Update tasks/text.py

tasks/text.py

@@ -9,7 +9,7 @@ from .utils.emissions import tracker, clean_emissions_data, get_space_info
 
 router = APIRouter()
 
-DESCRIPTION = "Random Baseline"
+DESCRIPTION = "FrugalDisinfoHunter Model"
 ROUTE = "/text"
 
 @router.post(ROUTE, tags=["Text Task"], 
@@ -18,9 +18,7 @@ async def evaluate_text(request: TextEvaluationRequest):
     """
     Evaluate text classification for climate disinformation detection.
     
-    Current Model: Random Baseline
-    - Makes random predictions from the label space (0-7)
-    - Used as a baseline for comparison
+    Current Model: FrugalDisinfoHunter
     """
     # Get space info
     username, space_url = get_space_info()
@@ -55,46 +53,29 @@ async def evaluate_text(request: TextEvaluationRequest):
     # YOUR MODEL INFERENCE CODE HERE
     # Update the code below to replace the random baseline by your model inference within the inference pass where the energy consumption and emissions are tracked.
     #--------------------------------------------------------------------------------------------   
-    
-    #--------------------------------------------------------------------------------------------
-    # Load your model and tokenizer from Hugging Face
-    #--------------------------------------------------------------------------------------------
 
     model_name = "Zen0/FrugalDisinfoHunter"  # Model identifier from Hugging Face
     tokenizer = AutoTokenizer.from_pretrained(model_name)
     model = AutoModelForSequenceClassification.from_pretrained(model_name)
 
-    #--------------------------------------------------------------------------------------------
-    # Load the dataset
-    #--------------------------------------------------------------------------------------------
-
-    # Assuming 'quotaclimat/frugalaichallenge-text-train' is the dataset you are working with
-    dataset = load_dataset("quotaclimat/frugalaichallenge-text-train")
-
-    # Access the test dataset (you can change this if you want to use a different split)
-    test_dataset = dataset['test']  # Assuming you have a 'test' split available
-
-    #--------------------------------------------------------------------------------------------
-    # Tokenize the text data
-    #--------------------------------------------------------------------------------------------
-
-    # Tokenize the test data (the text field contains the quotes)
     test_texts = test_dataset["text"]  # The field 'text' contains the climate quotes
-
     inputs = tokenizer(test_texts, padding=True, truncation=True, return_tensors="pt", max_length=512)
+    
+    dataset = load_dataset("quotaclimat/frugalaichallenge-text-train")
 
-    #--------------------------------------------------------------------------------------------
-    # Inference
-    #--------------------------------------------------------------------------------------------
+    # Access the test dataset
+    test_dataset = dataset['test']  
 
-    # Run inference on the dataset using the model
     with torch.no_grad():  # Disable gradient calculations
         outputs = model(**inputs)
         logits = outputs.logits
 
-    # Get predictions from the logits (choose the class with the highest logit)
+    # Get predictions from the logits
     predictions = torch.argmax(logits, dim=-1).cpu().numpy()  # Convert to numpy array for use
 
+    # Get true labels for accuracy calculation
+    true_labels = test_dataset["label"]  # Extract true labels from the dataset
+
     #--------------------------------------------------------------------------------------------
     # YOUR MODEL INFERENCE STOPS HERE
     #--------------------------------------------------------------------------------------------