Downloads · 30 days
333
1% of all-time downloads
jashepp/Equinox-31B-i1-MXFP4-GGUF
Equinox-31B-i1-MXFP4-GGUF is a text generation model from jashepp. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as apache-2.0.
Equinox draws its name from the balance between extremes. Trained on a balanced blend of Wayfarer 2's unforgiving dark adventures and Hearthfire's quiet slice-of-life storytelling, Equinox is equally at home in perilo…
Downloads · 30 days
333
1% of all-time downloads
All-time downloads
23.8K
Public
Repo size
93.5 GB
Likes
4
Public
Click a slice to open those files.
.gguf93.5 GB · 100%
From the Hugging Face model README
Equinox draws its name from the balance between extremes. Trained on a balanced blend of Wayfarer 2's unforgiving dark adventures and Hearthfire's quiet slice-of-life storytelling, Equinox is equally at home in perilous dungeons and candlelit conversations.
Equinox 31B was trained with two epochs of SFT (Supervised Fine-Tuning) on top of Gemma 4 31B Instruct.
This repository contains custom, highly optimized, multi-tier mixed precision GGUF weights for LatitudeGames/Equinox-31B.
In order of quality:
| Filename | Size | Quants |
|---|---|---|
| Equinox-31B-MXFP4_Q8_0_F16-Imatrix.gguf | 22.9 GB | MXFP4_MOE + Q8_0 + F16 |
| Equinox-31B-MXFP4_Q6_K_Q8_0-Imatrix.gguf | 19.5 GB | MXFP4_MOE + Q6_K + Q8_0 |
| Equinox-31B-MXFP4_Q8_0-Imatrix.gguf | 17.7 GB | MXFP4_MOE + Q8_0 |
| Equinox-31B-MXFP4_Q6_K-Imatrix.gguf | 17.1 GB | MXFP4_MOE + Q6_K |
| Equinox-31B-MXFP4-Only-Imatrix.gguf | 16.3 GB | MXFP4_MOE Only |
Updated 2026-07-07:
Equinox-31B-MXFP4-Only-Imatrix.ggufUpdated 2026-06-28:
MXFP4_Q8_0 & MXFP4_Q6_K
model-quants-imatrix style namingThe following datasets were used for the imatrix:
I am currently remaking this imatrix with 2048 ctx, 3000 chunks, with smaller chunks & sets of text
Old/Current: Custom Target Matrix (2048 ctx with 600 chunks), with all of the following:
Standard global quantization presets (like stock MXFP4) compress the backbone layers uniformly, which degrades the quality of models.
This repository provides multiple distinct manual configuration layouts to balance precision and memory constraints:
Equinox-31B-MXFP4_Q8_0_F16-Imatrix.gguf - Designed for maximum quality preservation, this layout implements a strict 3-Tier Precision Matrix:
token_embd.weight - Protects the critical input/output vocabulary mappings. Dramatically prevents text degradation.attn_q, attn_k, attn_v, attn_output, attn_qkv - All core attention layers. Keeps the model's primary attention mechanism high quality.ffn_down, ffn_gate, ffn_up - Downsamples the massive, heavy feed-forward network layers to hardware-optimized 4-bit microscaling blocks.Equinox-31B-MXFP4_Q8_0-Imatrix.gguf - Designed for a slightly leaner memory profile, this layout utilizes 2-Tier Precision:
token_embd), final logit projections (output), layer normalizations, and structural layers fall back to the heavy Q8_0 format.attn_k, attn_q, attn_output, ffn_down, ffn_gate, ffn_up - The core attention projections and dense feed-forward network blocks are target-quantized directly to MXFP4 to leverage highly optimized hardware kernels.Equinox-31B-MXFP4_Q6_K_Q8_0-Imatrix.gguf - Designed for a tighter VRAM footprint while preserving quality, this layout implements a strict 3-Tier Precision Matrix:
token_embd.weight - Keeps the critical input/output vocabulary mappings. Lessens the text degradation.attn_q, attn_k, attn_v, attn_output, attn_qkv - All core attention layers. Keeps the model's primary attention mechanism at decent quality.ffn_down, ffn_gate, ffn_up - Downsamples the massive, heavy feed-forward network layers to hardware-optimized 4-bit microscaling blocks.Equinox-31B-MXFP4_Q6_K-Imatrix.gguf - Designed for an even tighter VRAM footprint while preserving the mathematical depth of the logical layers, this layout utilizes 2-Tier Precision:
token_embd), final logit projections (output), layer normalizations, and structural layers fall back to the heavy Q6_K format.attn_k, attn_q, attn_output, ffn_down, ffn_gate, ffn_up - The core attention projections and dense feed-forward network blocks are target-quantized directly to MXFP4 to leverage highly optimized hardware kernels.Equinox-31B-MXFP4-Only-Imatrix.gguf - Using only MXFP4, this shrinks the model down to 16.3 GB. The quality is not the best, but it can still do decent work.
MXFP4, for speed and a low VRAM footprint.These files were converted via llama-quantize utilizing the following manual recipe parameters:
llama-quantize \
--tensor-type "token_embd\.weight=F16" \
--tensor-type "blk\..*\.(ffn_down|ffn_gate|ffn_up)\.weight=MXFP4" \
--imatrix "imatrix.gguf" \
"Equinox-31B_Unified_BF16.gguf" \
"Equinox-31B-MXFP4_Q8_0_F16-Imatrix.gguf" \
Q8_0
llama-quantize \
--tensor-type "blk\..*\.(attn_k|attn_q|attn_output|ffn_down|ffn_gate|ffn_up)\.weight=MXFP4" \
--imatrix "Equinox-31B-combined-imatrix.gguf" \
"Equinox-31B_Unified_BF16.gguf" \
"Equinox-31B-MXFP4_Q8_0-Imatrix.gguf" \
Q8_0
llama-quantize \
--tensor-type "token_embd\.weight=Q8_0" \
--tensor-type "blk\..*\.(ffn_down|ffn_gate|ffn_up)\.weight=MXFP4" \
--imatrix "imatrix.gguf" \
"Equinox-31B_Unified_BF16.gguf" \
"Equinox-31B-MXFP4_Q6_K_Q8_0-Imatrix.gguf" \
Q6_K
llama-quantize \
--tensor-type "blk\..*\.(attn_k|attn_q|attn_output|ffn_down|ffn_gate|ffn_up)\.weight=MXFP4" \
--imatrix "imatrix.gguf" \
"Equinox-31B_Unified_BF16.gguf" \
"Equinox-31B-MXFP4_Q6_K-Imatrix.gguf" \
Q6_K
llama-quantize \
--tensor-type "token_embd\.weight=MXFP4" \
--tensor-type "blk\..*\.(attn_k|attn_q|attn_v|attn_output|ffn_down|ffn_gate|ffn_up)\.weight=MXFP4" \
--imatrix "imatrix.gguf" \
"Equinox-31B_Unified_BF16.gguf" \
"Equinox-31B-MXFP4-Only-Imatrix.gguf" \
MXFP4_MOE
I'm doing this as a side hobby, with my AMD 5900X, 64GB DDR4, RTX 3060 12GB & RTX 5060 Ti 16GB.
I've found this model works best up to 35k context before it starts hallucinating, with it worsening when it reaches 45k, 64k, etc.
It works well for SillyTavern with CharMemory and MessageLimit.
It also works well enough for story-writing analysis.
As a passionate developer, I'm always programming, automating, or experimenting with new ideas.
I love building open-source tools, trying out new web tech, and creating things that don't yet exist, including local AI & quantizing models.
I love sharing these creations to give back to the community.
If my projects have saved you time or helped you out, consider supporting my work below!
Equinox-31B base model.BF16 GGUF quants of Equinox-31B.