Bitfusion Ubuntu 14 TensorFlow
Bitfusion.io | 2017.04Linux/Unix, Ubuntu 14.04 - 64-bit Amazon Machine Image (AMI)
tensorflow serving is not compiled
don't claim to have tensorflow serving "pre-installed" if you only have the repo downloaded. the whole benefit of using custom amis is to not have to compile everything myself. waste of time.
- Leave a Comment |
- Mark review as helpful
tensorflow_model_server not installed
We needed a server to serve our models, but tensorflow_model_server is not installed.
There is a folder "serving" where everything seems to be copied, but nothing is executable.
not working on G2 & P2 with latest aws updates
Peer access is not working on GPUs. Tensorflow backend is not able to distribute computing on multiple GPUs . Possibly requires updated cuda libraries.
Not working for Other type of instances.
It is not working for G2 and P2 instances. I tried with t2Medium. it works fine with only tthat instance type.
Life is so much easier with Bitfusion
Before trying Bitfusion I was spending so much time to configure my remote server for installing machine learning libraries and python modules. I am pleasantly surprised to see that every pre-installation and configuration for my deep learning codes is already done in Bitfusion Tensorflow, with a reasonable price.
Works out of the box as advertised.
Tested working ML application running python, CUDA, and tensorflow. Minimal setup was needed to get my code running. Works great with a p2 instance.
Easy to use and out of the box solution
I spent less time on setting up an ami and more time on the model. Easy to use and out of the box solution to using Tensorflow on GPU :)
Works really well - saves a lot of hassle
Tried putting tensorflow + gpu support all by myself on Ubuntu 12.04. Turned out to be a mess. This AMI works seamlessly and saved a lot of time. Proved to be a really useful because I was towards the fag end of my semester and was struggling with the dependency installs.
Productivity multiplier for my DeepLearning project
After wasting long hours on compiling TensorFlow on GPU instance, I switched to using BitFusion AMI on a trial. This saved me lot of trouble, and also saved me the cost of keeping a once compiled machine for long time. I can launch a new instance with working TensorFlow out-of-the-box. This has helped me immensely for Udacity Machine Learning NanoDegree.
Works like a charm!
I used Bitfusion's g2.2x Ubuntu 14 TensorFlow image and it worked great out of the box... I also found the trial period very useful, since I had a problem similar to cifar10 and during the trial period I was able to make sure that this image pulls off cifar10 easily. All I had to run was a single line of command and it started training a cifar10 model, using the GPU. I was able to witness 8x speedup when compared to my weak CPU. Yesterday I came back for my actual problem with my dataset size being nearly 8 times greater than that of cifar10. With some minor changes to cifar10 codebase, I was able to train a model for my problem successfully. I think I can wait for some more time before buying myself a physical GPU. Infrastructure-as-Service rocks!