ORPO (#1419)

* orpo trainer

* rl handling for orpo

* support for remove_unused_columns

* orpo fixes

* fix loader for orpo

* chore: lint

* fix default for remove_unused_columns

* roll ORPO into the main AxolotlTrainer so it can be compatible with some of the other techniques like relora

* better handling of system message for orpo

* revert system prompt changes for chat templtes

* no need for else condition

* split dataset parsing into it's own component

docs/rlhf.md

@@ -34,6 +34,21 @@ datasets:
 rl: ipo
 ```
 
+#### ORPO
+
+Paper: https://arxiv.org/abs/2403.07691
+
+```yaml
+rl: orpo
+orpo_alpha: 0.1
+remove_unused_columns: false
+
+chat_template: chatml
+datasets:
+  - path: argilla/ultrafeedback-binarized-preferences-cleaned
+    type: orpo.chat_template
+```
+
 #### Using local dataset files
 ```yaml
 datasets: