Qwen3.5-4B-MNN (mirror with MTP head)
Published by TokForge for on-device inference.
TokForge
- Website: https://tokforge.ai
- Discord: https://discord.gg/Acv3CBtfVm
- Google Play: https://play.google.com/store/apps/details?id=dev.tokforge
- iOS TestFlight: https://testflight.apple.com/join/jnufjzRr
Runs on-device in the TokForge app.
This repo mirrors taobao-mnn/Qwen3.5-4B-MNN at commit aa966261175a532d9906fa165c9d506d617320a9 byte for byte
(every file below has the same sha256 as the original; the MNN conversion and all credit for it belong to the
taobao-mnn team, the model to the Qwen team) and adds an optional multi-token-prediction (MTP) head next to the
unchanged trunk, used by TokForge on the MNN CPU route.
Trunk files (identical to taobao-mnn/Qwen3.5-4B-MNN @ aa96626117)
| file | bytes | sha256 |
|---|---|---|
config.json |
652 | 92853033efe602f95efca3e1c05cd8b108f973c8beed417843a9671f8147ed8d |
export_args.json |
1,040 | fb827952ef5e5b8133d931d19123171f3681132932d3d7682c0169da7d562b92 |
llm.mnn |
3,651,096 | 45eebc9dc433a55f0388453250a05a728a206650294be92f48e4fdd1b0348e0f |
llm.mnn.json |
9,172,204 | d2e7f54aea01444ec1bd89edc0a3b9b5203cc00fdf909474c877a79a45147b1b |
llm.mnn.weight |
2,629,387,626 | a31d50e5668ce6b22655ff4750cfd48b972790ae883f2c154ba34b9173b38de1 |
llm_config.json |
8,693 | 8da9ee370f627577f95875a6bf915e2358d6a6d7a1eb3d16272987379dd9ceca |
tokenizer.txt |
6,465,727 | 7e75de1f279a10b65bd9dc1a5207205cb8993823861c4c42bbbd74e48e1c23a4 |
visual.mnn |
488,096 | b9a865dee1ceffc83eeca8ee0a8e7d5b911960b18ac93003d1135ea903a2bc31 |
visual.mnn.weight |
196,768,960 | cc51f3aa1e026c853bdd22e9188c51b121b7bdcfcef6030f4254ef297b615ab5 |
MTP head (optional add-on)
Exported from Qwen/Qwen3.5-4B @ 851bf6e806efd8d0a36b00ddf55e13ccb7b8cd0a with MNN-TokForge llmexport --mtp --mtp_only --quant_bit 4 --quant_block 64
and a 32k draft vocabulary (mtp-v32k.mnn.vocab, int32 token ids). Load with speculative_type: "mtp",
mtp_model: "mtp-v32k.mnn", draft_predict_length: 1. The trunk verifies every drafted token, so output quality is
the trunk's own.
| file | bytes | sha256 |
|---|---|---|
mtp-v32k.mnn.vocab |
132,160 | 80af0c499240994c092bb3e79fe4d32a4e9e8dbc82fbbf8d728158f11fbae1b3 |
mtp-v32k.mnn.weight |
128,337,094 | a0fa0d3c556321dfdeacb477599000c0f729d5997914f54a5418226f27db6d06 |
mtp-v32k.mnn |
34,024 | 23fb866877ee07a9e162a2cfbe7a3d414b92b9dec7b0af7dc0844d5a3b23d18c |
License: Apache-2.0, as the original model and conversion.
Credits
- Qwen3.5-4B by the Qwen team (Alibaba Cloud) (Hugging Face). Used under the Apache-2.0 license.
- The MNN conversion (every trunk file here is byte-identical to it) by the taobao-mnn team (Hugging Face). Used under the Apache-2.0 license.
- The MTP head is exported from Qwen/Qwen3.5-4B (Apache-2.0) by TokForge with llmexport from TokForge's fork of MNN.