Skip to content

docs(readme): TurboKV precision known issue on M5 (not reproduced on M6) #174

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
solderzzc merged 2 commits into from
Sep 24, 2026
Merged

docs(readme): TurboKV precision known issue on M5 (not reproduced on M6) #174

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
a709b77
docs(readme): TurboKV precision known issue on M5 (not reproduced on M6)
Sep 24, 2026
a5b8f36
docs(readme): link #175 / #176 from the known issues; recommend --str…
Sep 24, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 3 additions & 1 deletion README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -121,7 +121,9 @@ TurboKV barely changes speed on this model. Only 16 of its 64 layers use full at
3. **An auto-detected VLM that failed to load exited the server.** `Qwen3.6-35B-A3B-UD-MLX-4bit` ships a `preprocessor_config.json` without `image_mean`. SwiftLM now falls back to text-only unless you pass `--vision`.
4. **Vision-capable models skipped chunked prefill.** On the older mlx-swift-lm pin, a text-only prompt on the VLM path ran through the model in a single pass. It's fixed by the mlx-swift-lm bump in #167. Every number in this section was measured on `main` with that bump.

> ⚠️ **Known issues:** `--gpu-layers N` (CPU/GPU layer partitioning) hits a Metal GPU timeout on the first request (repro: `--model mlx-community/gemma-4-26b-a4b-it-4bit --gpu-layers 23`). QAT-quantized Gemma 4 MTP assistants (`…-qat-assistant-4bit`) fail with `unhandledKeys pre_projection/post_projection`; use `gemma-4-26B-A4B-it-assistant-bf16`.
> ⚠️ **`--turbo-kv` precision on M5 (not reproduced on M6):** on an Apple M5, Qwen3.8-27B-4bit with `--turbo-kv` gets exact long-range lookups wrong from somewhere between 2K and 5K prompt tokens. Asked how many numbered lines a prompt has, it answers "1,000" or "14" instead of 315 / 500 / 700. Without `--turbo-kv` it answers correctly, and on the M6 both modes are 24/24 correct from 2K to 11.8K tokens with the same prompt. The likely cause is a GPU-family-dependent path in TurboKV's dequant or attention kernels. Until it's fixed, avoid `--turbo-kv` on M5 when exact recall matters. Tracked in [#175](https://github.com/SharpAI/SwiftLM/issues/175).
>
> ⚠️ **Known issues:** `--gpu-layers N` (CPU/GPU layer partitioning) hits a Metal GPU timeout on the first request, on both M5 and M6 (repro: `--model mlx-community/gemma-4-26b-a4b-it-4bit --gpu-layers 23`). Tracked in [#176](https://github.com/SharpAI/SwiftLM/issues/176); the fix is [SharpAI/mlx-swift#17](https://github.com/SharpAI/mlx-swift/pull/17). Even once it's fixed, CPU-resident MoE layers are very slow (~0.4 tok/s prefill), so on a 32 GB Mac try `--stream-experts` first (Qwen3.6-35B-A3B: 13.2 tok/s decode, 7.7 GB GPU). QAT-quantized Gemma 4 MTP assistants (`…-qat-assistant-4bit`) fail with `unhandledKeys pre_projection/post_projection`; use `gemma-4-26B-A4B-it-assistant-bf16`.

Reproduce:

Expand Down
Loading