ds4 · Issues· 681 open

[RFC] Evaluate PrismML Bonsai 2 support in ds4

Shutdown checkpoint is never found on restart when the conversation began after another live conversation in the same process

`--batched-session N`: a request that shares only the system prompt with a live slot evicts it instead of taking an empty slot

Server: multimodal sessions lose all KV reuse on a tail token mismatch — no common-prefix rewind, no disk-KV fallback, no continued disk stores

ds4-server: expose tensor-parallel roles (field report: 2x M4 Max TB5, TP 22.7 t/s vs server-pipeline ~20)

Premature compacting when using ds4-agent

Show and tell: DS Menu Bar - lightweight macOS control for ds4-server

ROCm build fails: host-side `rsqrtf()` in `rocm/ds4_rocm_deepseek4_vision.cuh`

Support for QWEN3.8-Flash-Next

GLM-5.3-Flash-Q4_K Metal SSD-streaming (M4 Max, 128 GiB): host-side per-layer bookkeeping dominates token time (15 % GPU), and the async selected-id load is a no-op

Progressive JPEG: AC refinement scan desynchronises mid-stream, ds4_image_decode_memory fails on files libjpeg decodes cleanly

test_deepseek41_metal: new router check fails on M1 Ultra (pre-M5); sqrt(softplus) compared at 3e-6 but 1.0f+exp is ~0.7% imprecise

ds4-agent: `-p` one-shot mode fails with "failed to start line editor"

V4 Flash 0731 (Q4K chat-v2 imatrix): greedy / low-temp decoding enters a verbatim repetition loop on long analytical answers

FYI: DeepSeek v4.1 Flash weights are available