Update to fix bug for csv files with real commas in text#214
Open
ianrowan wants to merge 1 commit into
Open
Conversation
Currently the load_dataset method will load the following csv row example: "I ate pie, ice cream, and cookies\n" in line 37 into reader list as reader=[["I ate pie", "ice cream", "and cookies"]] due to the commas read as columns delimiters. When reader[0] is iterated as row in line 39 only row[0](I ate pie) is taken which would yield "<|startoftext|>I ate pie"<|endoftext|>" This is because the commas created multiple list items in row. Proposed change is to join each entire row list to use full text from each csv row. in this case ''.join(row) yields "<|startoftext|>I ate pie ice cream and cookies"<|endoftext|>" as it compresses the list into a single string. I have tested both of these scenarios in the code base as well.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Currently the load_dataset method will load the following csv row example: "I ate pie, ice cream, and cookies\n" in line 37 into reader list as
reader=[["Iate pie", "ice cream", "and cookies"]]` due to the commas read as columns delimiters.When
reader[0]is iterated as row in line 39 onlyrow[0](I ate pie) is taken which would yield"<|startoftext|>I ate pie"<|endoftext|>"This is because the commas created multiple list items in row.
Proposed change is to join each entire row list to use full text from each csv row. in this case
''.join(row)yields"<|startoftext|>I ate pie ice cream and cookies"<|endoftext|>"as it compresses the list into a single string. I have tested both of these scenarios in the code base as well.