Downloads · 30 days
0
chibifire/hailo-ugen300-artifacts
hailo-ugen300-artifacts is a machine learning model from chibifire. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Shelved binaries, scripts and logs from work on the UGen300 (Hailo-10H) accelerator.
Downloads · 30 days
0
Access
Public
Updated Oct 11, 2026
Repo size
7 GB
Likes
0
Public
Click a slice to open those files.
.onnx8.6 GB · 100%
From the Hugging Face model README
Shelved binaries, scripts and logs from work on the UGen300 (Hailo-10H) accelerator.
The repo keeps the state of the UGen300 work as it was left, so the work can restart without a new export.
It holds ONNX exports of the Qwen3-VL-4B vision tower with the EditScore merger LoRA merged, raw and onnxsim-simplified, at two opsets.
It also holds a DeiT control model that translates cleanly, the QZO rung ladder's HAR and parameter archives, and compiled HEF files.
The hailo/, vit/ and mezo/ folders mirror the working directories: probes, benches, result files, SDK logs and the built hailort_shim.dll.
The apparatus that made these files (shim, probes, rung ladders, export_vit.py) is in its own repo, linked below.
logbook-qwen3vl-vit-translate-walls.md in https://github.com/weftspun/request-for-discussion.hf download chibifire/hailo-ugen300-artifacts --local-dir hailo-ugen300-artifacts
hf download chibifire/hailo-ugen300-artifacts --include "vit/qwen3vl_vit*.onnx" --local-dir hailo-ugen300-artifacts
Run the files with the scripts in the apparatus repo.
Apache 2.0, as named in the header. See the Apache License 2.0 terms. The base model Qwen/Qwen3-VL-4B-Instruct is also Apache 2.0.