selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

Unlimited-OCR-2

Side-by-side specs, memory needs, licence, runtimes and benchmarks. Add up to 4 models.

--vram
any
--gpus
--quant
--add

On your GPU

  • Unlimited-OCR-2
Property
Overview
TaskOCR
Parameters3.3B
Active parameters1.1B
Context length128K
Input → output
Released26 Oct 2026
LicenceMIT
Gated downloadno
Memory and speed
Fits your hardwarepick a GPU above≈4.5 GB
VRAM at Q4≈2.8 GB
VRAM at Q8≈4.5 GB
VRAM at FP16≈8.0 GB
Est. generation speedpick a GPU above—
Popularity
Downloads / 30 days604.6K
Likes714
Runs with
llama.cpp
SGLang
Transformers
vLLM
Ollama tag—
Architecture
ArchitectureUnlimitedOCRForCausalLM
Layers12
Attention / KV heads10 / 10
Vocabulary129 280
Weights dtypebfloat16
GGUF availableyes

Keep comparing

Memory figures are estimates for an 8K context at the chosen quantisation (Q4_K_M for LLMs by default). Speeds are rough, memory-bound estimates. Self-reported benchmark scores come from the model authors.