facebookresearch / FixRes
This repository reproduces the results of the paper: "Fixing the train-test resolution discrepancy" https://arxiv.org/abs/1906.06423
Summary
 |
|||
| Main Code: 3,576 LOC (37 files) = PY (100%) Secondary code: Test: 0 LOC (0); Generated: 0 LOC (0); Build & Deploy: 0 LOC (0); Other: 606 LOC (5); |
|||
| Duplication: 85% | |||
| File Size: 0% long (>1000 LOC), 63% short (<= 200 LOC) | |||
| Unit Size: 3% long (>100 LOC), 45% short (<= 10 LOC) | |||
| Conditional Complexity: 2% complex (McCabe index > 50), 68% simple (McCabe index <= 5) | |||
|
|
Logical Component Decomposition: primary (5 components) | ||
|
|
2 years, 5 months old
|
||
|
|
0% of code updated more than 50 times Also see temporal dependencies for files frequently changed in same commits. |
||
| Goals: Keep the system simple and easy to change (4) | |||
Latest commit date: 2020-12-26
|
0
commits
(30 days)
0
contributors
(30 days) |
|