# Request: Offline AI Model Conversion Tool Access for Traffic Sign Detection POC **URL:** \ **Category:** AI\ **Created:** [January 4, 2026, 5:03am UTC](https://forum.amebaiot.com/t/request-offline-ai-model-conversion-tool-access-for-traffic-sign-detection-poc/4634 "2026-01-04T05:03:37Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![krish\_tyger](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.amebaiot.com/krish_tyger/32/3012_2.png) [@krish\_tyger](https://forum.amebaiot.com/u/krish_tyger)\ **Post date:** [January 4, 2026, 5:03am UTC](https://forum.amebaiot.com/t/request-offline-ai-model-conversion-tool-access-for-traffic-sign-detection-poc/4634/1 "2026-01-04T05:03:37Z")
Hello, I’m working on a traffic sign detection POC using AMB82-MINI and need access to the offline AI model conversion toolkit (Acuity Toolkit). **What I need:** - Access to ameba-ai-offline-toolkit GitHub repository - Acuity conversion tools to convert TFLite/H5 model to .nb format **Current status:** - Have AMB82-MINI board working - Camera and DEFAULT\_IMGCLASS tested successfully - Have traffic sign detection model (TFLite format) - Need to convert to .nb format for deployment **What I tried:** - Emailed [AmebaAIoT@realtek.com](mailto:AmebaAIoT@realtek.com) with subject “Offline AI Model” - no response yet - Online conversion tool at [amebaiot.com](http://amebaiot.com) - not receiving email with converted model **Request:** Could someone from Realtek team please grant access or provide alternative download method for the offline conversion toolkit? My use case: ADAS lite system for road signs detection Thank you! ---
**Author:** ![KevinKL](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.amebaiot.com/kevinkl/32/2632_2.png) [@KevinKL](https://forum.amebaiot.com/u/KevinKL)\ **Post date:** [January 5, 2026, 2:24am UTC](https://forum.amebaiot.com/t/request-offline-ai-model-conversion-tool-access-for-traffic-sign-detection-poc/4634/2 "2026-01-05T02:24:18Z")
Hello, We have replied to your email request for the offline toolkit. Regarding your online model conversion issue, may I know what model are you using currently? Would you mind to share your model with us so that we can help to verify the problem? Thank you. ---
**Author:** ![krish\_tyger](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.amebaiot.com/krish_tyger/32/3012_2.png) [@krish\_tyger](https://forum.amebaiot.com/u/krish_tyger)\ **Post date:** [January 8, 2026, 9:19am UTC](https://forum.amebaiot.com/t/request-offline-ai-model-conversion-tool-access-for-traffic-sign-detection-poc/4634/3 "2026-01-08T09:19:58Z")
Hello Kevin, Thank you for reaching out. I am currently working with a **custom YOLOv7-tiny** model (Pytorch) for road sign detection on the **AmebaPro2 (AMB82-mini)**. 1. **ONNX Layer Names:** When converting my YOLOv7-tiny `.pt` to `.onnx` for the offline tool, which specific **output nodes** should I target to avoid the ‘No Bounding Box’ error?" 2. **Widescreen Grid Math:** “For a custom **576x320** resolution, does the offline toolkit automatically handle the non-square grid mapping, or do I need to manually adjust the **Pegasus import** script’s outputs?” 3. **Reparameterization Necessity:**"Is the `reparam_yolov7-tiny.py` script required even for offline conversion, or does the toolkit handle reparameterization during the `.nb` export? ---
**Author:** ![KevinKL](https://sea2.discourse-cdn.com/flex016/user_avatar/forum.amebaiot.com/kevinkl/32/2632_2.png) [@KevinKL](https://forum.amebaiot.com/u/KevinKL)\ **Post date:** [January 9, 2026, 9:35am UTC](https://forum.amebaiot.com/t/request-offline-ai-model-conversion-tool-access-for-traffic-sign-detection-poc/4634/4 "2026-01-09T09:35:54Z")
Hi @krish_tyger , 1. You might need to modify the `models/yolo.py` file in the YOLOv7 repo. Kindly refer to the guide [NN Example (Yolov7\_tiny)](https://ameba-doc-ai-video-analytics-doc.readthedocs-hosted.com/en/latest/user_manual/NN_quantization_discussion/NN_yolov7_example.html) to modify the forward function within class Detect. Please take note of the part of Convert PyTorch to onnx. You may also refer to this related forum thread, [YOLOv7-tiny .nb (NBG) model inference shows no bounding boxes on Amb82-mini](https://forum.amebaiot.com/t/yolov7-tiny-nb-nbg-model-inference-shows-no-bounding-boxes-on-amb82-mini/4588) 2. No, you do not need to manually adjust the pegasus import script’s outputs, it’ll be automatically handled by the toolkit. 3. Yes, reparameterization script is needed as the toolkit does not handle reparameterization process. Thank you.