MOSS Image Text Encoder Heretic needs about 7.4 GB of memory at Q4_K_M (a 5.8 GB file) with an 8K context window, so it fits on a 8 GB GPU. This repository has 12 quants, from 3.9 GB (Q2_K) to 18.4 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.4 GB.
MOSS Image Text Encoder Heretic – 9.2B Qwen3.5 model for vision and prompt enhancement
MOSS Image Text Encoder Heretic is a 9.2B parameter Qwen3.5 model adapted for vision and prompt enhancement. It removes refusals from the original MOSS encoder while keeping all other weights identical. The model supports coding, reasoning, tools, and vision tasks.
Good at
- Vision‑to‑text encoding
- Prompt enhancement for MOSS‑Image
- Adult‑content handling without refusals
Limits
- No vision support beyond encoding
- Large context requires high VRAM
- License is Apache‑2.0
Best for
- Generating image captions
- Enhancing prompts for image generation
- Encoding text for DiT models
Who it’s for. Developers building image‑generation pipelines that need a robust text encoder. Researchers testing prompt‑enhancement without refusal behavior. Users of MOSS‑Image who want a decensored encoder.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 3.9 GB | 5.5 GB | Fits | Get file → | |
| Q3_K_S | 4.4 GB | 6.0 GB | Fits | Get file → | |
| Q3_K_M | 4.7 GB | 6.3 GB | Fits | Get file → | |
| Q3_K_L | 5.0 GB | 6.7 GB | Fits | Get file → | |
| IQ4_XS | 5.4 GB | 7.0 GB | Fits | Get file → | |
| Q4_K_S | 5.5 GB | 7.1 GB | Fits | Get file → | |
| Q4_K_M | 5.8 GB | 7.4 GB | Tight | Get file → | |
| Q5_K_S | 6.5 GB | 8.1 GB | Too big | Get file → | |
| Q5_K_M | 6.6 GB | 8.2 GB | Too big | Get file → | |
| Q6_K | 7.6 GB | 9.2 GB | Too big | Get file → | |
| Q8_0 | 9.8 GB | 11.4 GB | Too big | Get file → | |
| F16 | 18.4 GB | 20.0 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_S (5.49 GB file, 7.1 GB memory). For 12 GB VRAM, Q5_K_S (6.47 GB file, 8.1 GB memory) or Q5_K_M (6.64 GB file, 8.2 GB memory). For 16 GB VRAM, Q6_K (7.56 GB file, 9.2 GB memory) or Q8_0 (9.79 GB file, 11.4 GB memory).
What GPU do you need for MOSS Image Text Encoder Heretic?
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: no quant fits at an 8K context.
- 6 GB VRAM: the best quant that fits is Q3_K_S (4.4 GB file, about 6.0 GB in use).
- 8 GB VRAM: the best quant that fits is Q4_K_M (5.8 GB file, about 7.4 GB in use).
- 12 GB VRAM: the best quant that fits is Q8_0 (9.8 GB file, about 11.4 GB in use).
- 16 GB VRAM: the best quant that fits is Q8_0 (9.8 GB file, about 11.4 GB in use).
- 24 GB VRAM: the best quant that fits is F16 (18.4 GB file, about 20.0 GB in use).
Which MOSS Image Text Encoder Heretic quant should you choose?
Q4_K_M is the usual starting point for MOSS Image Text Encoder Heretic: it balances file size and quality well. Here it is a 5.8 GB file.
The Q8_0 file is 9.8 GB, about 1.7 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.
If memory is tight, the smallest file is Q2_K at 3.9 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about MOSS Image Text Encoder Heretic
How much VRAM does MOSS Image Text Encoder Heretic need?
About 7.4 GB at Q4_K_M with an 8K context window. That is the 5.8 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 MOSS Image Text Encoder Heretic on an 8 GB GPU?
Yes, with Q4_K_M: the 5.8 GB file needs about 7.4 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of MOSS Image Text Encoder Heretic GGUF?
It depends on the quant: Q2_K is 3.9 GB, Q3_K_S is 4.4 GB, Q3_K_M is 4.7 GB, Q3_K_L is 5.0 GB, IQ4_XS is 5.4 GB, Q4_K_S is 5.5 GB and 6 more.
What context length does MOSS Image Text Encoder Heretic support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.4 GB at 32K.
Where can I download MOSS Image Text Encoder Heretic GGUF?
The files are on Hugging Face in the mradermacher/MOSS-Image-Text-Encoder-Heretic-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 MOSS Image Text Encoder Heretic use?
Hugging Face lists the license as apache-2.0. 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 this model on 8 GB VRAM?
Yes, use the Q4_K_S quant which requires about 7.1 GB of VRAM at 8k context.
Does this model support vision tasks?
It provides vision‑to‑text encoding but does not perform image generation or vision inference beyond encoding.
Is the model licensed for commercial use?
The model is released under the Apache‑2.0 license, which permits commercial use.