Fix eos_token_id to align with vocabulary of this model (#6)

- Fix eos_token_id to align with vocabulary of this model (214ede753741cba42ba95344e4e18f9ab68dce86)


Co-authored-by: Mike Ravkine <[email protected]>

configuration_RW.py

@@ -38,7 +38,7 @@ class RWConfig(PretrainedConfig):
         initializer_range=0.02,
         use_cache=True,
         bos_token_id=1,
-        eos_token_id=2,
+        eos_token_id=11,
         apply_residual_connection_post_layernorm=False,
         hidden_dropout=0.0,
         attention_dropout=0.0,