Yesterday
-
Add got-ocr2-f16-r2.gguf: reconverted F16 (old got-ocr2-f16.gguf has a zeroed 8 MiB block in embed_tokens; kept) [CrispEmbed 145e5d724d48]
3 days ago
-
License: weights inherit tongyi-qianwen-research (see NOTICE, LICENSE)
-
docs: add provenance / EU AI Act Art. 53 note
-
Upload got-ocr2-ref-full.gguf with huggingface_hub
-
Upload got-ocr2-q8_0.gguf with huggingface_hub
-
Upload README.md with huggingface_hub
-
Upload got-ocr2-q4_k.gguf with huggingface_hub
-
Upload README.md with huggingface_hub
-
docs: F16-decoder requirement + accurate sizes/parity (#25)
-
fix: keep 0.5B decoder at F16 (all-q4_k broke OCR); decoder f16 + vision q8_0 (#25)
-
fix: keep 0.5B decoder at F16 (q8_0 decoder broke OCR, cos 0.936); vision q8_0 (#25)
-
Upload README.md with huggingface_hub
-
Upload got-ocr2-f16.gguf with huggingface_hub
-
Upload got-ocr2-q4_k.gguf with huggingface_hub
-
Upload got-ocr2-q8_0.gguf with huggingface_hub
-
initial commit
