Humanizer needs about 20.8 GB of memory at Q4_K_M (a 7.4 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 25 quants, from 0.0 GB to 9.8 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 51.5 GB.
12B model that rewrites AI drafts to read like human writing
Humanizer is a 12B gemma4 model trained to rewrite AI‑written drafts in English and Chinese. It preserves all factual details such as numbers, dates, names, and quotes while adding no new content. The model runs locally and is available in multiple GGUF quantizations.
Good at
- Rewriting AI drafts to read human
- Preserving factual details
- Local offline operation
Limits
- Requires 8‑k context for full performance
- Large VRAM needed for higher‑bit quants
- No vision support
Best for
- Email and essay rewriting
- Report and forum post editing
- Content polishing
Who it’s for. Developers and content creators who need to transform AI‑generated text into natural‑sounding drafts. The model is suitable for offline use on machines with sufficient VRAM and RAM.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.0 GB | 13.4 GB | Too big | Get file → | |
| IQ1_S | 3.0 GB | 16.4 GB | Too big | Get file → | |
| IQ1_M | 3.2 GB | 16.6 GB | Too big | Get file → | |
| IQ2_XXS | 3.6 GB | 17.0 GB | Too big | Get file → | |
| IQ2_XS | 3.9 GB | 17.3 GB | Too big | Get file → | |
| IQ2_S | 4.1 GB | 17.5 GB | Too big | Get file → | |
| IQ2_M | 4.4 GB | 17.8 GB | Too big | Get file → | |
| Q2_K_S | 4.5 GB | 17.9 GB | Too big | Get file → | |
| Q2_K | 4.8 GB | 18.2 GB | Too big | Get file → | |
| IQ3_XXS | 4.8 GB | 18.2 GB | Too big | Get file → | |
| IQ3_XS | 5.3 GB | 18.7 GB | Too big | Get file → | |
| IQ3_S | 5.5 GB | 18.9 GB | Too big | Get file → | |
| Q3_K_S | 5.5 GB | 18.9 GB | Too big | Get file → | |
| IQ3_M | 5.7 GB | 19.1 GB | Too big | Get file → | |
| Q3_K_M | 6.1 GB | 19.5 GB | Too big | Get file → | |
| Q3_K_L | 6.6 GB | 20.0 GB | Too big | Get file → | |
| IQ4_XS | 6.6 GB | 20.0 GB | Too big | Get file → | |
| IQ4_NL | 7.0 GB | 20.4 GB | Too big | Get file → | |
| Q4_0 | 7.0 GB | 20.4 GB | Too big | Get file → | |
| Q4_K_S | 7.0 GB | 20.4 GB | Too big | Get file → | |
| Q4_K_M | 7.4 GB | 20.8 GB | Too big | Get file → | |
| Q4_1 | 7.7 GB | 21.0 GB | Too big | Get file → | |
| Q5_K_S | 8.3 GB | 21.7 GB | Too big | Get file → | |
| Q5_K_M | 8.5 GB | 21.9 GB | Too big | Get file → | |
| Q6_K | 9.8 GB | 23.2 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (≈20.8 GB memory at 8k context) or Q4_1 (≈21 GB). If you have 12 GB VRAM, Q5_K_M (≈21.9 GB) is a good balance. For 16 GB VRAM, Q6_K (≈23.2 GB) offers the best quality. File sizes are listed in the GGUF files section.
What GPU do you need for Humanizer?
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: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: the best quant that fits is the main file (0.0 GB file, about 13.4 GB in use).
- 24 GB VRAM: the best quant that fits is Q6_K (9.8 GB file, about 23.2 GB in use).
Which Humanizer quant should you choose?
Q4_K_M is the usual starting point for Humanizer: it balances file size and quality well. Here it is a 7.4 GB file.
If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Humanizer
How much VRAM does Humanizer need?
About 20.8 GB at Q4_K_M with an 8K context window. That is the 7.4 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 Humanizer on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 13.4 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Humanizer GGUF?
It depends on the quant: file is 0.0 GB, IQ1_S is 3.0 GB, IQ1_M is 3.2 GB, IQ2_XXS is 3.6 GB, IQ2_XS is 3.9 GB, IQ2_S is 4.1 GB and 19 more.
What context length does Humanizer support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 12.9 GB at 8K and 51.5 GB at 32K.
Where can I download Humanizer GGUF?
The files are on Hugging Face in the mradermacher/humanizer-i1-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 Humanizer 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 Humanizer on 8 GB VRAM?
Yes, with the Q4_K_M or Q4_1 quant, which require about 20.8–21 GB of VRAM at 8k context. Lower‑bit quants are not available for this model.
Does Humanizer support Chinese text?
Yes, it is trained on both English and Chinese drafts and preserves factual details in both languages.
Is there a vision version of Humanizer?
No, Humanizer is a text‑only model; no mmproj files are provided.