Import Torch/PyTorch #8
Comments
thequicksort
added a commit
that referenced
this issue
Apr 15, 2020
Add pytorch dependency. #8 Co-authored-by: Jessica Dunstan <jdunstan@cs.washington.edu>
kdoroschak
pushed a commit
that referenced
this issue
Apr 27, 2020
Add pytorch dependency. #8 Co-authored-by: Jessica Dunstan <jdunstan@cs.washington.edu>
kdoroschak
added a commit
that referenced
this issue
Apr 28, 2020
* begin quantifier refactor; absolutely does not run attempting to get a basic test set up for calc_time_until_capture. test comes from 20180618 data, run on the original version. moved voltage utility function to segmenter. * rewrite calc_time_until_capture to be conceptually easier to follow * fix relative import . * remove orig version of calc_time_until_capture * add init so pytest can find tests; minor flake compliance * add test for get_overlapping_regions; fixed edge case bug * docstrings * update parameterized fn * remove calc_time_until_capture_blockages * minor flake change, star import, & add test file * open_pore --> _channel; docstrings raw utils; rm multi path choice raw utils * python 3 import * fix import * method to create the capture fast5 + tests * rm plotting in utils * rm example test * add filters (+tests) & find capture fn (untested) * make fn call more generic * window-only segmentation core fn * functional segmenter + tests + parallel * finish doc for segment, move fn to raw_signal_utils * rm placeholder test * reorganize module names for downstream changes * Build docker image * Upload the docker image as a build artifact * rewrite calc_time_until_capture to be conceptually easier to follow * fix relative import . * remove orig version of calc_time_until_capture * add init so pytest can find tests; minor flake compliance * add test for get_overlapping_regions; fixed edge case bug * Build the application's docker image 🐳 (#14) Run "nix-build" or "nix-build -A docker" to build a docker image. Run "nix-build -A app" to build Poretitioner locally without Docker. #14 * Adding build steps to the Readme * Added more build steps * Update Readme.md * Added NanoporTER paper * Update Readme.md * docstrings * update parameterized fn * remove calc_time_until_capture_blockages * minor flake change, star import, & add test file * open_pore --> _channel; docstrings raw utils; rm multi path choice raw utils * python 3 import * fix import * Import pytorch and torchvision (#25) Add pytorch dependency. #8 Co-authored-by: Jessica Dunstan <jdunstan@cs.washington.edu> * method to create the capture fast5 + tests * Update Readme.md * Added continuous integration badge * Add uninstall steps * rm plotting in utils * rm example test * add filters (+tests) & find capture fn (untested) * make fn call more generic * window-only segmentation core fn * functional segmenter + tests + parallel * finish doc for segment, move fn to raw_signal_utils * rm placeholder test * reorganize module names for downstream changes Co-authored-by: J.D <thequicksort@gmail.com> Co-authored-by: Jessica Dunstan <jdunstan@cs.washington.edu>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Description
This task is to bring in PyTorch and PyTorch vision as run-time dependencies.
There's some difficulty, especially in determining hardware constraints and CUDA vs. non-CUDA.
After refactoring we may be able to isolate this more.
The text was updated successfully, but these errors were encountered: