Amazon Rekognition vs. Pytorch

Amazon Rekognition is well suited for all image and videos analysis. Also, deep learning projects for image and video can easily be done using this. It is very easy to use so a beginner can also use it.

Incentivized

They have created Pytorch Lightening on top of Pytorch to make the life of Data Scientists easy so that they can use complex models they need with just a few lines of code, so it's becoming popular. As compared to TensorFlow(Keras), where we can create custom neural networks by just adding layers, it's slightly complicated in Pytorch.

Incentivized

• Image tagging is very good.
• Object detection is precise.
• Video tagging has become very easy using Rekognition.

Incentivized

• flexibility
• Clean code, close to the algorithm.
• Fast
• Handles GPUs, multiple GPUs on a single machine, CPUs, and Mac.
• Versatile, can work efficiently on text/audio/image/tabular datasets.

Incentivized

• The cost is bit more for small scale companies.
• For text processing, OCR is not available in amazon rekognition.
• Only facial search is available in image search.

Incentivized

• Since pythonic if developing an app with pytorch as backend the response can be substantially slow and support is less compares to Tensorflow

Incentivized

Very much suitable for many applications where the image processing features are secondary and independent of any domain. This makes it a general solution and the recognition features are returned in a JSON object in response to the API called made which is a very simple process.

Incentivized

The big advantage of PyTorch is how close it is to the algorithm. Oftentimes, it is easier to read Pytorch code than a given paper directly. I particularly like the object-oriented approach in model definition; it makes things very clean and easy to teach to software engineers.

Incentivized

Support of Amazon is great. So far we are having a great experience using it.

Incentivized

It's easy to implement and its documentation is great compared to CNTK.

Incentivized

Pytorch is very, very simple compared to TensorFlow. Simple to install, less dependency issues, and very small learning curve. TensorFlow is very much optimised for robust deployment but very complicated to train simple models and play around with the loss functions. It needs a lot of juggling around with the documentation. The research community also prefers PyTorch, so it becomes easy to find solutions to most of the problems. Keras is very simple and good for learning ML / DL. But when going deep into research or building some product that requires a lot of tweaks and experimentation, Keras is not suitable for that. May be good for proving some hypotheses but not good for rigorous experimentation with complex models.

Incentivized

• The best thing it made our work fast.
• It reduces the chance of failure of a project which is based on image or video processing.
• We can assign anyone in our team to work on this as it very easy to use.

Incentivized

• The ability to make models as never before
• Being able to control the bias of models was not done before the arrival of Pytorch in our company

Incentivized