deepinterpolation
deepinterpolation copied to clipboard
AllenInstitute
This PR makes it possible for users to specify filepaths that are not specific to the Allen Institute LIMS when instantiating a MovieJSONGenerator. This will be useful for outside users,...
The calculation of `img_per_movie` [here](https://github.com/AllenInstitute/deepinterpolation/blob/master/deepinterpolation/generator_collection.py#L1042) is incorrect. There is no requirement that the same number of frames per movie are sampled, and this calculation assumes this. Update this to allow...
The name of `lims_ids` [here](https://github.com/AllenInstitute/deepinterpolation/blob/master/deepinterpolation/generator_collection.py#L1040) is ambiguous. It should be renamed to `movie_ids` or `experiment_ids`
Currently deepinterpolation runs model.fit with the default verbosity of "auto". This does logging of training status in interactive mode, which uses tqdm. It clutters the logs and is not useful...
Hello there! We try to run the example tiny ophys training on our computers. By now, it is running perfectly fine, when we just use the CPU, but we face...
Doing long inference with Tensorflow 2.7, Python 3.9 can cause gpu out of memory errors like so: > 2021-12-01 13:52:12.086386: I tensorflow/core/common_runtime/bfc_allocator.cc:1074] 1 Chunks of size 28311552 totalling 27.00MiB >...
Hi there! Thanks for the tool and the support. Minor bug in the examples while transitioning to cli I suppose in the current versions of the scripted cli_example_* (all of...
Moving the latests code to a separate linux box, I experienced issues installing tensorflow 2.4.4 with python 3.7.x due to the h5py locked version. Further testing is required to remove...
the API layer reads and validates arguments. It currently must then write those to disk so the inference class can get a path as an argument. It would be preferable...
This is confusing but will break external dependencies. Will need to add proper notice for the next release.