vllm/docs/models/extensions/fastsafetensor.md

6 lines
418 B
Markdown

Loading Model weights with fastsafetensors
===================================================================
Using fastsafetensors library enables loading model weights to GPU memory by leveraging GPU direct storage. See [their GitHub repository](https://github.com/foundation-model-stack/fastsafetensors) for more details.
For enabling this feature, set the environment variable ``USE_FASTSAFETENSOR`` to ``true``