facebookresearch / fairscale
PyTorch extensions for high performance and large scale training.
Summary
    |
|||
| Main Code: 18,003 LOC (192 files) = PY (77%) + PYI (20%) + CU (1%) + CUH (<1%) + CFG (<1%) + TOML (<1%) + YML (<1%) + CPP (<1%) + H (<1%) + IN (<1%) | |||
| Duplication: 5% | |||
| File Size: 17% long (>1000 LOC), 46% short (<= 200 LOC) | |||
| Unit Size: 0% long (>100 LOC), 73% short (<= 10 LOC) | |||
| Conditional Complexity: 1% complex (McCabe index > 50), 69% simple (McCabe index <= 5) | |||
|
|
Logical Component Decomposition: primary (13 components) | ||
|
|
1 year, 6 months old
|
||
|
|
9% 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) | |||
|
Features of interest:
TODOs
18 files |
|||
Latest commit date: 2022-01-20
|
13
commits
(30 days)
6
contributors
(30 days) |
|