/

/

google-research

Обзор Центр заботыВойти

google-research

Ветки: 311 Коммиты: 4490 Теги: 0

..

google-research

/

nopad_inception_v3_fcn

Fix a couple minor bugs with the training code for 129 receptive field.

3 года назад

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

download_and_convert_data.py

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

eval_inception_v3_fcn.py

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

finetune_on_flowers.sh

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

inception_base_129.py

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

inception_base_911.py

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

inception_v3_fcn.py

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

network_params.py

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

requirements.txt

Update some imports and dependencies needed to work in a third party environment.

3 года назад

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

train_inception_v3_fcn.py

Open-sourcing the code for "CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor". https://arxiv.org/abs/2312.07661

9 месяцев назад

README.md

Inception v3 in Fully Convolutional Network mode used in the following paper: An Augmented Reality Microscope with Real-time Artificial Intelligence Integration for Cancer Diagnosis.

This is a variant of inception v3 by removing all the paddings and adopting the principle of fully convolutional network design. This change allows the network to be trained and inference run with different patch size while having the same inference results.

This folder can be downloaded using subversion: svn export https://github.com/google-research/google-research/trunk/nopad_inception_v3_fcn.

An example uptraining script is provided at finetune_on_flowers.sh. Note the setup instructions detailed in the file.

Использование cookies

Мы используем файлы cookie в соответствии с Политикой конфиденциальности и Политикой использования cookies.

Нажимая кнопку «Принимаю», Вы даете АО «СберТех» согласие на обработку Ваших персональных данных в целях совершенствования нашего веб-сайта и Сервиса GitVerse, а также повышения удобства их использования.

Запретить использование cookies Вы можете самостоятельно в настройках Вашего браузера.