What I like best about Tensorflow is its flexibility and power. It's like a swiss army knife for machine learning and deep learning. You can build anything from simple models to complex neural networks for computer vision, NLP and more. The pre-built models and tools for transfer learning make it easier to get started, and the support for deployment across platforms, mobile, web and cloud is super convenient.

Additionally, the community is massive. So many tutorials, open source projects and helpful forums, you will never feel stuck. Once you get the hang of it, the possibilities are endless. Review collected by and hosted on G2.com.

The learning curve, it can feel pretty overwhelming at first, especially for beginners. The syntax can get complex, and debugging isn't always straightforward.

Another thing is it can be heavy and a bit slow compared to some other frameworks, especially when you are just experimenting or working on smaller projects. Setting up the environment is also a hassle, plus you need to be careful with versions as well. Review collected by and hosted on G2.com.

Graphs of tensorflow:TensorFlow has better computational graph visualizations. Which are better when compared to other libraries like PyTorch and Theano.

Scalability : It can run on CPUs, GPUs for high-performance computations.

Community :Tesnorflow has a very good community support ,so when we are stuck in a problem we can always ask our questions to the community for solutions which we get most of the time.

Easy to start : Also the extensive documentation of tensorflow helps to start up with it ease for an active development.Also because of python it is easy to implement. Review collected by and hosted on G2.com.

Slow : TensorFlow lacks in both speed and usage when it is compared to its competitors.

Limited GPU Support:Only NVIDIA and Python are supported by TensorFlow for GPU programming. It has no additional support of ther GPUs Review collected by and hosted on G2.com.

That it’s versatile and helps in various ml related task Review collected by and hosted on G2.com.

It’s not always compatible with the hardware or other libraries used. Sometimes it isn’t compatibility with lambda versions too Review collected by and hosted on G2.com.

It is highly scalable, making it a great choice for both small-scale projects and large distributed systems. It offers TensorFlow Lite (for mobile & edge devices), TensorFlow.js (for web applications), and TensorFlow Extended (TFX) (for production-level deployment) that means a complete suite of requirements can be fulfilled by it. A rich ecosystem of developers and customer support is great. Review collected by and hosted on G2.com.

Debugging issues, understanding execution flow, and handling tensors can be quite problematic for beginners. Compatibility Issues & Frequent API Changes is the worst. Review collected by and hosted on G2.com.

I have used this to develop face detection applications. Review collected by and hosted on G2.com.

At first, I found it quite difficult because the information was not readily accessible to me. Review collected by and hosted on G2.com.

What I like most about TensorFlow is its versatility—whether for deep learning, computer vision, or natural language processing, it offers a comprehensive set of tools that make model development scalable and efficient. Review collected by and hosted on G2.com.

One downside is its steep learning curve. While powerful, TensorFlow can be complex for beginners, and some functions require extensive documentation review or prior experience with machine learning frameworks. Review collected by and hosted on G2.com.

I love TensorFlow as a data scientist I use it on daily basis. I have used it to train and tweak models. I create Production grade ML pipelines. I can get low level control support, I can integrate custom loss functions. There are thousands of examples with which we can use TensorFlow easily with new models. Review collected by and hosted on G2.com.

For small datasets, it is kind of a heavy library, so should not be used here. For learning and quick deployment work as well I would prefer keras. Review collected by and hosted on G2.com.

I liked using TensorFlow due to its end-to-end interface. The data model building using Keras to powerful visualization supported the machine learning pipeline throughout my projects. TensorFlow heads built in tools for optimisation which was a huge plus and saves a lot of time. Review collected by and hosted on G2.com.

I have felt issues with it sometimes because for embedded applications it can be quite heavy and complicated especially while converting some models to lite version with unsupported operations. Resolving or debugging such issues often need deep research or asking forums and trial and error methods. Review collected by and hosted on G2.com.

Best thing about TensorFlow is that it supports both deep learning and traditional ML models, with deployment on multiple platforms like CPUs, GPUs etc. We have used this package in our python framework to work on phone call sentiment analysis project by leveraging deep learning models like LSTMs and CNNs . It helps us in text preprocessing, word embeddings, and training neural networks for sentiment classification. It is easy to integrate with Google cloud AI. Review collected by and hosted on G2.com.

It has very steep learning curve because of complex APIs available and its computational graphs make debugging less intuitive. Hence, we are leveraging pytorch as well for our sentiment analysis project because of pyhtonic syntax and less programming involved. Review collected by and hosted on G2.com.

I have used the library for my machine learning applications and it helped to speed up my programming tasks Review collected by and hosted on G2.com.

it doesn't have a proper reinforcement learning library Review collected by and hosted on G2.com.