https://github.com/triagemd/tensorflow-builds
Tensorflow binaries and Docker images compiled with GPU support and CPU optimizations.
https://github.com/triagemd/tensorflow-builds
bazel cuda cudnn docker gpu machine-learning nvidia python tensorflow tensorflow-serving
Last synced: 9 months ago
JSON representation
Tensorflow binaries and Docker images compiled with GPU support and CPU optimizations.
- Host: GitHub
- URL: https://github.com/triagemd/tensorflow-builds
- Owner: triagemd
- License: mit
- Archived: true
- Created: 2017-08-20T13:08:33.000Z (over 8 years ago)
- Default Branch: master
- Last Pushed: 2019-07-30T18:39:42.000Z (over 6 years ago)
- Last Synced: 2024-08-04T10:06:47.005Z (over 1 year ago)
- Topics: bazel, cuda, cudnn, docker, gpu, machine-learning, nvidia, python, tensorflow, tensorflow-serving
- Language: Dockerfile
- Homepage:
- Size: 58.6 KB
- Stars: 15
- Watchers: 5
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
**Warning:** This repo is now unmaintained and unnecessary.
Use the official `tensorflow/serving` Docker images instead: https://hub.docker.com/r/tensorflow/serving
---
# tensorflow-builds
Tensorflow binaries and Docker images compiled with GPU support and CPU optimizations.
We now pull optimized Tensorflow wheels from [mind/wheels](https://github.com/mind/wheels) rather than building our own.
## Updating Tensorflow version
When a new version of Tensorflow is released:
1. Update the versions at the top of `script/distribute*`.
2. On a *very powerful* machine with Docker, set `$GITHUB_TOKEN` to [a GitHub access token with repo permissions](https://help.github.com/articles/creating-a-personal-access-token-for-the-command-line/) and run `script/distribute-binaries` (seriously, crank the CPU all the way up if using the cloud).
3. Update the binary links below from https://github.com/triagemd/tensorflow-builds/releases.
4. Push to master.
---
| Docker builds |
|-|
| [](https://travis-ci.org/triagemd/tensorflow-builds) |
| Binary | |
|-|-|
| **Tensorflow Serving 1.7:** |
| CPU optimized, with Intel MKL | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.7.0/tensorflow_model_server_optimized |
| GPU, CUDA 9.0, without CPU optimization | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.7.0/tensorflow_model_server_gpu |
| GPU, CUDA 9.1, with CPU optimization, with Intel MKL | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.7.0/tensorflow_model_server_optimized_gpu |
| **Tensorflow Serving 1.6:** |
| CPU optimized, with Intel MKL | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.6.0/tensorflow_model_server_optimized |
| GPU, CUDA 9.0, without CPU optimization | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.6.0/tensorflow_model_server_gpu |
| GPU, CUDA 9.1, with CPU optimization, with Intel MKL | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.6.0/tensorflow_model_server_optimized_gpu |
| **Tensorflow Serving 1.5:** |
| CPU optimized, with Intel MKL | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.5.0/tensorflow_model_server_optimized |
| GPU, CUDA 9.0, without CPU optimization | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.5.0/tensorflow_model_server_gpu |
| GPU, CUDA 9.1, with CPU optimization, with Intel MKL | https://github.com/triagemd/tensorflow-builds/releases/download/tf-serving-1.5.0/tensorflow_model_server_optimized_gpu |
| Docker image | |
|-|-|
| **Python + Intel MKL-DNN:** |
| python2.7-mkl | [](https://hub.docker.com/r/triage/python2.7-mkl/) |
| python3.6-mkl | [](https://hub.docker.com/r/triage/python3.6-mkl/) |
| **Python + CUDA:** |
| python2.7-cuda9.0 | [](https://hub.docker.com/r/triage/python2.7-cuda9.0/) |
| python3.6-cuda9.0 | [](https://hub.docker.com/r/triage/python3.6-cuda9.0/) |
| python2.7-cuda9.1 | [](https://hub.docker.com/r/triage/python2.7-cuda9.1/) |
| python3.6-cuda9.1 | [](https://hub.docker.com/r/triage/python3.6-cuda9.1/) |
| **Python + CUDA + Intel MKL-DNN:** |
| python2.7-cuda9.1-mkl | [](https://hub.docker.com/r/triage/python2.7-cuda9.1-mkl/) |
| python3.6-cuda9.1-mkl | [](https://hub.docker.com/r/triage/python3.6-cuda9.1-mkl/) |
| **Tensorflow:** |
| python2.7-tensorflow | [](https://hub.docker.com/r/triage/python2.7-tensorflow/) |
| python3.6-tensorflow | [](https://hub.docker.com/r/triage/python3.6-tensorflow/) |
| python2.7-tensorflow-gpu | [](https://hub.docker.com/r/triage/python2.7-tensorflow-gpu/) |
| python3.6-tensorflow-gpu | [](https://hub.docker.com/r/triage/python3.6-tensorflow-gpu/) |
| **Tensorflow with CPU optimizations:** |
| python2.7-tensorflow-optimized | [](https://hub.docker.com/r/triage/python2.7-tensorflow-optimized/) |
| python3.6-tensorflow-optimized | [](https://hub.docker.com/r/triage/python3.6-tensorflow-optimized/) |
| python2.7-tensorflow-optimized-gpu | [](https://hub.docker.com/r/triage/python2.7-tensorflow-optimized-gpu/) |
| python3.6-tensorflow-optimized-gpu | [](https://hub.docker.com/r/triage/python3.6-tensorflow-optimized-gpu/) |
| **Tensorflow Serving:** |
| python2.7-tensorflow-serving | [](https://hub.docker.com/r/triage/python2.7-tensorflow-serving/) |
| python2.7-tensorflow-serving-optimized | [](https://hub.docker.com/r/triage/python2.7-tensorflow-serving-optimized/) |
| python2.7-tensorflow-serving-gpu | [](https://hub.docker.com/r/triage/python2.7-tensorflow-serving-gpu/) |
| python2.7-tensorflow-serving-optimized-gpu | [](https://hub.docker.com/r/triage/python2.7-tensorflow-serving-optimized-gpu/) |