🚀 FastSpeech2 + Linformer: Smaller, Faster and Stronger.#66
Merged
dathudeptrai merged 2 commits intomasterfrom Jul 8, 2020
Merged
🚀 FastSpeech2 + Linformer: Smaller, Faster and Stronger.#66dathudeptrai merged 2 commits intomasterfrom
dathudeptrai merged 2 commits intomasterfrom
Conversation
dathudeptrai
added
enhancement 🚀
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR is an combination of FastSpeech2 and recent paper Linformer: Self-Attention with Linear Complexity. Linformer paper state that:
That mean attention_head_size don't need equal to hidden_size // n_head. Here i set attention_head_size = 8 rather than 384/2 = 192 as v1. The model is smaller and faster 25% while the quality and performance is the same, even better. Especially, when sequence length is large we can gain more :)). Paper also state that
k here == attention_head_size * n_head and n is sequence length
I will let you try and see.