dispatches the best kernel implementation based on CPU vector
        architecture. To disable them, build with
        `--define=tensorflow_mkldnn_contraction_kernel=0`.
    *   `tf.linspace(start, stop, num)` now always uses "stop" as last value
        (for num > 1)
    *   Added top-k to precision and recall to keras metrics.
    *   Add a ragged size op and register it to the op dispatcher

- github.com/gliderlabs/ssh: [v0.1.1](https://github.com/gliderlabs/ssh/tree/v0.1.1)
- github.com/go-critic/go-critic: [1df3008](https://github.com/go-critic/go-critic/tree/1df3008)
- github.com/go-lintpack/lintpack: [v0.5.2](https://github.com/go-lintpack/lintpack/tree/v0.5.2)
- github.com/go-ole/go-ole: [v1.2.1](https://github.com/go-ole/go-ole/tree/v1.2.1)
- github.com/go-toolsmith/astcast: [v1.0.0](https://github.com/go-toolsmith/astcast/tree/v1.0.0)