🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the Apple M1 Ultra 128GB run DEQ-Transformer (Medium, Adaptive Embedding)?
Yes — here's how fast and up to what context, with real measured numbers.
DEQ-Transformer (Medium, Adaptive Embedding) on the Apple M1 Ultra 128GB
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1732k | 2472.7-2981.8 t/s | ~30.9-37.3 (slow) |
| Q8 | yes | 1733k | 4945.5-5963.6 t/s | ~61.8-74.5 (slow) |
| Q4 | yes | 1734k | 8831.2-10649.4 t/s | ~110.4-133.1 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
→ All models on the Apple M1 Ultra 128GB
→ DEQ-Transformer (Medium, Adaptive Embedding) specs
→ Other hardware
FAQ
Can the Apple M1 Ultra 128GB run DEQ-Transformer (Medium, Adaptive Embedding)?▶
Yes. At Q4 it generates ~8831.2-10649.4 tok/s and fits up to 1734k context; beyond that it spills to SSD swap and slows to ~110.4-133.1 tok/s.
How much context fits?▶
At Q4, up to about 1734k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.