Statements (19)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
gptkb:software
|
| gptkbp:compatibleWith |
gptkb:PyTorch
|
| gptkbp:developedBy |
gptkb:Facebook_AI_Research
|
| gptkbp:firstReleased |
2020
|
| gptkbp:focusesOn |
distributed training
model parallelism optimizations for PyTorch |
| https://www.w3.org/2000/01/rdf-schema#label |
fairscale
|
| gptkbp:license |
gptkb:MIT_License
|
| gptkbp:programmingLanguage |
gptkb:Python
|
| gptkbp:provides |
Fully Sharded Data Parallel (FSDP)
Memory optimizations Optimizer State Sharding Pipeline Parallelism Sharded Data Parallel |
| gptkbp:repository |
https://github.com/facebookresearch/fairscale
|
| gptkbp:usedFor |
large-scale deep learning
|
| gptkbp:bfsParent |
gptkb:Databricks_Runtime_11.3_ML
|
| gptkbp:bfsLayer |
8
|