caffe

From the paper: > We used the Caffe implementation (Jia et al., 2014) of the VGG19 deep neural network (DNN) model (Simonyan & Zisserman, 2015; available from https://github.com/BVLC/caffe/wiki/Model - Zoo) . All visual images were resized to 224 × 224 pixels to compute outputs by the VGG19 model . The VGG19 model consisted of a total of sixteen convolutional layers and three fully connected layers. They used a VGG-19 network for the feature detection, so in my basic understanding the results of this could already be improved by switching to a ResNet.

Caffe has a thing called exactly "model zoo"[1], which is also a collection of pre-trained models. The submitted site looks like a borrowed idea (+ borrowed name). [1]: https://github.com/BVLC/caffe/wiki/Model-Zoo

From the abstract: > Sharing the trained DL models has become a trend that is ubiquitous in various fields. Apart from the Model Zoo[1], are there any other sites for sharing serialized models? [1] https://github.com/BVLC/caffe/wiki/Model-Zoo

You actually don't even need the iGPU for simple MNIST training. The CPU is enough. Check out the sample MNIST training example with Tensorflow. If you are asking about using the iGPU for a bit more power, most frameworks don't support them. The only frameworks with which you can use with iGPUs are which have OpenCL support, like caffe[1] (even for those you need to have a recent enough CPU) Tensorflow either uses CPU or CUDA (Not even AMD, but the support is coming for openCL) [1]:https://github.com/BVLC/caffe/tree/opencl

Is bringing the OpenCL caffe backend[1] to Caffe2 on the roadmap? [1] https://github.com/BVLC/caffe/tree/opencl

Yes, that would be OpenCL/SPIR-V. But you still need to put in the work to optimise convolution kernels for specific hardware. TF has openCL support in the works, AFAIK outsourced to Codeplay - https://www.codeplay.com/portal/tensorflow%E2%84%A2-for-open... Caffe has an OpenCL branch, works on AMD but about quarter of the speed of equivalent NVidia hardware using CuDNN 4/5, mostly due to unoptimised kernels - https://github.com/BVLC/caffe/tree/opencl

Nop, check the stats of Caffe for instance: https://github.com/BVLC/caffe This is much older than the 9th of August.

Caffenet also offers a set of pre-trained models in their "model zoo": https://github.com/BVLC/caffe/wiki/Model-Zoo

Pretty sure most frameworks support "one weird trick" style learning. Eg Caffe: https://github.com/BVLC/caffe/blob/master/docs/multigpu.md

the openCL branch of caffe has had a lot of work recently and is usable now. Still a bit slower than with CUDA/CuDNN but works https://github.com/BVLC/caffe/tree/opencl

Caffe has an OpenCL backend - https://github.com/BVLC/caffe/tree/opencl