Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU-NVFP4
NVFP4 checkpoint of DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU. NVFP4 here means W4A16 with FP8 scales, group size 16, weight only. Linear layers are NVFP4, vision tower, linear attention path, lm_head and MTP head are kept in BF16 at the source. No calibration data was used. This matches the pattern used for other Qwen3.8 Cold Fusion NVFP4 builds.
HF repo: esatapedico/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU-NVFP4
GGUF family: esatapedico/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU-NVFP4-GGUF (8 tiers, see the GGUF model card)
At a glance
| Field | Value |
|---|---|
| Format | compressed-tensors nvfp4-pack-quantized |
| Quantization | W4A16, group size 16, FP8 E4M3 scales, weight only |
| Kept in BF16 | vision tower, linear attention path, lm_head, embeddings, MTP head |
| Calibration | none |
| Base | DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU (Qwen3.8-27B, Apache 2.0) |
Quantized set covers MLP on all 64 layers plus Q, K, V, O on the 16 full attention layers. The DeltaNet path stays BF16 in this checkpoint and is normalized to NVFP4 in the GGUF tier builds (448 tensor backbone).
Checkpoint
- Single
model.safetensorsabout 26 GB config.jsonquantization_config.format=nvfp4-pack-quantized,quant_method=compressed-tensors,Qwen3_5ForCausalLM, 64 layers, hybrid GatedDeltaNet, 262144 context, MTP headtokenizer.jsonintact, chat template intact- NVFP4 tensors verified via metadata checks
Provenance
Derivative of DavidAU/Qwen3.8-27B-TURBO-Fable-Cold-Fusion-735-882-Heretic-Uncensored-NM-DAU (Apache 2.0) which itself derives from Qwen/Qwen3.8-27B. See the GGUF model card for full attribution. This checkpoint is the source for the GGUF family, which contains 1,122 tensors per file with a shared 448 tensor NVFP4 backbone.
Follow along & support
I post updates on new conversions, benchmarks, and what I'm working on over on my site. Follow along there to keep up with new releases and the work in progress. If you'd like to support more of it, you will find the options there. I do this on consumer hardware and like seeing how far it goes. More is on the way.
🌐 esatapedico.com. Updates, work-in-progress, and ways to support.
License
apache-2.0
Card written with AI assistance.