (no rand_crop) . Or comparative experiments with other models, for example, does nnU-Net also use random crop size when testing?

Hi, for the test, we also use rand_crop_size, which is set to be the same as during the training. nnU-Net using a sliding window inference mechanism where the prediction is also predicted on a patch-wise basis but all the patches in the image are involved for inference. Our method requires a point as a prompt, as a result, we cannot predict for those patches far away from the prompt. Only patches with prompts inside can be used for prediction. Actually, this is also one advantage of interactive segmentation, we have knowledge about where the foreground is by prompts.

yun263214678 commented 1 year ago

2023-10-14 19:53:25,603 - - Test metrics Dice: 0.58578235 2023-10-14 19:53:25,603 - - Test metrics NSD: 0.7426014035058957 Hi, thanks a lot for your work, I tested colon with --num_prompts 1 , but in the test.py, can we use rand_crop_size in test data? which means that a volume has more than one point, and should the testing phase be better at the original size? (no rand_crop) . Or comparative experiments with other models, for example, does nnU-Net also use random crop size when testing?

Hi, for the test, we also use rand_crop_size, which is set to be the same as during the training. nnU-Net using a sliding window inference mechanism where the prediction is also predicted on a patch-wise basis but all the patches in the image are involved for inference. Our method requires a point as a prompt, as a result, we cannot predict for those patches far away from the prompt. Only patches with prompts inside can be used for prediction. Actually, this is also one advantage of interactive segmentation, we have knowledge about where the foreground is by prompts.

Thanks for your reply, and I also want to know that if we need same crop size for different models in both train and test step, does the comparative experiment need to be set like this, because it seems that crop size has a great impact on the result during testing.