Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

some confusions about soft-em #4

Open
yyht opened this issue Apr 13, 2021 · 0 comments
Open

some confusions about soft-em #4

yyht opened this issue Apr 13, 2021 · 0 comments

Comments

@yyht
Copy link

yyht commented Apr 13, 2021

hi, since you waveglow propose to use a soft-em version of vqvae, the core implementation is:
"
def _square_distance(x, code_book):
x = tf.cast(x, tf.float32)
code_book = tf.cast(code_book, tf.float32)
x_sg = tf.stop_gradient(x)
x_norm_sq = tf.reduce_sum(tf.square(x_sg), axis=-1, keepdims=True) # [b, 1]
code_book_norm_sq = tf.reduce_sum(tf.square(code_book), axis=-1, keepdims=True) # [V, 1]
scalar_prod = tf.matmul(x_sg, code_book, transpose_b=True) # [b, V]
dist_sq = x_norm_sq + tf.transpose(code_book_norm_sq) - 2 * scalar_prod # [b, V]

return tf.cast(dist_sq, x.dtype.base_dtype)

dist_sq = _square_distance(x, code_book)
q = tf.stop_gradient(tf.nn.softmax(-.5 * dist_sq))
discrete = tf.one_hot(tf.argmax(-dist_sq, axis=-1), depth=bottleneck_size, dtype=code_book.dtype.base_dtype)
dense = tf.matmul(discrete, code_book)
dense = dense + x - tf.stop_gradient(x)
def _get_losses(x, x_mask, dense, dist_sq, q):
x = tf.cast(x, tf.float32)
x_mask = tf.cast(x_mask, tf.float32)
dense = tf.cast(dense, tf.float32)
dist_sq = tf.cast(dist_sq, tf.float32)
q = tf.cast(q, tf.float32)
disc_loss = tf.reduce_sum(tf.reduce_sum(tf.square(x - tf.stop_gradient(dense)), -1)*x_mask) / (1e-10+tf.reduce_sum(x_mask))
# # M-step
em_loss = -tf.reduce_sum(tf.reduce_sum(-.5 * dist_sq * q, -1)*x_mask) / (1e-10+tf.reduce_sum(x_mask))
return disc_loss, em_loss
disc_loss, em_loss = _get_losses(x, x_mask, dense, dist_sq, q)
"
, however, the tensor2tensor has a different implementation:
https://github.com/tensorflow/tensor2tensor/blob/master/tensor2tensor/layers/vq_discrete.py

  1. multisample to get mean of soft-alignment
  2. when calculate em-loss, it has a different loss funtion type compare to your "M-step"
    .
    Could you hepl me with it?
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant