-
Hello,
during some experiments with classifying network I stumbled upon an issue where, if provided some pictures as a reference, the quantization process (I think) emits incorrect weights to divi…
-
I am trying to create a new model in Keras, based on an existing caffe model. The caffe model's last layer output is [N,50,50,2] which has softmax activation. After this it performs a slice to provide…
-
Hi! Thanks for sharing your code!
When running the code on my device it gets stuck after showing the Brownie logo (i can hear the voice as well). It fails on reading the kmodel,
`task = kpu.load("…
-
(This is an issue to be confirmed)
Related: https://github.com/kiali/kiali/pull/1711
I believe a unit in charts model set as "ms" for millisecond is going to be scaled as "kms" when x1000.
Solu…
jotak updated
5 years ago
-
nncase is failing when using a graph with intermediate tensors in: 32x128x128 out: 64x128x128.
[Sample tflite graph](https://github.com/kendryte/nncase/files/3795970/detect.tflite.zip)
`./ncc co…
-
@sunnycase
When I run Yolo model with Kendryte KPU.c compiled in debug mode, I get a list of yolo layers and the time taken for execution of each layer. However when I use the Ultra light fast face …
-
When I do an attack with the foolbox, for 6 pictures everything works fine, but the i get the following error:
> File "/home/wouter/anaconda3/lib/python3.6/site-packages/numpy/core/fromnumeric.py",…
-
If one wants to contribute in implementing/porting an op which is currently executed on CPU to KPU, what other background is required other than the knowledge of C/C++.
Can you provide some startin…
-
Create a graph with input node 256x256x3, an output node 128x128x32, 3x3 kernel, and **stride=2** [Example tflite attached - detect.tflite.zip](https://github.com/kendryte/nncase/files/3790490/detect.…
-
Hello guys,
we just try to convert our tflite model and got this error:
Fatal: Not supported tflite opcode: LOGISTIC
In the readme, LOGISTIC is one of the ops supported by the toolkit. I had tried …