mangadict-models is an open-weight model from Craig Kolb, released under Apache License 2.0. Its published files total 151.4 MB.
ONNX models used by the mangadict browser extension to find and read Japanese text in manga pages.
Model Card
By Craig Kolb, published under apache-2.0, revision f30a1f4cdec5.
ONNX models used by the mangadict browser extension to find and read Japanese text in manga pages. They are repackaged from the models below for in-browser inference with onnxruntime-web; see those repos for training details and their licenses (all Apache-2.0). Use tagged revisions (e.g. v1); main may change.
Read Craig Kolb's full model card
ONNX models used by the mangadict browser extension to find and read Japanese text in manga pages. They are repackaged from the models below for in-browser inference with onnxruntime-web; see those repos for training details and their licenses (all Apache-2.0).
| Path | Source | Notes |
|---|---|---|
detector/v4-s_int8.onnx |
ogkalu/comic-text-and-bubble-detector (detector-v4-s_int8.onnx) |
Unmodified copy. RT-DETR; classes: bubble, text_bubble, text_free. |
ocr-2025-kv/encoder.onnx, cross.onnx, step.onnx, config.json |
jzhang533/manga-ocr-base-2025, a fine-tune of kha-white/manga-ocr-base | fp32 re-export with a KV cache: encoder (image → features), cross (per-image cross-attention keys/values), step (one decoder step: previous token + cache → logits + updated cache). Greedy decoding matches the original model. |
ocr-2025/vocab.txt |
kha-white/manga-ocr-base | Character vocabulary (shared by base and 2025). |
Use tagged revisions (e.g. v1); main may change.
Identity and Version
- Repository
- craigkolb/mangadict-models
- Publisher
- Craig Kolb
- Task
- Not stated by the source
- Modality
- Other
- Library
- Not stated by the source
- Parameters
- Not stated by the source
- Languages
- ja
- Revision
- f30a1f4cdec52dfac4c7862ec5c9f41f1ff66b01
- First published
- 2026-09-28
- Last updated
- 2026-09-28
Files and Weights
8 files, 151.4 MB in total. The weights are 4 files totalling 151.4 MB in onnx.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| detector/v4-s_int8.onnx | Weights | 11.1 MB | 5fe9e4f576e4 |
| ocr-2025-kv/cross.onnx | Weights | 10.0 MB | 94a628f0b942 |
| ocr-2025-kv/encoder.onnx | Weights | 22.3 MB | 1a7fd905b83c |
| ocr-2025-kv/step.onnx | Weights | 107.9 MB | 5cb58f567686 |
| ocr-2025-kv/config.json | Configuration | 63 B | — |
| README.md | Documentation | 1.3 KB | — |
| .gitattributes | Repository | 1.5 KB | — |
| ocr-2025/vocab.txt | Tokenizer | 24.1 KB | — |
License and Download
- License
- apache-2.0
- Access
- Open weights, no gate
- Download size
- 151.4 MB
Released by Craig Kolb through its official repository on Hugging Face. Read the license.
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 151.4 MB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About mangadict-models
Can I use mangadict-models commercially?
Yes. mangadict-models is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.