This month both QNAP and Synology Announced the availability of new packages built around Machine Learning and AI.
QNAP First .
QNAP is targeting the Ameatures programmers and data scientists with its QuAI Developer's package. A package that can help the data scientists build ,train and deploy their machine Learning Models right on QNAP.
This was highly anticipated ever since QNAP started to ship out machines with GPU Support. Possible applications of a GPU on QNAP could be - Cryptocurrency mining ( yes, you can do that on QNAP), Gaming ( with HDMI port connected to your TV) and off course training your ML model.
Its worth Noting that QuAI is just package with all the popular libraries built in - a Container. As such, you don't need to have QuAI to write your ML code. You could easily run a Virtual machine on QNAP and download whichever library you work with. For Example I could have Ubuntu running in a container ( or a Virtual machine) on QNAP and install python, NLTK and Tensorflow on Ubuntu.
What QuAI does facilitate is -
A) it makes creating the environment fast and easy. Within minutes you could have your entire framework and environment setup.
B) It can make use of the GPU.
You have option of installing QuAI with major popular libraries like Tensorflow, Caffe, Caffe2, MS CNTK and MXNET. How much processing power you get for training and deploying your ML model depends on your GPU.
Now about Synology. Synology has Applied Machine learning and released the beta version of Synology Moments. Moments is an app that can group your photos into similar groups based on the faces, places and objects. Technically it is Unsupervised Clustering done by your NAS , that's AI !
So far the options for managing photos on NAS relied heavily on EXIF metadata , and manual tags. For example if you took a photo with your phone with the GeoTag photo settings and saved it to your NAS , you would have a photo which has Exif data of the Location it was clicked on and a Date. With this NAS would give you the option to filter the photos according to date or Location. That's as far as you can go with this approach. If you wanted anything more, you would have to TAG the photo manually.
With Moments it changes for Synology users. Now the machine will scan the photos for know patterns , and for patterns within the images and cluster them together. It will further tag the photos too so you can search based on the tags. On the models with slower processors, only facial recognition is supported , and higher models support both face and subject recognition. You must have node.js installed on Synology for moments to work.
Its interesting to see the NAS vendors offering packages built around machine learning. This is just the beginning and I can foresee a lot many more applications on NAS coming up. Both QNAP and Synology Surveillance apps are going to deploy ML for pattern recognition for sure. Right now the only limiting factor for these vendors is the amount of processing power they have.
QNAP First .
QNAP is targeting the Ameatures programmers and data scientists with its QuAI Developer's package. A package that can help the data scientists build ,train and deploy their machine Learning Models right on QNAP.
This was highly anticipated ever since QNAP started to ship out machines with GPU Support. Possible applications of a GPU on QNAP could be - Cryptocurrency mining ( yes, you can do that on QNAP), Gaming ( with HDMI port connected to your TV) and off course training your ML model.
Its worth Noting that QuAI is just package with all the popular libraries built in - a Container. As such, you don't need to have QuAI to write your ML code. You could easily run a Virtual machine on QNAP and download whichever library you work with. For Example I could have Ubuntu running in a container ( or a Virtual machine) on QNAP and install python, NLTK and Tensorflow on Ubuntu.
What QuAI does facilitate is -
A) it makes creating the environment fast and easy. Within minutes you could have your entire framework and environment setup.
B) It can make use of the GPU.
You have option of installing QuAI with major popular libraries like Tensorflow, Caffe, Caffe2, MS CNTK and MXNET. How much processing power you get for training and deploying your ML model depends on your GPU.
Now about Synology. Synology has Applied Machine learning and released the beta version of Synology Moments. Moments is an app that can group your photos into similar groups based on the faces, places and objects. Technically it is Unsupervised Clustering done by your NAS , that's AI !
So far the options for managing photos on NAS relied heavily on EXIF metadata , and manual tags. For example if you took a photo with your phone with the GeoTag photo settings and saved it to your NAS , you would have a photo which has Exif data of the Location it was clicked on and a Date. With this NAS would give you the option to filter the photos according to date or Location. That's as far as you can go with this approach. If you wanted anything more, you would have to TAG the photo manually.
With Moments it changes for Synology users. Now the machine will scan the photos for know patterns , and for patterns within the images and cluster them together. It will further tag the photos too so you can search based on the tags. On the models with slower processors, only facial recognition is supported , and higher models support both face and subject recognition. You must have node.js installed on Synology for moments to work.
Its interesting to see the NAS vendors offering packages built around machine learning. This is just the beginning and I can foresee a lot many more applications on NAS coming up. Both QNAP and Synology Surveillance apps are going to deploy ML for pattern recognition for sure. Right now the only limiting factor for these vendors is the amount of processing power they have.
