GPT prompt/target pairs (pairs-per-target=2) from clustered speakers

.gitattributes

@@ -59,3 +59,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
 train_manifest.jsonl filter=lfs diff=lfs merge=lfs -text
+gpt_pairs_train.jsonl filter=lfs diff=lfs merge=lfs -text

gpt_pairs_train.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4f254744e8942d534ce1225783aa2b58dd67376b03439a8bea57c6c59b97c9f
+size 1210788695