Daedalus-150M: A Convolution-Attention Hybrid Designed for CPU Inference
Quick summary
arXiv:2608.20210v1 Announce Type: cross Abstract: Small language models are usually built like large ones and then squeezed onto a CPU afterwards. We did the opposite: we fixed the target first, one user, one token at a time, 4-bit weights, ordinary CPU, and chose the architecture to suit it. The result keeps full attention in only 6 of its 18 blocks. The other 12 use short convolutions whose memory is two timesteps wide no matter how long the conversation gets, so two thirds of the network never re-reads a growing cache. Trained from scratch on 59.9B tokens, the model scores 47.31 on a five-t
Key takeaways
- arXiv:2608.20210v1 Announce Type: cross Abstract: Small language models are usually built like large ones and then squeezed onto a CPU afterwards.
- We did the opposite: we fixed the target first, one user, one token at a time, 4-bit weights, ordinary CPU, and chose the architecture to suit it.
- The result keeps full attention in only 6 of its 18 blocks.
Why it matters
This model development creates a new option for users and a new testing obligation for developers. A fixed evaluation set comparing quality, cost and failure behavior is more useful than launch claims.
Member comments