A LoRA adapter that turns openbmb/MiniCPM-V-4.7-1B (open-weight ~1.5B VLM, Apache-2.0) into a focused document-OCR model for Devanagari (Hindi) + English text, including the conjuncts and mixed-script layouts where the base model is weakest. pure-Devanagari, mixed Hindi+English+digits, conjunct-heavy vocabulary (क्षेत्र, ज्ञान, त्रिभुज, …) gradient checkpointing, NaViT variable-shape slicing (4× downsample, ≤36 slices) Measured with eval/runeval.py (NF4, seed 0) + eval/metrics.py (NFC-normalized, whitespace-stripped Levenshtein CER). Per-bucket mean CER on the 50-image set (base → LoRA): The 50-image set is genuinely out-of-distribution (real fonts/layouts vs synthetic training), so the…
SAVRN Model Hub · Models by Task
Document Text Extraction Models
1 open-weight document text extraction models in the SAVRN Model Hub, with Zosma AI publishing the most.
1Models
1Publishers
1Licenses
Most Downloaded
| Model | Publisher | Parameters | License | Monthly downloads | Cheapest GPUs at 16-bit |
|---|---|---|---|---|---|
| zorro-ocr | Zosma AI | — | apache-2.0 | — | — |
Licenses
| License | Models | Commercial use |
|---|---|---|
| apache-2.0 | 1 | Yes |
Who Publishes Them
| Publisher | Models |
|---|---|
| Zosma AI | 1 |
All 1 Models
Questions
Which Document text extraction models are most downloaded?
By monthly downloads reported by the Hugging Face Hub: .