Hugging Face's logo
Lucas Silva 's picture
25 1 26

Lucas Silva PRO

YuLexuan30
Fataj's profile picture marduk191's profile picture rooox's profile picture
ยท

AI & ML interests

None yet

Recent Activity

reacted to abdurrahmanbutler's post with ๐Ÿค— 12 minutes ago
๐Ÿš€ ๐—œ๐—ป๐˜๐—ฟ๐—ผ๐—ฑ๐˜‚๐—ฐ๐—ถ๐—ป๐—ด ๐—ž๐—ฎ๐—ป๐—ผ๐—ป ๐Ÿฎ ๐—˜๐—ป๐—ฟ๐—ถ๐—ฐ๐—ต๐—ฒ๐—ฟ: ๐˜๐—ต๐—ฒ ๐˜„๐—ผ๐—ฟ๐—น๐—ฑโ€™๐˜€ ๐—ณ๐—ถ๐—ฟ๐˜€๐˜ ๐—ต๐—ถ๐—ฒ๐—ฟ๐—ฎ๐—ฟ๐—ฐ๐—ต๐—ถ๐—ฐ๐—ฎ๐—น ๐—ด๐—ฟ๐—ฎ๐—ฝ๐—ต๐—ถ๐˜๐—ถ๐˜‡๐—ฎ๐˜๐—ถ๐—ผ๐—ป ๐—บ๐—ผ๐—ฑ๐—ฒ๐—น Today weโ€™re publicly releasing Kanon 2 Enricher, and with it, an entirely new class of AI model that weโ€™re calling a hierarchical graphitization model. This is fundamentally different from both universal extraction models and generative models. As a hierarchical graphitization model, Kanon 2 Enricher natively outputs a ๐—ธ๐—ป๐—ผ๐˜„๐—น๐—ฒ๐—ฑ๐—ด๐—ฒ ๐—ด๐—ฟ๐—ฎ๐—ฝ๐—ต rather than tokens, which makes it architecturally incapable of hallucinating or inventing text that wasnโ€™t present in the input. What that enables in practice is unlike any other model or ML architecture on the market: โ€ข ๐—ก๐—ผ ๐—ต๐—ฎ๐—น๐—น๐˜‚๐—ฐ๐—ถ๐—ป๐—ฎ๐˜๐—ถ๐—ผ๐—ป๐˜€ ๐Ÿค– It cannot hallucinate. All references and links are stored as spans, meaning exact character offsets anchored to the original text. โ€ข ๐—›๐—ถ๐—ฒ๐—ฟ๐—ฎ๐—ฟ๐—ฐ๐—ต๐—ถ๐—ฐ๐—ฎ๐—น ๐˜€๐—ฒ๐—ด๐—บ๐—ฒ๐—ป๐˜๐—ฎ๐˜๐—ถ๐—ผ๐—ป, ๐—ป๐—ผ๐˜ ๐—ท๐˜‚๐˜€๐˜ ๐—ฒ๐˜…๐˜๐—ฟ๐—ฎ๐—ฐ๐˜๐—ถ๐—ผ๐—ป ๐Ÿ“‘ It deconstructs a documentโ€™s full nested hierarchy, down to chapters, sections, clauses, schedules, signatures, and even singular sentences, and classifies each span with dozens of contextual features. โ€ข ๐—˜๐—ป๐˜๐—ถ๐˜๐˜† ๐—ฒ๐˜…๐˜๐—ฟ๐—ฎ๐—ฐ๐˜๐—ถ๐—ผ๐—ป, ๐—ฑ๐—ถ๐˜€๐—ฎ๐—บ๐—ฏ๐—ถ๐—ด๐˜‚๐—ฎ๐˜๐—ถ๐—ผ๐—ป, ๐—ฎ๐—ป๐—ฑ ๐—น๐—ถ๐—ป๐—ธ๐—ถ๐—ป๐—ด ๐Ÿ”— It resolves what references actually point to, then links entities, citations, and cross-references into a single coherent graph. โ€ข ๐—š๐—ฟ๐—ฎ๐—ฝ๐—ต-๐—ณ๐—ถ๐—ฟ๐˜€๐˜ ๐—ฒ๐—ณ๐—ณ๐—ถ๐—ฐ๐—ถ๐—ฒ๐—ป๐—ฐ๐˜† ๐Ÿƒโ€โžก๏ธ Small enough to run locally on a consumer PC with sub-second latency, and it stays reliable on long documents where front To read more about our new model, check out our latest Hugging Face article: https://huggingface.co/blog/isaacus/introducing-kanon-2-enricher
liked a Space 22 days ago
multimodalart/FLUX.2-dev-turbo
reacted to marksverdhei's post with ๐Ÿค— about 1 month ago
Dear Hugging Face team, can we please have a way to archive hf repositories / spaces? I have a bunch of spaces that used to work but don't any more due to the hf space implementations changing and i think it would be good if I could archive those like in GitHub. React to this post if you want to see this feature! ๐Ÿ’ก
View all activity

Organizations

lorafrenzi's profile picture