Downloads · 30 days
382
20% of all-time downloads
gearwave00001/Huihui-Qwen3.8-27B-Quark-MXFP4-MagicQuant-GGUF
Huihui-Qwen3.8-27B-Quark-MXFP4-MagicQuant-GGUF is a text generation model from gearwave00001. Use it when you need the model to write or continue text. It is set up for quark. The card lists the license as apache-2.0.
The repository contains huihui-ai/Huihui-Qwen3.8-27B-abliterated converted to the AMD Quark MXFP4 format using the AMD specified 128 calibration samples @ 512 sequences. The resulting gearwave00001/Huihui-Qwen3.8-27B-…
Downloads · 30 days
382
20% of all-time downloads
All-time downloads
1.9K
Public
Repo size
34.8 GB
Likes
0
Public
Click a slice to open those files.
.gguf34.8 GB · 100%
From the Hugging Face model README
The repository contains huihui-ai/Huihui-Qwen3.8-27B-abliterated converted to the AMD Quark MXFP4 format using the AMD specified 128 calibration samples @ 512 sequences. The resulting gearwave00001/Huihui-Qwen3.8-27B-Quark-AWQ-MXFP4 was then converted to BF16 gguf and then MQ-IQ4_XS gguf using https://github.com/magiccodingman/MagicQuant.
magiccodingman/Qwen3.8-27B-MXFP4-MagicQuant-GGUF
MagicQuant is a benchmark driven GGUF hybrid discovery and validation system focused on finding real, practical GGUF quants specific to each architecture.
Whether it's a pure baseline model built by llama.cpp, learned tensor configurations from Unsloth, or a custom built MagicQuant hybrid, the model table below shows quants that have won dominance checks, survived collapse spaces, and/or were found to be nonlinearly better. Instead of dumping every quant type possible, MagicQuant tests, validates, and brutally murders anything deemed unworthy.
<details> <summary>Support MagicQuant</summary>I’m a solo developer working full time for myself to achieve my dream. I build open source code on the side. If you like any of my work, buying me a coffee is always appreciated. Otherwise, I hope you enjoy, maybe give me a star or something. Or just send me good vibes. Either way, thank you!
Click here to see ways to support - BTC, Paypal, GitHub sponsors.
</details> <details> <summary>Clone Notice</summary>This repository did not run through the full MagicQuant discovery pipeline. It is a clone of the final survivor tensor configurations from /mnt/NVME2/AI/MagicQuant-campaign/magicquant.clone-source.json, rebuilt and benchmarked locally for this model.
The archived MagicQuant JSON files in magicquant-manifest/ are copied from the source release for durability. The clone benchmark JSON and the table below are from this clone run, so those metrics reflect the rebuilt outputs in this repository.
| Name | Provider | KLD | Size (GB) | Download |
|---|---|---|---|---|
| MQ-IQ4_XS_1 | MagicQuant recipe clone | 0.126940 | 14.98 | Link |
External/custom baselines are normalized into MagicQuant's controlled comparison flow. MagicQuant rebuilds a learned baseline under native-source / MagicQuant-controlled conditions, including its own imatrix handling, so hybrids or external baselines (like Unsloth) can be judged on a more equal footing. That does not mean MagicQuant proved the original upstream artifact or upstream imatrix was worse. These comparisons exist for internal hybrid-search consistency and equal playing field comparisons, not as a universal judgment of the original creator's exact release artifact.
MagicQuant learns tensor quantization assignments and rebuilds from local source weights. It does not automatically reproduce a provider's additional weight transformations, calibration recipes, custom processing, or other techniques unless explicitly supported. These results are not a byte-for-byte reproduction or a test of the provider's original GGUF.
Easier to digest explanation:
MagicQuant compares and benchmarks the models quant to tensor configurations, but not the original artifact. And there's different reasons MagicQuant chooses to lift up a winning quant, not all winners are purely "better". It depends heavily on a variety of factors. Though choices are always documented in the repo under the manifest folder. You can always view what and why decisions were made by the automated system.
So, MagicQuant can confidently tell you, "under the same quantization to tensor configurations and identical imatrix, with this benchmark, I deemed this a winner".
</details> <details> <summary>Re-Uploading External Provider Baselines</summary>By default, if an external provider like Unsloth is deemed the winner, the repo should generally link directly to the original provider instead of re-hosting the quant. External GGUFs are normally only re-uploaded when a specific winning variant does not already exist (e.g. Heretic models or similar).
</details>