ML engineer, Scikit-learn core-dev, working on PyTorch at Facebook AI
Highlights
- 10 discussions answered
Block or Report
Block or report NicolasHug
Report abuse
Contact GitHub support about this user’s behavior. Learn more about reporting abuse.
Report abusePinned Loading
1,161 contributions in the last year
Less
More
Activity overview
Loading
Contribution activity
November 2021
Created 13 commits in 3 repositories
Created a pull request in pytorch/vision that received 28 comments
Add Kitti and Sintel datasets for optical flow
This PR Adds the Kitti and Sintel datasets for optical flow. Here are a few design decisions, following our initial discussion (in this PR): Unli…
+373
−1
•
28
comments
Opened 9 other pull requests in 2 repositories
pytorch/vision
1
open
7
merged
- NOMRG Comments on prototype weights and model builders
- Add HD1K dataset for optical flow
- Change expected directory from "Kitti" to "KittiFlow" for KittiFlow dataset
- Add pass_name='both' for Sintel dataset
- Fix .flo file decoding
- Support big endian for .flo file decoding
- Add FlyingChairs dataset for optical flow
- Add FlyingThings3D dataset for optical flow
pytorch/hub
1
merged
Reviewed 28 pull requests in 5 repositories
pytorch/vision
22 pull requests
- Add Sintel Dataset to the dataset prototype API
- Fix clang_format CI failure
- Add HD1K dataset for optical flow
- Fix failing CI job for android
-
Set seed on
test_nms_refto reduce flakiness - Update the error message on DeformConv2d
- Reverted folder.py back to using complete path to file for make_dataset
- [FBcode->GH] Removed type annotations from rcnn
- add prototype utilities to read arbitrary numeric binary files
- Remove unused import os in datasets
- disable weight download and state dict loading for model tests
- Fix .flo file decoding
- Add FlyingThings3D dataset for optical flow
- Add FlyingChairs dataset for optical flow
- Add Kitti and Sintel datasets for optical flow
-
use
store_test_resultsfor prototype tests - Fix TORCH_HOME environment variable in docs
- remove hard coded input from category file generation
- Update requirements.txt
- add missing 0.11/_modules/torchvision via git add -f
- use toplevel directories
- Expose misc ops at package level
scikit-learn/scikit-learn
2 pull requests
pytorch/pytorch
2 pull requests
microsoft/recommenders
1 pull request
pytorch/hub
1 pull request
Created an issue in princeton-vl/RAFT that received 2 comments
Valid mask prevents x and y to be 0 in resize_sparse_flow_map()
Hi, and thanks a lot for your work and for publishing the code!
In resize_sparse_flow_map(), the new valid mask is define like as:
RAFT/core/uti…
2
comments