Support DeepSeek V4 Flash Vision Experimental

This commit is contained in:
Georg Bauer committed 2026-10-04 20:32:19 +02:00
1 parent 7c18e738cd
commit 9bbe12e6b3
52 files changed
+4988 -100

No files matched your search

+4
View File
@@ -40,6 +40,10 @@ the Metal kernels and small native integration layers are adapted from
sampling, cancellation, idle unload, and a bounded resident session pool.
Supported acceleration includes SSD expert streaming, DeepSeek legacy MTP,
DSpark, GLM MTP, directional steering, and model-specific memory admission.
- **DeepSeek V4 Flash Vision Experimental.** Model Manager offers the
Vision Experimental main model, required image encoder, and matching DSpark
drafter. Image input works in chat, tool results, and the HTTP APIs, with
image-aware KV reuse. The 0731 drafter is not interchangeable with this model.
- **Local HTTP APIs.** A configurable listener on `127.0.0.1:4000` exposes
Models, Chat Completions, Completions, Anthropic Messages, and Responses APIs
with streaming, tool calls, usage accounting, cancellation, and prefix-aware