Parameter input rotary-freq #263
Open
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
…ding frequency parameter
|
To test my changes I ran training with a small model and put a breakpoint where the parameter is used in rotary.py. I set a different value both in command line or in the json file and it was correctly passed on. Training was successful and tests passed. (except one in grad_accum and one in tokenize but not due to this PR, just access right issues).
|
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Outdated suggestions cannot be applied.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 allows to change the rotary positional embedding frequency parameter.
This is useful given the more recent approaches:
LLaMA 1&2 used 10000 which is the default value here.
LLaMA 3 uses 500000, Mistral uses 1000000
LLaMA long context extension works usually increase from 10000 to 100000.