NapYang/Qwen3.8-27B-pi-3.80bpw-fp16

image-text-to-text model by NapYang

Open original source ↗

Older snapshot: check the original source for current information.

From the publisher

Model card & documentation

Source preview

Read the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.

A per-layer quantized Qwen3.8-27B multimodal model for Apple Silicon via MLX. Rationale: K/V projections are most sensitive to quantization error (attention scores), so they retain 8-bit. Q projection tolerates 3-bit. Linear-attention fused projections use 4-bit as a middle ground. Rationale: First and last layers are most distribution-sensitive → higher precision (5-bit). Core body aggressively compressed to 3-bit. gateproj uses larger groupsize (128) for smoother statistics. All three animated SVGs below were generated in one shot from the same prompt: The only difference is the reasoning effort (low / medium / xhigh).…

NapYang/Qwen3.8-27B-pi-3.80bpw-fp16 on Hugging Face A short preview, not the full document.

Read the full model card ↗ · Preview checked 2026-11-21T12:29:45.339Z

Inside the original model card — Document outline
  1. Qwen3.8-27B-pi-3.80bpw-fp16 (MLX)
  2. Per-Layer Quantization Strategy
  3. Global Defaults
  4. Attention Layers — Two Distinct Paths
  5. MLP Layers — U-Shaped Precision Profile
  6. MTP (Multi-Token Prediction) Module
  7. Vision Encoder
  8. Model Card Visuals: Animated Pelican on Motorcycle (3 Reasoning Efforts)
  9. Low reasoning effort
  10. Medium reasoning effort
  11. Xhigh reasoning effort
  12. Usage (MLX)

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

NapYang/Qwen3.8-27B-pi-3.80bpw-fp16 is a image-text-to-text model repository published by NapYang on Hugging Face. The source reports the mlx library.

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
mlx
Recent downloads (30 days)
0
Cumulative likes
0
Hugging Face trending score
0
Reported safetensors parameters
27,781,427,952
Architecture
Qwen3_5ForConditionalGeneration
License
apache-2.0
Access
Not gated by Hugging Face
Created
2026-10-07T06:37:55.000Z
Last modified
2026-10-07T06:40:00.000Z

Compatibility and lineage

Reported languages: en, zh

Reported base models: bytkim/Qwen3.8-27B-pi

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.

Model card and usage instructions

Model files and configuration

Community discussion

Source and freshness

Source: Hugging Face Hub. Metadata observed 2026-10-07T06:40:04.949Z. Daily imports are snapshots, not real-time monitoring.

Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.

Open original source

Related models

microsoft/Florence-2-base

ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF

Qwen/Qwen3.8-27B-FP8

Qwen/Qwen3.8-27B

zai-org/GLM-5.3-Flash

Qwen/Qwen3.5-9B