extensive ecosystem of integrations specially with technologies like Keras and Numpy that help immensely in model development. Easy deployment and model training is also a good to have feature for various platforms. Review collected by and hosted on G2.com.

steeper learning curve for a beginner like me, not well documented (you will find the majority of stuff needed in docs but not those intricacies that some integrations required), for that I had to refer to stack overflow. Review collected by and hosted on G2.com.

TensorFlow is an best platform for machine learning and deep learning models to train and process without taking care of another stuff and it also give us variants of models to use. Review collected by and hosted on G2.com.

Right now TensorFlow uses lot's of memory and computing power which is needed but at any how if it would be reduced it will be best for the models to train. Review collected by and hosted on G2.com.

I love how flexible and powerful it is—being a Google-backed platform means it’s packed with top-tier capabilities. For coders just diving into AI, it’s a fantastic place to start. Review collected by and hosted on G2.com.

The learning curve is steep. It took me a while to get comfortable with development and fully grasp all of TensorFlow’s features. But once you push through the initial hurdles, the possibilities are seriously impressive. Review collected by and hosted on G2.com.

Since tensorflow is from google so, its has a strong background.

It has a very large community for support

I can always refer to their community forums and blogs for help and support.

It has multi-language support other than Python, C++, Java.

It is very easy to use thanks to guides provided by tensorflow community. Review collected by and hosted on G2.com.

Debugging is a very hectic task in tensorflow.

It also is very resource intensive: Since it requires significant hardware resources (GPU/TPU) for training large models. Review collected by and hosted on G2.com.

TensorFlow is mainly designed for Python, but it also enables us to run ML models on Node.js environments, whether in a browser or any other runtime. That's why, as a JavaScript developer, I find it friendly to integrate any machine learning models directly into the web. Review collected by and hosted on G2.com.

All the features of TensorFlow are not available in the JavaScript version so sometimes we need to depend on Python to train the models before converting them to JS. That's why, I would recommend having the strong fundamentals of Python along with learning this awesome library is a must. Review collected by and hosted on G2.com.

Good for prototyping and rapid model development. Review collected by and hosted on G2.com.

Old versions are not compatible with new versions and there are lot of compatibility issues. Review collected by and hosted on G2.com.

Its is very good for quickly building and testing ml models.It is open source. Review collected by and hosted on G2.com.

well for dislike there was a time in project in which due to its technical limitation i was unable to use it and i had to switch to pytorch for that. Review collected by and hosted on G2.com.

At this point, TensorFlow has become a gigantic collection of high-level API's to do any sort of data manipulation and analytics work. It provides any functionality you need to quickly run neural networks on your data while also giving access to lower level features for custom processing. Review collected by and hosted on G2.com.

Due to its vastness, it is very intimidating even for intermediate professionals. It requires careful study and focus on the task at hand to use it effectively. Review collected by and hosted on G2.com.

One of the best things about TensorFlow is its flexibility and scalability. It allows me to design and deploy machine learning models easily, whether I am implementing small scale projects or integrating large scale projects and production ready systems. TensorFlow is frequently used in these type of projects. Review collected by and hosted on G2.com.

Debugging TensorFlow errors is quite difficult because stack traces are not good sometimes. Issues with tensors, shapes, or mismatched data types sometimes require deep dives into error messages that is not very helpful. Review collected by and hosted on G2.com.

TensorFlow is extremely versatile in terms of building Machine Learning models. It also helps you visualize and give insights on how a particular model is performing. It is also very easy to integrate it with Google Cloud, making model training and deployment quite simple. Review collected by and hosted on G2.com.

Although TensorFlow has a quite steep learning curve, you can grow on it with time. But certain aspects like debugging and working with APIs still feel quite complex. Although the performance is great, GPU optimization is something I haven't been able to get hold of. Review collected by and hosted on G2.com.