cdiamond/Qwen3.8-27B-iMatrix-NVFP4-MTP-GGUF
image-text-to-text model by cdiamond
From the publisher
Model card & documentation
Source previewRead the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.
I built this quant because the ready-made FP4 file answered the wrong question. It was fast, but on my short WikiText-2 control it scored 6.4949 PPL. Plain Q40 scored 6.3798. The first higher-quality hybrid went too far the other way: good perplexity, 34.19 tok/s, and no comfortable room for 256K plus vision. This is the build that survived both gates. It is a 17.1 GB, 5.01 BPW mixed-precision GGUF of Qwen/Qwen3.8-27B. It keeps large, tolerant matrices in native NVFP4 and spends more bits on selected attention, Gated DeltaNet, and late FFN tensors. The trained MTP layer remains embedded in the same GGUF. This is not a fine-tune. I built the private calibration workload from 5,472 messages across 296 real sessions in my Hermes agent setup. The 153,600 processed tokens covered coding, infrastructure work, tool calls, and mixed Polish-English conversation. I used llama-imatrix to rank tensor sensitivity; the corpus and raw…
Read the full model card ↗ · Preview checked 2026-10-29T02:07:42.876Z
Inside the original model card — Document outline
- Qwen3.8-27B iMatrix NVFP4 MTP GGUF
- Files
- Quant recipe
- Measured results
- The llama.cpp branch behind 55.402 tok/s
- Run it on current llama.cpp
- The measured max-throughput profile
- Why MTP stops at eight
- Known limitation: batch invariance
- Provenance and license
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Documentation belongs to its respective authors. Reported project/model license: apache-2.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
Model overview
cdiamond/Qwen3.8-27B-iMatrix-NVFP4-MTP-GGUF is a image-text-to-text model repository published by cdiamond on Hugging Face. A library was not reported.
This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.
Model facts
- Task
- image-text-to-text
- Library
- Not reported
- Recent downloads (30 days)
- 3,201,804
- Cumulative likes
- 46
- Hugging Face trending score
- 2
- Reported safetensors parameters
- Not reported
- Architecture
- Not reported
- License
- apache-2.0
- Access
- Not gated by Hugging Face
- Created
- 2026-08-17T22:29:42.000Z
- Last modified
- 2026-08-17T22:40:35.000Z
Compatibility and lineage
Language coverage was not reported.
Reported base models: Qwen/Qwen3.8-27B
Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.
Use and evaluate
Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.
No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.
Source and freshness
Source: Hugging Face Hub. Metadata observed 2026-11-01T06:51:30.840Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.