1. Redundancy exponential
    Setup an infinite loop thru the models.
    Wrote about 1700 poems today.

    Info: bdp.glia.ca/
    Code : github.com/jhave/Wavenet-for-Poem-Generation
    Poems (in txt files) : github.com/jhave/Wavenet-for-Poem-Generation/commit/f457caa2dfe2da59b12d7826f5d689d0906b037f
    FIle: 20170206-2253-59m44s_SILENT

    # vimeo.com/202872109 Uploaded
  2. Neural Nets trained on 11k PoetryFoundation corpus

    Info: bdp.glia.ca
    Code: github.com/jhave/Wavenet-for-Poem-Generation

    Screencapture: 20170124-0052_39-18_fromModel-93542_to_136479
    Duration: 1 hour, 26 minutes, 21 seconds

    # vimeo.com/200910826 Uploaded
  3. Neural Nets trained on 11k PoetryFoundation corpus

    Info: bdp.glia.ca
    Code: github.com/jhave/Wavenet-for-Poem-Generation

    Screencapture: 20170123-2104_38-59from50ktoEnd
    Duration: 1 hour, 28 minutes, 46 seconds

    # vimeo.com/200764749 Uploaded
  4. 1h53m silent realtime screen-capture of approx. 240 neural net models generating poems (x 2).
    __________________________
    Poems are generated side-by-side to allow for comparative viewing.
    __________________________
    Models were trained on an 11k poem corpus for 6 weeks from 26-10-2016 to 11-12-2016 in Hong Kong at the School of Creative Media (on a G5 using Wavenet-for-Poem-Generation).
    __________________________
    Poems were generated on 14-01-2017 in Montreal on a 2012 Macbook Air.
    __________________________
    Suggested viewing: put it on your extra monitor and let it run in background. Consult occasionally. Note the neural-net suffered a minor stroke around the 130K epoch; poems after that display an increasing disintegration of capacity.
    __________________________
    bdp.glia.ca/3-survivors-1397-models-16548-txt-files-wavenet-for-poem-generation-secondary-results-after-training-for-6-weeks-continuously/
    __________________________
    39-18 (2016-10-26T18-39-18)
    __________________________
    Dilation Channels : 8
    __________________________
    Born: 26 October 2016 at 03:29
    Died: Sunday, 11 December 2016 at 11:28
    Models: 458
    Exploded: approx. 130k
    Epochs: 145070
    Size: 80.37GB
    __________________________
    Video File: 20170113-2252-DOUBLE-Laptop-39-18from50k-150kend-1h53m.mov

    # vimeo.com/199601832 Uploaded
  5. 2h28m silent realtime screen-capture of 473 neural net models generating poems (473 x 2 = 946 poems).
    __________________________
    Poems are generated side-by-side to allow for comparative viewing.
    __________________________
    Models were trained on an 11k poem corpus for 6 weeks from 26-10-2016 to 11-12-2016 in Hong Kong at the School of Creative Media (on a G5 using Wavenet-for-Poem-Generation).
    __________________________
    Poems were generated on 14-01-2017 in Montreal on a 2012 Macbook Air.
    __________________________
    Suggested viewing: put it on your extra monitor and let it run in background. Consult occasionally. Note the neural suffered a minor stroke around the 108K epoch; results after that display an increasing disintegration of capacity.
    __________________________
    bdp.glia.ca/3-survivors-1397-models-16548-txt-files-wavenet-for-poem-generation-secondary-results-after-training-for-6-weeks-continuously/
    __________________________
    26-03 (2016-10-26T15-26-03)
    __________________________
    Dilation Channels : 32
    __________________________
    Born: 26 October 2016 at 03:29
    Died: Sunday, 11 December 2016 at 11:28
    Models: 464
    Exploded: 108+
    Epochs: 145070
    Size: 98.1GB
    __________________________
    Video File: 20170115-1054-DOUBLE-LaptopALL-26-03-2h28m

    # vimeo.com/199582593 Uploaded

BDP

David (Jhave) Johnston Plus

Browse This Channel

Shout Box

Heads up: the shoutbox will be retiring soon. It’s tired of working, and can’t wait to relax. You can still send a message to the channel owner, though!

Channels are a simple, beautiful way to showcase and watch videos. Browse more Channels.