Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

SSD300 Inference Tutorial (weights.h5 shape error)

See original GitHub issue

I’m using ssd300_inference.ipynb for the first time.

I’m just following the steps and I imported from gDrive the weights trainval35k: SSD300.

The code gives me the following error:

ValueError: Layer #25 (named "conv4_3_norm_mbox_conf"), weight <tf.Variable 'conv4_3_norm_mbox_conf/kernel:0' shape=(3, 3, 512, 84) dtype=float32_ref> has shape (3, 3, 512, 84), but the saved weight has shape (324, 512, 3, 3)

Can you help me?

Issue Analytics

State:
Created 4 years ago
Comments:5

Top GitHub Comments

2reactions

anystroemcommented, Apr 25, 2019

You might have to change the number of classes in the model from 20 to 80 if you use weights trained for MS COCO. I had the same error as you have and that fixed it for me.

0reactions

MaxxTrcommented, Dec 29, 2020

Actually I have the same issue. Do you know what was the reason for such problem ?

Top Results From Across the Web

Dimension mismatch error in Jetson-inference ssd model ...

The mask shape is coming from shape of labels. I am unable to track where is the labels's shape getting set to [4,...

Acceleration of deep convolutional neural networks on ...

The most common layers are convolutional layers consisting of weights that projects the input to create a weighted output. Figure 2 shows an ......

Object Detection on Large-Scale Egocentric Video Dataset

They tested the model on the dataset created by Ren and Philipose [19] and outperformed [21] by 19% in segmentation error rate. The...

Object Detection and Tracking on a Raspberry Pi using ...

As even the most light-weight state of the art object detection models, i.e. Tiny-YOLO and. SSD300 with MobileNet, were considered too computationally ...

Object Detection using PyTorch and SSD300 with VGG16 ...

In this tutorial, we will be using an SSD300 (Single Shot Detector) deep learning object detector along with the PyTorch framework for ...