Dataiku vs. Gluon vs. Pytorch

Dataiku is an awesome tool for data scientists. It really makes our lives easier. It is also really good for non technical users to see and follow along with the process. I do think that people can fall into the trap of using it without any knowledge at all because so much is automated, but I dont think that is the fault of Dataiku.

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

• Allows users to collaborate and monitor individual tasks
• Caters to both types of analysts, coders and non-coders, alike
• Integrate graphs and plots with visualization tools such as Tableau

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 integrated windows of frontend and backend in web applications make it cumbersome for the developer.
• When dealing with multiple data flows, it becomes really confusing, though they have introduced a feature (Zones) to cater to this issue.
• Bundling, exporting, and importing projects sometimes create issues related to code environment. If the code environment is not available, at least the schema of the flow we should be able to import should be.

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

The user experience is very good. Everything feels intuitive and "flows" (sorry excuse the pun) so nicely, and the customization level is also appropriate to the tool. Even as a newer data scientist, it felt easy to use and the explanations/tutorials were very good. The documentation is also at a good level

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

The open source user community is friendly, helpful, and responsive, at times even outdoing commercial software vendors. Documentation is also top notch, and usually resolves issues without the need for human interactions. Great product design, with a focus on user experience, also makes platform use intuitive, thus reducing the need for explicit support.

Incentivized

Anaconda is mainly used by professional data scientists who have profound knowledge of Python coding, mainly used for building some new algorithm block or some optimization, then the module will be integrated into the Dataiku pipeline/workflow. While Dataiku can be used by even other kinds of users.

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

• Customer satisfaction
• Timely project delivery

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