Downloads · 30 days
112
9% of all-time downloads
cstr/smoldocling-GGUF
smoldocling-GGUF is a machine learning model from cstr. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
GGUF conversions of ds4sd/SmolDocling-256M-preview for CrispEmbed inference.
Downloads · 30 days
112
9% of all-time downloads
All-time downloads
1.2K
Public
Repo size
2 GB
Likes
1
Trending 1
Click a slice to open those files.
.gguf1 GB · 100%
From the Hugging Face model README
GGUF conversions of ds4sd/SmolDocling-256M-preview for CrispEmbed inference.
Ultra-compact document conversion model (256M params). Generates DocTags structured markup from page images — OCR, layout, tables, formulas, code, charts.
| File | Quant | Size | Notes |
|---|---|---|---|
smoldocling-f16.gguf | F16 | 491 MB | Full precision |
smoldocling-q8_0.gguf | Q8_0 | 261 MB | Recommended |
smoldocling-q4_k.gguf | Q4_K | 153 MB | Max compression |
Parity vs HF reference: vision cos=0.9998, connector cos=0.9999.
# CLI
./crispembed -m smoldocling-q8_0.gguf --ocr document.png
# Server
./crispembed-server --ocr smoldocling-q8_0.gguf --port 8080
curl -X POST http://localhost:8080/math/ocr -F "[email protected]"
from crispembed import CrispMathOcr
ocr = CrispMathOcr("smoldocling-q8_0.gguf")
doctags = ocr.recognize("document.png")
print(doctags) # <doctag><text>...</text>...</doctag>
Apache-2.0 — same as the base model.
Original model by Docling Team, IBM Research. GGUF conversion and inference engine by CrispEmbed.