Rex Frame 3B needs about 2.6 GB of memory at IQ4_XS (a 1.8 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 5 quants, from 1.3 GB (IQ3_XXS) to 3.3 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 1.2 GB.
Rex Frame 3B – vision‑language model for comic and manga page analysis
Rex Frame 3B is a 3.09B parameter model fine‑tuned from Rex‑Omni to read comic and manga pages. It detects panels, characters, text and bubbles, transcribes dialogue with speaker attribution, and provides visual descriptions of characters and scenes. The model runs on a single consumer GPU or on llama.cpp on a laptop.
Good at
- Detects panels, characters, text and bubbles
- Transcribes dialogue with speaker pointers
- Provides visual descriptions of characters and scenes
Limits
- No vision support beyond comic pages
- Requires mmproj vision encoder
- License type is other
Best for
- Comic page OCR and transcription
- Character search and alt‑text generation
- Scene description for accessibility
Who it’s for. Developers building comic reading tools, accessibility solutions, or content analysis pipelines. It is suitable for users with a single GPU or laptop who need a lightweight vision‑language model.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ3_XXS | 1.3 GB | 2.1 GB | Fits | Get file → | |
| IQ3_M | 1.6 GB | 2.4 GB | Fits | Get file → | |
| IQ4_XS | 1.8 GB | 2.6 GB | Fits | Get file → | |
| Q6_K_XL | 2.5 GB | 3.3 GB | Fits | Get file → | |
| Q8_0 | 3.3 GB | 4.1 GB | Fits | Get file → |
For 8 GB VRAM, use IQ3_XXS (1.31 GB, ~2.1 GB memory at 8k context) or IQ3_M (1.59 GB, ~2.4 GB). For 12 GB VRAM, IQ4_XS (1.8 GB, ~2.6 GB) is suitable. If more VRAM is available, Q6_K_XL (2.55 GB, ~3.3 GB) or Q8_0 (3.29 GB, ~4.1 GB) provide higher precision.
What GPU do you need for Rex Frame 3B?
These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: the best quant that fits is Q6_K_XL (2.5 GB file, about 3.3 GB in use).
- 6 GB VRAM: the best quant that fits is Q8_0 (3.3 GB file, about 4.1 GB in use).
Which Rex Frame 3B quant should you choose?
IQ4_XS (1.8 GB) is the middle option in this repository. The usual sweet spot between file size and quality, and the quant most people start with.
If memory is tight, the smallest file is IQ3_XXS at 1.3 GB, with the quality trade-offs that come with fewer bits. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Rex Frame 3B
How much VRAM does Rex Frame 3B need?
About 2.6 GB at IQ4_XS with an 8K context window. That is the 1.8 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run Rex Frame 3B on an 8 GB GPU?
Yes, with Q8_0: the 3.3 GB file needs about 4.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of Rex Frame 3B GGUF?
It depends on the quant: IQ3_XXS is 1.3 GB, IQ3_M is 1.6 GB, IQ4_XS is 1.8 GB, Q6_K_XL is 2.5 GB, Q8_0 is 3.3 GB.
What context length does Rex Frame 3B support?
The GGUF header lists a trained context length of 128,000 tokens (125K). The KV cache grows with the context you use: about 0.3 GB at 8K and 1.2 GB at 32K.
Where can I download Rex Frame 3B GGUF?
The files are on Hugging Face in the miranda1000/Rex-Frame-3B-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does Rex Frame 3B use?
Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run Rex Frame 3B on 8 GB VRAM?
Yes, the IQ3_XXS or IQ3_M quantizations fit within 8 GB VRAM, using approximately 2.1–2.4 GB at 8k context.
Does Rex Frame 3B support other languages?
The model is fine‑tuned for comic and manga pages and does not specify language support beyond the training data.
What vision encoder does Rex Frame 3B use?
It uses the F16 mmproj vision encoder, shared across all quantizations.