Release: 0.9-soft
This release reflects the state of Fathom more or less as it was for the paper published paper in September 2016. We are currently developing a somewhat more user-friendly version, which you can track in the GitHub issue tracker. If you're eager to use Fathom as it is, please let us know. We might be able to help you get started.
This paper contains a description of the workloads, performance characteristics, and the rationale behind the project:
R. Adolf, S. Rama, B. Reagen, G.Y. Wei, D. Brooks. "Fathom: Reference Workloads for Modern Deep Learning Methods." (Arxiv) (DOI)
| Name | Description |
|---|---|
| Seq2Seq | Direct language-to-language sentence translation. State-of-the-art accuracy with a simple, language-agnostic architecture. |
| MemNet | Facebook's memory-oriented neural system. One of two novel architectures which explore a topology beyond feed-forward lattices of neurons. |
| Speech | Baidu's speech recognition engine. Proved purely deep-learned networks can beat hand-tuned systems. |
| Autoenc | Variational autoencoder. An efficient, generative model for feature learning. |
| Residual | Image classifier from Microsoft Research Asia. Dramatically increased the practical depth of convolutional networks. ILSVRC 2015 winner. |
| VGG | Image classifier demonstrating the power of small convolutional filters. ILSVRC 2014 winner. |
| AlexNet | Image classifier. Watershed for deep learning by beating hand-tuned image systems at ILSVRC 2012. |
| DeepQ | Atari-playing neural network from DeepMind. Achieves superhuman performance on majority of Atari2600 games, without any preconceptions. |
We've begun to put together some actual documentation, and it's already better than what we had:
Please bear with us as the docs get fleshed out.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent