Reviews from AWS Marketplace
0 AWS reviews
-
5 star0
-
4 star0
-
3 star0
-
2 star0
-
1 star0
External reviews
External reviews are not included in the AWS star rating for the product.
Simple way to build complex models
What do you like best about the product?
I like how easy tensor flow makes it to build a simple neural network. You can have a model up and running in minutes, but tensorflow still provides advanced users with the ability to customize models a lot.
What do you dislike about the product?
I wish that the documentation for tensorflow was more detailed. Sometimes I have a hard time finding answers to some questions that I have about tensorflow.
What problems is the product solving and how is that benefiting you?
I am trying to make more accurate predictive models in the credit industry. Tensorflow has helped to improve the accuracy of existing models built using other software by a few percent.
- Leave a Comment |
- Mark review as helpful
Best deep learning software
What do you like best about the product?
Good documentation, large support community
What do you dislike about the product?
learning curve with backgrounds like Java, DotNet and Javascript etc
What problems is the product solving and how is that benefiting you?
Deep learning models to predict certain outcomes
Recommendations to others considering the product:
Ability to build models and train them easily with Tensorflow. Best tool for data scientists
The machine learning swiss army knife
What do you like best about the product?
The ability to easily write code that scales up to multiple CPUs/GPUs. Since version 1.0, there are higher level modules that provide Keras-like functionalities (e.g. layers, metrics). Moreover, it is definitely production ready. The best thing, though, is the Tensorboard tool that can be used to debug the implemented algorithms.
What do you dislike about the product?
Setting up GPU support on Windows can be tricky, depending on how convoluted the environment is. While Tensorboard is a great tool, sometimes its pages get "stuck" and force to reload the tensorboard server again. The learning curve might be steeper than Keras, which hides most of the complexity away.
What problems is the product solving and how is that benefiting you?
TensorFlow it's a nice, low/mid level API to build production-level software that deals with machine learning. It's stable, so building products at scale for production use is not a problem. This is being used as the back-end to train voice recognition models.
Recommendations to others considering the product:
I'd strongly suggest using Keras for prototyping before jumping straight into TensorFlow. Getting started with keras is much simpler and paves the way to implementing the final product with TensorFlow.
Best deep learning open source library out there
What do you like best about the product?
TensorFlow uses python which is conenient.Its numerical compability with NumPy is great. We can store the models and reuse them is very convenient
What do you dislike about the product?
The learning curve is a bit steep.The tutorials on the website can be more explicitly explained.
What problems is the product solving and how is that benefiting you?
I use tensor flow to solve complex classification problems especially delaying with Images,as Image processing and machine/deep learning go hand in hand for classification problems
Recommendations to others considering the product:
Get yore basics of python,numpy and ML strong before using TensorFlow
Amazing
What do you like best about the product?
The easy environment to adapt.
Working online community is awesome...
Thank you guys
Working online community is awesome...
Thank you guys
What do you dislike about the product?
Some dependencies are that easy to work with,
RNNs are still a bit lacking, compared to Theano.
RNNs are still a bit lacking, compared to Theano.
What problems is the product solving and how is that benefiting you?
designing a hardware to help the blind peoples recognize there loved ones
Recommendations to others considering the product:
Theano
Great tool, worth the steep learning curve
What do you like best about the product?
With TensorFlow you can do pretty much anything you want in the broad area of machine learning. But the amazing thing about it is that you can also use this software to deal with math problems outside of machine learning. The use of computation graphs along with TensorBoard makes model visualization very intuitive.
What do you dislike about the product?
The learning curve can be quite steep if, like me, you start with no knowledge of TensorFlow's computational model philosophy. Once you get the hang of things, it can be quite rewarding.
Also, as the software is updated quite frequently, it seems the documentation is not as accurate as it could be, leading to quite a few headaches.
Also, as the software is updated quite frequently, it seems the documentation is not as accurate as it could be, leading to quite a few headaches.
What problems is the product solving and how is that benefiting you?
I am working with time series forecasting models, and TensorFlow allows for pretty fast prototyping while allowing a huge degree of freedom in terms of model features that you can incorporate.
Recommendations to others considering the product:
Be VERY patient. It may look a bit overwhelming, but it is an amazing tool to master and a great philosophy of computation to understand and follow.
Deep learning made easy with TensorFlow
What do you like best about the product?
The ease of building deep learning models and the high level API's. One can build all kind of architectures pertaining to deep learning. One can use it to build models to solve computer vision problems, perform speech analytics, text analytics, seq2seq models. It supports major algorithms such as ConvNets, RNN, LSTMs, Seq2Seq, It also includes some of the pretrained models which can be customized and trained with new datasets. It can also scale to use multi -gpu systems out of the box without much configuration.
What do you dislike about the product?
Some of the benchmark results show it doesn't train fast, I hope the team is working on making it faster. Also it doesn't include other ML models for comparison.
What problems is the product solving and how is that benefiting you?
We are trying to use multi-layer neural network in customer analytics space. We use various models and TensorFlow is easy to implement using high level api.
Personally, I have used it in building forecasting models, sentiment analysis, text analytics, language translation, general adversarial networks. The implementation for all the models were very easy and it provided great benefits in terms of using multi-gpu systems efficiently.
Personally, I have used it in building forecasting models, sentiment analysis, text analytics, language translation, general adversarial networks. The implementation for all the models were very easy and it provided great benefits in terms of using multi-gpu systems efficiently.
Recommendations to others considering the product:
It has good documentation, and there are many online courses to train the resources. There are new features being added regularly and has great flexibility in terms of use common API's and if needed one can make use of computational graphs to build and customize the models as per the need.
A very great tool for engineers to work on Deep Learning.
What do you like best about the product?
There is a very great ecosystem around TensorFlow, such as TensorBoard for visualizing computation graph, TensorFlow Serving for manage model in production, TFSlim for simplification of building neural network, and so on. TensorFlow is also evolving very fast and it is nearly going to get 1.0 release.
What do you dislike about the product?
It is a little hard for new comer to learn. The API is not that friendly for non-research background people.
What problems is the product solving and how is that benefiting you?
We use it to solve computer vision problem as a proof of concept. It worked fine.
Recommendations to others considering the product:
People who want to use TensorFlow should have enough machine learning background, especially deep learning. To use it in a GPU cluster is not that easy. Learning to use its API also cost a lot of time. But to my extent, TensorFlow is really the best in engineering support among all the deep learning framework. Moreover, people could use Keras on TensorFlow to build network more easily.
showing 61 - 68