NPU documentation and tools

Do you mean that there yolo darknet detection already done and trainable ? (Not classification)

The training needs to use PC. the NPU of vim3 can accelerate the operation of the model and can not be used as training. For the training on PC, you can refer to the one on DOC, which is verified by my practice. SSD and Yolo models do not classify the picture sets, but they need to label the pictures.

Got it. Thanks

What about the FPS you can achieve, in your test? Tiny yolo or YoloV3 ?

In PC, the FPS have 30 , but just 10 in board. The yoyl_v3 model is a little bit stuck, which is related to the calculation method of the model. I am still optimizing it. At the same time, I am training a yolo_v2 model, which will run higher FPS

1 Like

Thank you
It Worked!

root@Khadas:/npu/aml_npu_sdk/acuity-toolkit/conversion_scripts# ./
./ line 14: …/bin/convertensorflow: cannot execute binary file: Exec format error

root@Khadas:/npu/aml_npu_sdk/acuity-toolkit/conversion_scripts# file …/bin/convertensorflow
…/bin/convertensorflow: symbolic link to ./acuitylib/convertensorflow


You can’t run this tool on a AARCH64 architecture board, you can only run it on a X86_64 PC.

1 Like

@RichardG This tool just run on PC and tensorflow 1.10 .You can train with tensorflow-gpu 1.1x,but you need tensorflow 1.10 when you train.

thank you …


(venv) root@Khadas:/npu/darknet# pip3 install opencv-python
ERROR: Could not find a version that satisfies the requirement opencv-python (from versions: none)
ERROR: No matching distribution found for opencv-python


You can try this:

$ sudo apt-get update
$ sudo apt-get install opencv3

Just finished our first test of the VIM3’s NPU! @Frank did most of the heavy lifting. I just videoed everything with my iPhone.


Can we download this demo?

@RichardG yes you can, refer to the instructions inside the video. The part where Frank goes to this forum to download it: NPU Demo and source code

thanks …

1 Like

Hello , I have push a yolo tiny model to my gitlab . You can clone it .

the source code I have push to my gitlab too.

It should be noted that this is a fairly primitive model. I haven’t optimized the parameters yet, so the recognition effect is not very good,FPS is between 40 and 45, I will do more optimization later. This is just for users to try

Hi all,
I’m looking for a guideline on how to create a dataset.txt file which contains my 200 sample image filenames. This dataset.txt file is supposed to be input of option --source-file of the Acuity ‘tensorzonex’ quantization tool.
Thank in advance

@thanhhc just need to add you image file path to dataset.txt . Only one path per line.

1 Like

Thank for your reply