reformatted code

fastchat/serve/monitor/classify/category.py

@@ -11,9 +11,7 @@
 import ast
 import re
 
-from utils import (
-    HuggingFaceRefusalClassifier
-)
+from utils import HuggingFaceRefusalClassifier
 
 
 class Category:
@@ -192,6 +190,9 @@ def __init__(self):
     def pre_process(self, conversation):
         conv = []
         for i in range(0, len(conversation), 2):
-            args = {"QUERY": conversation[i]["content"], "RESPONSE": conversation[i+1]["content"]}
+            args = {
+                "QUERY": conversation[i]["content"],
+                "RESPONSE": conversation[i + 1]["content"],
+            }
             conv.append(self.prompt_template.format(**args))
         return conv

fastchat/serve/monitor/classify/label.py

@@ -11,10 +11,7 @@
 
 from category import Category
 
-from utils import (
-    api_config,
-    chat_completion_openai
-)
+from utils import api_config, chat_completion_openai
 
 LOCK = threading.RLock()
 
@@ -58,7 +55,6 @@ def get_answer(
     output_log = {}
 
     for category in categories:
-
         if category.name_tag == "refusal_v0.2":
             refusal_classifier = category.classifier
 
@@ -69,10 +65,10 @@ def get_answer(
             batch_size = 16
             refusal_results = []
             for i in range(0, len(refusal_prompts), batch_size):
-                batch_prompts = refusal_prompts[i:i + batch_size]
+                batch_prompts = refusal_prompts[i : i + batch_size]
                 batch_results = refusal_classifier.classify_batch(batch_prompts)
                 refusal_results.extend(batch_results)
-            
+
             # If any query/resp classified as refusal, entire conversation is refusal
             output = any(refusal_results)
 

fastchat/serve/monitor/classify/utils.py

@@ -66,24 +66,24 @@ def chat_completion_openai(model, messages, temperature, max_tokens, api_dict=No
 class HuggingFaceRefusalClassifier:
     def __init__(self):
         print("Loading model and tokenizer...")
-        self.tokenizer = AutoTokenizer.from_pretrained("derixu/refusal_classifier-mlm_then_classifier_v3") #TODO: Migrate to LMSYS account and change path
-        self.model = AutoModelForSequenceClassification.from_pretrained("derixu/refusal_classifier-mlm_then_classifier_v3")
-        self.model.eval() 
+        self.tokenizer = AutoTokenizer.from_pretrained(
+            "derixu/refusal_classifier-mlm_then_classifier_v3"
+        )  # TODO: Migrate to LMSYS account and change path
+        self.model = AutoModelForSequenceClassification.from_pretrained(
+            "derixu/refusal_classifier-mlm_then_classifier_v3"
+        )
+        self.model.eval()
 
     def classify_batch(self, input_texts):
         inputs = self.tokenizer(
-            input_texts, 
-            truncation=True, 
-            max_length=512, 
-            return_tensors="pt", 
-            padding=True
+            input_texts,
+            truncation=True,
+            max_length=512,
+            return_tensors="pt",
+            padding=True,
         )
-        with torch.no_grad(): 
+        with torch.no_grad():
             outputs = self.model(**inputs)
             probabilities = torch.nn.functional.softmax(outputs.logits, dim=-1)
             pred_classes = torch.argmax(probabilities, dim=-1).tolist()
         return [bool(pred) for pred in pred_classes]
-
-
-
-