HuggingFaceBio releases FineDNA: 300B+ token DNA-and-text dataset behind Carbon-X, with full recipe
_lewtun · x · 2026-11-02
HuggingFaceBio published the FineDNA collection — the complete training data behind Carbon-X, billed as the best open model for DNA understanding — featuring 300B+ tokens of DNA and text plus mid-training, SFT and RL subsets, with the training recipe open-sourced on GitHub.
The collection also includes Carbon-X-Base (a 15B model mid-trained from Qwen3-14B on FineDNA), the SFT variant, a Carbon annotation database, mid-training evals and pure-DNA benchmarks — a rare fully open stack for AI-for-Science researchers.
Related event: Hugging Face Open-Sources FineDNA Dataset and Carbon DNA Model Family(11 posts)→
More from Research
- Legora CTO hand-writes 100k-param model weights with Fable, one-shot beats training — dair_ai · 2026-11-04
- 20-40% of alignment data is noisy: a loss-reduction scoring framework to fix it — burkov · 2026-11-04
- Scientists image quantum entangled photons in real time, and the result looks like a taijitu — CurieuxExplorer · 2026-11-04
- kalomaze: crude diffusion autoencoders may lose low-frequency info at the latent stage — kalomaze · 2026-11-04
- Structured captions boost text-to-music alignment, new Music DSG metric beats CLAP — affige_yang · 2026-11-04
- Fruit fly's 166,000-neuron circuit map is being used to drive robot behavior — CurieuxExplorer · 2026-11-04