license: mit
datasets:
- svhn
language:
- en
- zh
metrics:
- accuracy
pipeline_tag: object-detection
tags:
- legal
The SVHN-Recognition model is implemented using a deep convolutional neural network in PyTorch. The dataset utilized for training is derived from the Google Street View House Numbers (SVHN) dataset, where each image contains a sequence of Arabic digits ranging from 0 to 9 composing a house number. After thorough testing, the model demonstrates an impressive accuracy of 89% in recognizing multi-digit house numbers. By learning the features of digits within street view images, this deep learning model successfully achieves accurate recognition of multi-digit house numbers. The model not only showcases its ability to efficiently handle multi-digit numbers in real-world scenarios but also provides a reliable and flexible implementation solution within the PyTorch framework for house number recognition tasks.
Maintenance
GIT_LFS_SKIP_SMUDGE=1 git clone git@hf.co:MuGeminorum/SVHN-Recognition
Training curve
Mirror
https://www.modelscope.cn/models/MuGeminorum/SVHN-Recognition
Reference
[1] https://github.com/MuGeminorum/SVHN-Recognition
[2] Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks