[GH-ISSUE #37] why <|EOT|> ? #34

Open
opened 2026-10-26 21:01:56 +00:00 by bsakaguchi · 0 comments
Owner

Originally created by @BING-LLL on GitHub (Jul 5, 2024).
Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-MoE/issues/37

In finetune.py, <|EOT|> is added to the end of the output, but this is not a token in the tokenizer.

Originally created by @BING-LLL on GitHub (Jul 5, 2024). Original GitHub issue: https://github.com/deepseek-ai/DeepSeek-MoE/issues/37 In finetune.py, <|EOT|> is added to the end of the output, but this is not a token in the tokenizer.
Sign in to join this conversation.
No labels
pull-request
WIP
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/DeepSeek-MoE#34
No description provided.