Downloads · 30 days
208
9% of all-time downloads
Versi-StyleTune-31B-GGUF
Iloqt/Versi-StyleTune-31B-GGUF
Versi-StyleTune-31B-GGUF is a machine learning model from Iloqt. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
GGUF quants of Iloqt/Versi-StyleTune-head: a text-only Gemma 4 31B with Nimbz/Versipellis-31B as the base and the lmhead (output projection) grafted from Gryphe/Gemma-4-31B-StyleTune.
All-time downloads
2.4K
Public
Repo size
416 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.gguf416 GB · 100%
At a glance
- License
- apache-2.0
- Access
- Public
- Created
- Jun 17, 2026
- Updated
- Jul 17, 2026
- SHA
- 0ea02abe
Base models
Model card
From the Hugging Face model README
Versi-StyleTune-31B (GGUF)
GGUF quants of Iloqt/Versi-StyleTune-head: a text-only Gemma 4 31B with Nimbz/Versipellis-31B as the base and the lm_head (output projection) grafted from Gryphe/Gemma-4-31B-StyleTune.
All credits go to Nimbz and Gryphe for the original models, I only committed the merge.
Variants
Standard quants:
- Q2_K, Q3_K_M, Q4_K_M, Q5_K_M, Q6_K, Q8_0 — body-only quantization.
hb16 variants (head + embeddings kept at BF16):
- Q4_K_M-hb16, Q5_K_M-hb16, Q6_K-hb16
- Preserves the grafted StyleTune
lm_headandembed_tokensat full precision while quantizing the rest of the body to K-quant.
attn8-HB variants (Q8_0 attention + BF16 head + embeddings):
- Q4_K_M-attn8-HB, Q5_K_M-attn8-HB, Q6_K-attn8-HB
- Adds Q8_0 attention layers on top of the hb16 protection.
Notes
- Run with the Gemma 4 chat template; thinking off by default.
- Like all Gemma 4 models, benefits from repetition penalty or DRY to avoid token loops.
- License
- apache-2.0
- Created
- Jun 17, 2026
- Updated
- Jul 17, 2026