Closed stevevanhooser closed 5 months ago
I wonder if dropbox changed their API. this? https://www.dropbox.com/scl/fi/nuxgnjq2524uv2txp1mdb/2019_09_11_23_32_unet_single_1024_mean_absolute_error_Ai93-0450.h5?rlkey=sgyh0cctq0su0m1kx2imdgey1&dl=0
On Sat, May 4, 2024 at 11:38 AM Steve Van Hooser @.***> wrote:
The following file now shows You don't have access
— Reply to this email directly, view it on GitHub https://github.com/MATLAB-Community-Toolboxes-at-INCF/DeepInterpolation-MATLAB/issues/50, or unsubscribe https://github.com/notifications/unsubscribe-auth/AATAHT6A6IVDVL5EU2Q3E7DZAUTIVAVCNFSM6AAAAABHHA6NYKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGI3TSMJUHE4DOMY . You are receiving this because you are subscribed to this thread.Message ID: <MATLAB-Community-Toolboxes-at-INCF/DeepInterpolation-MATLAB/issues/50 @github.com>
-- Jérôme
Thanks Jerome..now I get a different error that says That didn't work for some reason
Thanks again Jerome, but I still get That didn't work for some reason
. (Not a very helpful error message.)
Google Drive seems a dead end, too (https://www.mathworks.com/matlabcentral/answers/553867-how-to-add-files-directly-from-google-drive-to-matlab-online)
OK, I am going to use our OpenDatasets account, as this is a "for the public good" purpose:
https://wds-matlab-community-toolboxes.s3.amazonaws.com/DeepInterpolation-MATLAB/Trained_models/ophys/2019_09_11_23_32_unet_single_1024_mean_absolute_error_Ai93-0450.h5
Branch newdemo
updated with the AWS link
Ok. I will also work to migrate those files to a model zoo or similar.
Is there already a list of models along with some description of how they were trained? I'm imagining users could write:
net = deepinterp.Net.ImageTimeSeries('pretrained','model','modelname1_512_512_30_30')
and load any of a set of named models, for example. This would make it really easy for users to apply your pretrained models.
Thanks!
Those are described in the readme there : https://github.com/AllenInstitute/deepinterpolation
But I have not set up a library or such.
Jerome
On Sun, May 5, 2024 at 1:57 PM Steve Van Hooser @.***> wrote:
Is there already a list of models along with some description of how they were trained? I'm imagining users could write:
net = deepinterp.Net.ImageTimeSeries('pretrained','model','modelname1_512_512_30_30')
and load any of a set of named models, for example. This would make it really easy for users to apply your pretrained models.
Thanks!
— Reply to this email directly, view it on GitHub https://github.com/MATLAB-Community-Toolboxes-at-INCF/DeepInterpolation-MATLAB/issues/50#issuecomment-2094948197, or unsubscribe https://github.com/notifications/unsubscribe-auth/AATAHTYFQNDJSOQ7KD234OTZA2MMRAVCNFSM6AAAAABHHA6NYKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAOJUHE2DQMJZG4 . You are receiving this because you commented.Message ID: <MATLAB-Community-Toolboxes-at-INCF/DeepInterpolation-MATLAB/issues/50/2094948197 @github.com>
-- Jérôme
The following file now shows
You don't have access
https://www.dropbox.com/sh/vwxf1uq2j60uj9o/AAC0sZWahCJFBRARoYsw8Nnra/2019_09_11_23_32_unet_single_1024_mean_absolute_error_Ai93-0450.h5?dl=0