vLLM
@vllm_project
A high-throughput and memory-efficient inference and serving engine for LLMs. Join to discuss together with the community!
๊ฐ์
March 2024
37 ํ๋ก์ ์ค 54.1K ํฌ
๐ Day-0 vLLM support for Sage-A1.5 (35B MoE, 3B active) from Shanghai AI Lab โ agentic intelligence you can serve locally today:
vllm serve InternScience/Sage-A1.5
Congrats @sageai_official ! ๐ฅ
๋ ๋ณด๊ธฐ
๐Sage-A1.5 by Shanghai AI Lab is here โ pushing 35B agentic intelligence to the next level ๐ฅ
๐ค 35B MoE (3B active)
๐งฉ Agentic Orchestration & Verification โ spawns parallel subagents & self-corrects on the fly
๐ RSI Data Flywheel โ evolved via real scientific discovery workflows & tens of thousands of synthesized environments
๐ Beats frontier giants on FS-Olympiad, FS-Research, CMT & PinchBench, and hits 61.2% on SWE-Bench Pro
๐ป Runs locally via vLLM๏ผ@vllm_project๏ผ, SGLang๏ผ@sgl_project๏ผ, llama.cpp & LM Studio (Q4_K_M ~20GB)
Scale the Horizon, Compose the Capabilities ๐
๐ Model:
@huggingface
@ModelScope2022
๐น๏ธ Try it free:
๋ ๋ณด๊ธฐ
