Set repeat_penalty to 0 by default

abetlen · abetlen · commit 0d751a69a78c · 2023-05-08T01:50:43.000-04:00
diff --git a/llama_cpp/server/app.py b/llama_cpp/server/app.py
@@ -146,7 +146,7 @@ def get_llama():
 )
 
 repeat_penalty_field = Field(
-    default=1.0,
+    default=0.0,
     ge=0.0,
     description="A penalty applied to each token that is already generated. This helps prevent the model from repeating itself.\n\n"
     + "Repeat penalty is a hyperparameter used to penalize the repetition of token sequences during text generation. It helps prevent the model from generating repetitive or monotonous text. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient.",

Original file line number	Diff line number	Diff line change
`@@ -146,7 +146,7 @@ def get_llama():`
`146`	`146`	`)`
`147`	`147`
`148`	`148`	`repeat_penalty_field = Field(`
`149`		`- default=1.0,`
	`149`	`+ default=0.0,`
`150`	`150`	`ge=0.0,`
`151`	`151`	`description="A penalty applied to each token that is already generated. This helps prevent the model from repeating itself.\n\n"`
`152`	`152`	`+ "Repeat penalty is a hyperparameter used to penalize the repetition of token sequences during text generation. It helps prevent the model from generating repetitive or monotonous text. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient.",`