|Docker image CPU|
|Docker image GPU|
|Docker image GPU+TORCH|
|Docker image GPU+TENSORRT|
DeepDetect Docker for GPU & CPU
DeepDetect docker images are available for CPU and GPU with a range of supported backends, from Pytorch to Caffe, TensorRT, NCNN, Tensorflow, …
Our Docker images are organized as follows:
- All Docker images are available from Docker Hub
- Every DeepDetect release generates images that are tagged with the release number
- Nighly images are generated that bear the
DeepDetect Docker Images, Releases & CI
Framing an automation challenge into a Machine Learning problem, training and selecting models, then serving then within applications are enough challenges for a day.
At Jolibrain we are servicing customers with many heterogeneous requirements and solutions. Automation problems range from time-series forecasting with deep nets to generative models for transforming objects in images, text moderation, code analysis, etc…
Since we are “machine learnists”, we are automating as much as possible everything that is not deep learning. We achieve this in two steps:
- We automate and bring into DeepDetect anything that is useful to us from training to serving, so we never do something twice.
- We have grown our CI/CD capabilities with two objectives:
- We avoid breaking stuff we’ve automated in the past
- We can deliver changes fast and update our customer installs
This is how many capabilities have come to be embedded into DeepDetect, from state-of-the-art time-series to object detection, OCR, CSV and sparse data handling.
Machine Learning functionalities per library
|OCR / Seq2Seq||Y||N||N||N||Y||N||N||N||N|
Releases & CI
DeepDetect releases come every few weeks or so. Our CI is not public but the results are on every PR.
Current release is shown below:
|Latest Release||Release Date||Commits since last release|