The major deep learning frameworks today are the following:
| Framework | Website |
|---|---|
| TensorFlow | www.tensorflow.org |
| Keras | keras.io |
| CNTK | microsoft.com/cognitive-toolkit |
| MXNet | mxnet.apache.org |
| Caffe2 | caffe2.ai |
| Torch | torch.ch |
| PyTorch | pytorch.org |
| Deeplearning4j | deeplearning4j.org |
| Intel Nervana | neon.nervanasys.com |
All can be implemented in production on distributed systems, with Deeplearning4j being the only one written natively in java. Keras is a user-friendly frontend to TensorFlow and CNTK. TensorFlow is by far the most popular – one measure of adoption is the number of stars the framework has on github. As of today, those numbers are:
| Framework | Jul 17 | Jun 18 | Mar 20 |
|---|---|---|---|
| TensorFlow | 62,199 | 103,087 | 141,493 |
| Keras | 17,086 | 30,726 | 47,017 |
| CNTK | 11,593 | 14,633 | 16,694 |
| MXnet | 10,217 | 14,218 | 18,423 |
| Caffe2 | 5,076 | 24,544 | archived |
| Torch | 7,038 | 7,739 | 8,476 |
| PyTorch | 5,647 | 16,494 | 36,485 |
| DL4j | 6,850 | 9,149 | 11,511 |
| Intel Nervana | 3,093 | 3,569 | 3,828 |
Table 1 – Framework vs stars on github
As we can see from the table and from Figure 1 below, TensorFlow is by far the most popular framework. Growth is particularly strong in TensorFlow, Keras, Caffe2 and PyTorch.
Figure 1 - The Dramatic Rise of TensorFlow
Putting aside the astounding rise of TensorFlow, all of these frameworks are being used extensively within the companies from which they are primarily being developed, e.g.,
| Company | Framework |
|---|---|
| Amazon | MXNet |
| Microsoft | CNTK |
| TensorFlow | |
| PyTorch | |
| Intel | Neon |
Keras is a frontend and DL4J is used in several enterprises, so they are all getting traction in the Enterprise. The DL4J commercial arm is here https://skymind.ai. Validation that these frameworks are being used in production comes from the fact that they are being used in these hyperscale companies, being applied at scale to some of the largest data sets in the world.
References:
Benchmarking state-of-the-art deep learning software tools, Feb 17, 2017
The Top 10 deep learning frameworks, Packt, May 25, 2017
Google’s latest platform play is artificial intelligence, and it’s already winning, The Verge May 18, 2017