Nicolò Monti
8 posts


free tours of Garbatella walking the paradogma for the first 10 who ask. pic related

Paradigma@paradigmainc
come join us in our natural habitat
English

github.com/erfanzar/Spect…
SpecTrax 0.1.0 with `sxregion_stage`
sxstage_region means multimodal MPMD can finally look like the model:
Vision path: V0 -> V1 -> V2 -> V3
Text path: T0 -> T1 -> T2 -> T3
One function. Separate logical pipelines. True forward/backward/scheduler MPMD underneath.
No fake stages, no SPMD cosplay.

English

Looks like Qwen wants to prevent distillation on their larger, closed-source models?
I somehow never looked too much at the top of Qwen3.5's CoTs, and it's mentioning instructions I never put in my prompt to prevent it from reciting its reasoning. Presumably an artifact from the RL done on the large teacher model?

English
Nicolò Monti retweetledi

Today we’re announcing Ternary Bonsai: Top intelligence at 1.58 bits
Using ternary weights {-1, 0, +1}, we built a family of models that are 9x smaller than their 16-bit counterparts while outperforming most models in their respective parameter classes on standard benchmarks.
We’re open-sourcing the models under the Apache 2.0 license in three sizes: 8B (1.75 GB), 4B (0.86 GB), and 1.7B (0.37 GB).

English
Nicolò Monti retweetledi

Just merged an external PR for Bonsai-8B support (1 bit LLM). Because tinygrad has the correct abstractions, it was 5 lines. huggingface.co/prism-ml/Bonsa… github.com/tinygrad/tinyg…
English

Nicolò Monti retweetledi
Nicolò Monti retweetledi
