Allow configurable sampling steps #318
Merged
Conversation
This file contains 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
The dimensions of the mock decoder output and the test input token ids did not match and were causing PyTorch to issue a warning/error in CI.
rlouf
reviewed
Oct 12, 2023
|
|
||
| """ | ||
| probs = torch.nn.functional.softmax(logits, dim=-1) | ||
| # next_token_ids = torch.multinomial(probs, num_samples=samples, generator=rng) |
There was a problem hiding this comment.
torch.multinomial accepts matrices for probs: https://pytorch.org/docs/stable/generated/torch.multinomial.html
There was a problem hiding this comment.
I have a few follow-up updates to add shortly; I'll put that in place alongside those.
There was a problem hiding this comment.
While torch.multinomial does work for matrices, we still need to check its results against the higher dimensional tensor inputs in the tests (and update the tests that rely on RNG seeds), so the change should come in a follow-up.
rlouf
reviewed
Oct 12, 2023
|
👍 :) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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 allows one to specify sampler functions that decide exactly how generated token sequences are sampled/produced from the logit values. In doing so, this PR also adds a greedy "sampler" function (i.e. one that chooses the max logit values).