fix: lint

pilot/model/adapter.py

@@ -85,6 +85,7 @@ class ChatGLMAdapater(BaseLLMAdaper):
 
 class GuanacoAdapter(BaseLLMAdaper):
     """TODO Support guanaco"""
+
     def match(self, model_path: str):
         return "guanaco" in model_path
 

pilot/model/guanaco_llm.py

@@ -6,16 +6,17 @@ import transformers
 from transformers import GenerationConfig
 from llm_utils import Iteratorize, Stream
 
+
 def guanaco_generate_output(model, tokenizer, params, device):
     """Fork from fastchat: https://github.com/KohakuBlueleaf/guanaco-lora/blob/main/generate.py"""
     prompt = params["prompt"]
     inputs = tokenizer(prompt, return_tensors="pt")
     input_ids = inputs["input_ids"].to(device)
-    temperature=0.5,
+    temperature = (0.5,)
-    top_p=0.95,
+    top_p = (0.95,)
-    top_k=45,
+    top_k = (45,)
-    max_new_tokens=128,
+    max_new_tokens = (128,)
-    stream_output=True
+    stream_output = True
 
     generation_config = GenerationConfig(
         temperature=temperature,
@@ -37,19 +38,13 @@ def guanaco_generate_output(model, tokenizer, params, device):
         # from https://github.com/oobabooga/text-generation-webui/blob/ad37f396fc8bcbab90e11ecf17c56c97bfbd4a9c/modules/text_generation.py#L216-L243.
 
         def generate_with_callback(callback=None, **kwargs):
-                kwargs.setdefault(
+            kwargs.setdefault("stopping_criteria", transformers.StoppingCriteriaList())
-                    "stopping_criteria", transformers.StoppingCriteriaList()
+            kwargs["stopping_criteria"].append(Stream(callback_func=callback))
-                )
-                kwargs["stopping_criteria"].append(
-                    Stream(callback_func=callback)
-                )
             with torch.no_grad():
                 model.generate(**kwargs)
 
         def generate_with_streaming(**kwargs):
-                return Iteratorize(
+            return Iteratorize(generate_with_callback, kwargs, callback=None)
-                    generate_with_callback, kwargs, callback=None
-                )
 
         with generate_with_streaming(**generate_params) as generator:
             for output in generator:
@@ -72,7 +67,7 @@ def guanaco_generate_output(model, tokenizer, params, device):
         )
 
     s = generation_output.sequences[0]
-    print(f"debug_sequences,{s}",s)
+    print(f"debug_sequences,{s}", s)
     output = tokenizer.decode(s)
-    print(f"debug_output,{output}",output)
+    print(f"debug_output,{output}", output)
     yield output.split("### Response:")[-1].strip()

pilot/model/llm_utils.py

@@ -53,6 +53,7 @@ def create_chat_completion(
         response = None
         # TODO impl this use vicuna server api
 
+
 class Stream(transformers.StoppingCriteria):
     def __init__(self, callback_func=None):
         self.callback_func = callback_func

pilot/server/chat_adapter.py

@@ -85,7 +85,7 @@ class CodeGenChatAdapter(BaseChatAdpter):
 
 
 class GuanacoChatAdapter(BaseChatAdpter):
-    """Model chat adapter for Guanaco """
+    """Model chat adapter for Guanaco"""
 
     def match(self, model_path: str):
         return "guanaco" in model_path