Atlas inference library

A shared inference library built on vLLM that serves evaluation, internal, production, and RL rollout workloads.

Also called Atlas inference library on vLLM.

source: 1
model: 1
lab adopt it: 1
strongest: used

How sources treat it

One count per evidence span, weakest treatment to strongest.

used 1

Documented in

Laguna-S-2.1

Evidence

1 span quoted from the sources, strongest treatment first.

Atlas serves every downstream consumer: automated evaluations, internal inference, customer-facing production traffic, but also crucially RL rollout generation.

usedinference servingPoolside

laguna m1 xs2 technical report·Poolside

Filed alongside

Other methods under software implementation :: inference engine.

vLLM SGLang TensorRT-LLM KTransformers Unsloth Dedicated serving engines Environment variable propagation to subprocesses litert-lm serve llama.cpp vLLM language-model-only mode vLLM or SGLang serving vLLM tensor parallelism vLLM-Ascend xLLM