snowflakedb/ArcticInference

arctic-inference-architecture

Orientation map for the ArcticInference vLLM plugin: what each feature does, which file implements it, how it is patched onto vLLM, and the knob that turns it on. Covers the ArcticPatch framework, the plugin bootstrap order, and the seven feature groups…

Langue du texte source : anglais

arctic-vllm-versioning

Explains how ArcticInference targets and enforces its vLLM version. Covers the target-version declaration (`VLLM_PATCH_VERSION` in `arctic_inference/utils.py`), the unpinned `vllm` install extra, the runtime plugin gate (`arctic_inference.vllm.plugin` +…

Langue du texte source : anglais

rebase-arctic-inference

Rebase the ArcticInference vLLM plugin from one pinned vLLM version to a newer one. Use when asked to rebase, bump, upgrade, or port ArcticInference to a new vLLM release (e.g. "rebase ArcticInference to vLLM 0.25", "bump the vllm pin"), or to reconcile…

Langue du texte source : anglais