search

devicehive / devicehive-audio-analysis

Apache License 2.0
196 stars 80 forks source link

repeating results of audioset paper #7

Open yunbin opened 6 years ago

yunbin commented 6 years ago
  1. Is your result close to what published by the audioset paper? How do you define hit rate? Is it accuracy?
  2. Is the feature you generated using vggish model close to their released feature? It seems very different. 3. They have quantized 8-bit feature and you used floating number feature? Is your model trained with their 8-bit feature?
igor-panteleev commented 6 years ago

Hi,

  1. Can you provide link to "audioset paper" i will take closure look.
  2. Model is trained with audioset 8-bit features. But I should mention that youtube-8m lib dequantize features back before feed them to model. So for final script i have modified vggish feature extractor to prevent unnecessary quantize/dequantize process and feed features directly. vggish quantize - here youtube-8m dequantize - here
yunbin commented 6 years ago

Thanks for the info. It turns out the release features don't match the code generated feature, such that the model doesn't work very well with audio data.

On Mon, Apr 2, 2018 at 8:26 AM, Igor Panteleev notifications@github.com wrote:

Hi,

  1. Can you provide link to "audioset paper" i will take closure look.
  2. Model is trained with audioset 8-bit features. But I should mention that youtube-8m lib dequantize features back before feed them to model. So for final script i have modified vggish feature extractor to prevent unnecessary quantize/dequantize process and feed features directly. vggish quantize - here https://github.com/tensorflow/models/blob/d640ab9cb1c06e7da39cca0399f14f8c6a54eb9c/research/audioset/vggish_postprocess.py#L84 youtube-8m dequantize - here https://github.com/igor-panteleev/youtube-8m/blob/master/utils.py#L28

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/devicehive/devicehive-audio-analysis/issues/7#issuecomment-377919131, or mute the thread https://github.com/notifications/unsubscribe-auth/AE9rS9Tr3aFO-YOhxCw4JVPzL0-Y94I9ks5tkhkSgaJpZM4S8HeG .

-- Yunbin