Feat/custom chat template #1233
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
winglian
reviewed
Jan 31, 2024
| @@ -615,6 +615,8 @@ rl: | |||
| chat_template: chatml | |||
| # Changes the default system message | |||
| default_system_message: You are a helpful assistant. Please give a long and detailed answer. # Currently only supports chatml. | |||
| # A custom chat template in Jinja format. Will overwrite `chat_template` and `default_system_message` | |||
There was a problem hiding this comment.
"May overwrite the default system message" since you could design the chat template to accept one.
winglian
reviewed
Jan 31, 2024
| @@ -218,7 +218,9 @@ def load_tokenizer(cfg): | |||
| LOG.debug(f"PAD: {tokenizer.pad_token_id} / {tokenizer.pad_token}") | |||
| LOG.debug(f"UNK: {tokenizer.unk_token_id} / {tokenizer.unk_token}") | |||
|
|
|||
| if cfg.chat_template: | |||
| if cfg.custom_chat_template: | |||
| tokenizer.chat_template = cfg.custom_chat_template | |||
There was a problem hiding this comment.
similar to the chatml replacement, we could consider some sort of standardized string in the chat template that could be replaced based on the configuration.
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Outdated suggestions cannot be applied.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.
Description
Adds a
custom_chat_templateparameter on the config that allows to set a custom chat template in Jinja format.Motivation and Context
It's convenient so those of use who use this won't need to edit the tokenizer config directly after the fine-tune has finished.
How has this been tested?
I did not write any test cases for this. I tested it manually on my own environment and it does seem to work.