vllm-project / speculators
A unified library for building, evaluating, and storing speculative decoding algorithms for LLM inference in vLLM
- Repository stars
- 942All-time #41,846
- Fork
- 285
- Last push
- updated 1 days ago
- Repository created
- since 2025
About
A unified library for building, evaluating, and storing speculative decoding algorithms for LLM inference in vLLMSummarised by AI from the README; the original prevails
Author description: A unified library for building, evaluating, and storing speculative decoding algorithms for LLM inference in vLLM
Basic listing based on official public information; functionality has not been independently evaluated.
- Pricing
- Pricing not verifiedCheck current official software, hosting, model and infrastructure pricing.