Tensorflow 1.5.0
github.com
Tensorflow 1.5.0
1–10 of 30 posts
Re: Tensorflow 1.5.0
#2The big feature is CUDA 9 and cuDNN 7 support, which promises double-speed training on Volta GPUs/FP16. (it should be noted that TF 1.5 does not support CUDA 9.1 yet, which I found out the hard way)
I updated my Keras container with the TF 1.5 RC, CUDA 9, and cuDNN 7 (https://github.com/minimaxir/keras-cntk-docker), but did not notice a significant speed increase on a K80 GPU (I'm unsure if Keras makes use of FP16 yet either).
Re: Tensorflow 1.5.0
#3The big feature is CUDA 9 and cuDNN 7 support, which promises double-speed training on Volta GPUs/FP16. (it should be noted that TF 1.5 does not support CUDA 9.1 yet, which I found out the hard way) I updated my Keras container with the TF 1.5 RC, CUDA 9, and cuDNN 7 ( https://github.com/minimaxir/keras-cntk-docker ), but did not notice a significant speed increase on a K80 GPU (I'm unsure if Keras makes use of FP16…
Speed in FP16 under k80 is almost as FP32, the architecture doesn't work well with FP16. As you said, you need Volta to notice improvements with FP16.
The other two main features are: Eager execution and TensorFlow Lite
Re: Tensorflow 1.5.0
#4The big feature is CUDA 9 and cuDNN 7 support, which promises double-speed training on Volta GPUs/FP16. (it should be noted that TF 1.5 does not support CUDA 9.1 yet, which I found out the hard way) I updated my Keras container with the TF 1.5 RC, CUDA 9, and cuDNN 7 ( https://github.com/minimaxir/keras-cntk-docker ), but did not notice a significant speed increase on a K80 GPU (I'm unsure if Keras makes use of FP16…
k80 is pretty old now, 1080ti have x5 perf of k80 for one tenth of the price
Re: Tensorflow 1.5.0
#5So is it an easy process to convert a tensorflow program from 1.4 to 1.5? When I tried converting something from 0.9 to 1.0 I couldn't figure it out.
Re: Tensorflow 1.5.0
#6So is it an easy process to convert a tensorflow program from 1.4 to 1.5? When I tried converting something from 0.9 to 1.0 I couldn't figure it out.
For most cases it should just be a drop in replacement. IIRC they promise not to break the API between point releases (except tf.contrib.* which may change or disappear entirely...)
Re: Tensorflow 1.5.0
#7The big feature is CUDA 9 and cuDNN 7 support, which promises double-speed training on Volta GPUs/FP16. (it should be noted that TF 1.5 does not support CUDA 9.1 yet, which I found out the hard way) I updated my Keras container with the TF 1.5 RC, CUDA 9, and cuDNN 7 ( https://github.com/minimaxir/keras-cntk-docker ), but did not notice a significant speed increase on a K80 GPU (I'm unsure if Keras makes use of FP16…
k80 is pretty old now, 1080ti have x5 perf of k80 for one tenth of the price
When I say K80, I mean running a GPU in the cloud, not running it locally (and as an aside, due to the crypto boom, buying a physical GPU for cheap is more difficult)
Re: Tensorflow 1.5.0
#8Earlier quoted context omitted.
k80 is pretty old now, 1080ti have x5 perf of k80 for one tenth of the price
When I say K80, I mean running a GPU in the cloud, not running it locally (and as an aside, due to the crypto boom, buying a physical GPU for cheap is more difficult)
oh yeah. what i meant was that as k80 s are old, they don't have the fancy features like being fast with fp16, which came later. k80 come from slightly before the deep learning explosion afaik
Re: Tensorflow 1.5.0
#9The big feature is CUDA 9 and cuDNN 7 support, which promises double-speed training on Volta GPUs/FP16. (it should be noted that TF 1.5 does not support CUDA 9.1 yet, which I found out the hard way) I updated my Keras container with the TF 1.5 RC, CUDA 9, and cuDNN 7 ( https://github.com/minimaxir/keras-cntk-docker ), but did not notice a significant speed increase on a K80 GPU (I'm unsure if Keras makes use of FP16…
You need Pascal hardware or later for FP16. Amazon's G3 Maxwell instances, which are newer than K80s, don't support it, either.
Re: Tensorflow 1.5.0
#10The big feature is CUDA 9 and cuDNN 7 support, which promises double-speed training on Volta GPUs/FP16. (it should be noted that TF 1.5 does not support CUDA 9.1 yet, which I found out the hard way) I updated my Keras container with the TF 1.5 RC, CUDA 9, and cuDNN 7 ( https://github.com/minimaxir/keras-cntk-docker ), but did not notice a significant speed increase on a K80 GPU (I'm unsure if Keras makes use of FP16…
You need Pascal hardware or later for FP16. Amazon's G3 Maxwell instances, which are newer than K80s, don't support it, either.
P3 instances have Volta (V100) GPUs.