thederby/Qwen3.5-4B-oQ4e-fp16-mtp
image-text-to-text model by thederby
Older snapshot: check the original source for current information.
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.
oQ4e mixed-precision MLX quantization of Qwen/Qwen3.5-4B, made with oMLX oQe (v0.7.1.dev1). This is the fp16 build. The multi-token-prediction (MTP) head is kept and the vision tower is included. The same family at every bit width and dtype (this repo in bold): Which dtype? bf16 (no suffix) is the default. fp16 builds store the language model's non-quantized weights and the quantization scales as float16 instead of bfloat16. They score the same at 4- and 5-bit and are marginally closer to the original at 6- and 8-bit. oMLX notes that float16 can give faster prefill on M1/M2, which lack native bfloat16. They are about 0.7 GB larger because the vision tower is kept in float32; vision quality was not measured. How these were measured: mean KL divergence of each quant against the original bf16 model, over 48 windows x 2048 tokens of English prose, Python code, Japanese and Korean text. Less than…
Read the full model card ↗ · Preview checked 2026-10-30T16:01:56.112Z
Inside the original model card — Document outline
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
thederby/Qwen3.5-4B-oQ4e-fp16-mtp is a image-text-to-text model repository published by thederby 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
- 4,659,865,088
- Architecture
- Qwen3_5ForConditionalGeneration
- License
- apache-2.0
- Access
- Not gated by Hugging Face
- Created
- 2026-10-10T06:32:50.000Z
- Last modified
- 2026-10-10T06:33:01.000Z
Compatibility and lineage
Language coverage was not reported.
Reported base models: Qwen/Qwen3.5-4B
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 2026-10-10T06:41:01.047Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.