Update tasks/text.py

tasks/text.py

@@ -53,29 +53,49 @@ 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 or local path
+    #--------------------------------------------------------------------------------------------
+
+    from transformers import AutoTokenizer, AutoModelForSequenceClassification
+    import torch
 
-    model_name = "Zen0/FrugalDisinfoHunter"  # Model identifier from Hugging Face
+    # Load model from Hugging Face (adjust if you uploaded it there)
+    model_name = "Zen0/FrugalDisinfoHunter"  # Replace with your model identifier if different
     tokenizer = AutoTokenizer.from_pretrained(model_name)
     model = AutoModelForSequenceClassification.from_pretrained(model_name)
 
+    #--------------------------------------------------------------------------------------------
+    # Load the dataset
+    #--------------------------------------------------------------------------------------------
+
+    # Assuming 'quotaclimat/frugalaichallenge-text-train' is the dataset you're working with
+    dataset = load_dataset(request.dataset_name)
+
+    # 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")
 
-    # Access the test dataset
-    test_dataset = dataset['test']  
+    #--------------------------------------------------------------------------------------------
+    # Inference
+    #--------------------------------------------------------------------------------------------
 
+    # 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
+    # Get predictions from the logits (choose the class with the highest logit)
     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
     #--------------------------------------------------------------------------------------------