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…
Organization · Verified on Hugging Face
Zosma AI
zosmaai
Models in Library1
Datasets in Library0
Models on Hugging Face2
Followers3
Compare the published evidence
This profile connects the 1 model and dataset record shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Published tasks: Document text extraction.
Reported licenses: apache-2.0. Open a record to review its specific license and conditions.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.