paired kto support (#1069)

2024-01-09 13:30:45 -05:00
parent 768d348f42
commit d7057ccd36
3 changed files with 6 additions and 1 deletions
--- a/README.md
+++ b/README.md
@@ -595,6 +595,9 @@ datasets:
      # For `completion` datsets only, uses the provided field instead of `text` column
      field:

+# use RL training: dpo, ipo, kto_pair
+rl:
+
 # Saves the desired chat template to the tokenizer_config.json for easier inferencing
 # Currently supports chatml and inst (mistral/mixtral)
 chat_template: chatml