I like TensorFlow because it makes it easier to build, train, and test machine learning models. I also like it because it has good documentation and support, which helps a lot when I’m learning or troubleshooting. Review collected by and hosted on G2.com.

I dislike about the tensor flow is it bit complicated at first and some errors are difficult to understand and setting up the model takes a lot of time Review collected by and hosted on G2.com.

I appreciate TensorFlow for its scalability and flexibility, which make it well suited for both small and large machine learning projects. I also value the robust performance it delivers, especially when working with deep learning models. The Keras API is a particular favorite because it supports rapid model development and noticeably boosts my productivity. I find TensorBoard invaluable for visualization and debugging, since it provides clear, detailed insight into the training process. The deployment ecosystem, including TensorFlow Lite, TensorFlow.js, and TensorFlow Serving, is another major strength, enabling efficient deployment across a range of platforms. I also like how straightforward the initial setup is through Python’s package installer, which makes it accessible and easy to start using. Overall, TensorFlow’s integration with a variety of other tools significantly improves my machine learning workflow. Review collected by and hosted on G2.com.

I find TensorFlow’s limitations on Windows to be a significant drawback. Compared with Linux, the Windows version doesn’t offer the same full feature set, which can affect performance and, at times, make GPU support more complicated. Overall, these constraints can get in the way of the experience and reduce TensorFlow’s usability for Windows users. Review collected by and hosted on G2.com.

I like the strong community sense, the fact that is production ready not just one of the so many gitlab repos out there Review collected by and hosted on G2.com.

TensorFlow can be a bit "verbose" at times, but I guess that is good for some Review collected by and hosted on G2.com.

I appreciate TensorFlow for its scalability and flexibility, especially through high-level APIs like Keras, which simplify complex processes and make building and training deep neural networks more manageable. The comprehensive ecosystem of tools and libraries it offers is invaluable, helping to abstract much of the underlying complexity typically involved in such tasks. Additionally, I find the community support around TensorFlow incredibly beneficial, providing a steady stream of updates, resources, and shared knowledge that enhance the overall usability of the platform. I also enjoy how easy the initial setup was by simply following the provided instructions. The integration of external programming tools with TensorFlow through APIs and specialized libraries contributes significantly to my workflow by managing tasks like visualization, model analysis, and deployment. Furthermore, transitioning to TensorFlow from PyTorch has been advantageous due to the appealing libraries such as Keras and TensorFlow Extended, which offer more varieties and functionalities that align with my needs. Review collected by and hosted on G2.com.

I find TensorFlow's C++ documentation limited. This lack of depth impacts my ability to fully leverage its capabilities and integrate them into complex systems. I believe the documentation could be improved by including more practical examples, better API reference details, clearer explanations of complex features like XLA, and guidance on build systems and common use cases. Review collected by and hosted on G2.com.

I appreciate TensorFlow for its scalability and flexibility, which makes it adept at handling both small and large-scale machine learning projects. I love the robust performance it offers, which is essential for deep learning models. The Keras API is a particular favorite of mine because it allows for rapid model development, enhancing my productivity significantly. I find TensorBoard invaluable for visualization and debugging, offering deep insights into model training processes. The deployment ecosystem that includes TensorFlow Lite, TensorFlow.js, and TensorFlow Serving is fantastic, allowing efficient model deployment across various platforms. I also appreciate the straightforward initial setup process using Python's package installer, making it accessible and easy to get started. The integration of TensorFlow with a variety of other tools enhances my machine learning workflow considerably. Review collected by and hosted on G2.com.

I find TensorFlow's limitations on Windows to be a significant drawback. The Windows version lacks the full feature set available on Linux, which affects performance and sometimes complicates GPU support. These constraints can hinder the overall experience and usability of TensorFlow for Windows users. Review collected by and hosted on G2.com.

I have been using tensorFlow for past 2 months as I have ML in my project ..previously i was using SciKit learn and then my friend recommended me the Tensorflow it was very efficient for doing all the complex neural network things which i am not able to do using SciKit and Keras also is integrated with it makes it more convenient to use for my projects. Review collected by and hosted on G2.com.

The tensorFlow was really efficient but my initial experience was not good enough .It took me lot of time to configure the system with it and the second most important problem which i faced was during debugging like if an error occurs then it takes a lot of time to understand the error and work on it ..And if i make a small change in the code then the whole model collapse making it more stressful and frustrating. Review collected by and hosted on G2.com.

Tensorflow with its documentation gives a very easy implementation. Its various models help ease of integration in both web and mobile platforms and it has a great customer support and community and I use it frequently with all my machine learning projects. Review collected by and hosted on G2.com.

The learn curve is pretty steep and especially working with high level Keras. Review collected by and hosted on G2.com.

Video related built in functions are a great addition Review collected by and hosted on G2.com.

Still computing power issue pertains, and the requirement of hardware Review collected by and hosted on G2.com.

I Like There Are wide range of features, and good community support and on stackoverflow support by dev also compatibility with both research and production environments make TensorFlow Extra Ordinary In My Opinions , Its is for both beginners and advanced users is a huge plus. most of CS student are used in their daily projects and easy to use by student and professional and easy to integration using python rich support and easy to implement in python files. Review collected by and hosted on G2.com.

It's hard for new users to learn at beginner stage and the instructions sets, even though there are a lot of things to learn like probability and statistic concepts to use efficiently, it can feel like too much. Fixing problems and debugging can also be tough for devs because the error messages are hard to understand and interpret but chat gpt can solve a lot of things for devs. Review collected by and hosted on G2.com.

I love how powerful and flexible TensorFlow is for building and training deep learning models. Keras makes it a bit easier and pre-trained models save a lot of time. Plus the community is great when I get stuck. Review collected by and hosted on G2.com.

The learning curve is steep. Especially for beginners. Sometimes the error messages are too complicated to understand and debugging is frustrating. Also it requires a lot of computing power which can be a problem if you don’t have high end hardware. Review collected by and hosted on G2.com.