add dpo training scripts

tyxiong23 · tyxiong23 · commit c2b022c7d600 · 2024-09-15T16:01:44.000-04:00
diff --git a/docs/LLaVA_OneVision_Chat.md b/docs/LLaVA_OneVision_Chat.md
@@ -108,7 +108,7 @@ This iterative process is repeated for `N=3` rounds in total, with each round re
     "chosen": "<chosen model response>",
     "rejected": "<rejected model response>"
   },
-  // ...
+  ...
 ]
 ~~~
 

Original file line number	Diff line number	Diff line change
@@ -108,7 +108,7 @@ This iterative process is repeated for `N=3` rounds in total, with each round re
`108`	`108`	`"chosen": "<chosen model response>",`
`109`	`109`	`"rejected": "<rejected model response>"`
`110`	`110`	`},`
`111`		`- // ...`
	`111`	`+ ...`
`112`	`112`	`]`
`113`	`113`	`~~~`
`114`	`114`