[
https://issues.apache.org/jira/browse/SINGA-232?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Xiangrui updated SINGA-232:
---------------------------
Description:
Implement Alexnet on Imagenet.
1. The model is following
https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf.
2. The data is created offline, including multiple training files, one test
file and one mean file. All of them are in binary format. This part is
implemented via writer, encoder in SINGA.
3. Loading data in multiple threads. This part needs reader, decoder and
transformer in SINGA.
4. This example need OpenCV support.
5. snapshot, jpgencoder, timer, binfile_reader are slightly modified.
was:
Implement Alexnet on Imagenet.
1. The model is following
https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf.
2. The data is created offline, including multiple training files, one test
file and one mean file. All of them are in binary format. This part is
implemented via writer, encoder in SINGA.
3. Loading data in multiple threads. This part needs reader, decoder and
transformer in SINGA.
4. This example need OpenCV support.
> Alexnet on Imagenet
> -------------------
>
> Key: SINGA-232
> URL: https://issues.apache.org/jira/browse/SINGA-232
> Project: Singa
> Issue Type: New Feature
> Reporter: Xiangrui
>
> Implement Alexnet on Imagenet.
> 1. The model is following
> https://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf.
> 2. The data is created offline, including multiple training files, one test
> file and one mean file. All of them are in binary format. This part is
> implemented via writer, encoder in SINGA.
> 3. Loading data in multiple threads. This part needs reader, decoder and
> transformer in SINGA.
> 4. This example need OpenCV support.
> 5. snapshot, jpgencoder, timer, binfile_reader are slightly modified.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
