feat: Add scaled dot-product self-attention, single head - #9
Open
ThomasHartDev wants to merge 1 commit into
Open
feat: Add scaled dot-product self-attention, single head#9ThomasHartDev wants to merge 1 commit into
ThomasHartDev wants to merge 1 commit into
Conversation
Implement Attention(Q,K,V) = softmax(QK^T / sqrt(d_k)) V with causal masking and a SelfAttentionHead that projects one sequence into Q, K, V.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Adds the Vaswani attention core in plain numpy:
softmax(QK^T / sqrt(d_k)) V, optional boolean masks (False positions go to -inf before softmax), and a single-head self-attention module that projects one sequence into Q, K, and V. Includes a causal lower-triangular mask helper and tests for shapes, scale vs saturation, retrieval, empty/single-token edges, and autoregressive masking.Closes #8.