Create confidence.py

confidence.py

@@ -0,0 +1,40 @@
+from fastapi import FastAPI
+from pydantic import BaseModel
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+import torch
+
+# Initialize FastAPI app
+app = FastAPI(title="Confidence Statement API", version="1.0")
+
+# Load the fine-tuned model and tokenizer
+model_name = "mjpsm/Confidence-Statement-Model-final"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForSequenceClassification.from_pretrained(model_name)
+
+# Define input format
+class InputText(BaseModel):
+    statement: str
+
+# Define prediction function
+def predict_statement(statement: str):
+    inputs = tokenizer(statement, return_tensors="pt", padding=True, truncation=True, max_length=128)
+    with torch.no_grad():
+        outputs = model(**inputs)
+    logits = outputs.logits
+    probabilities = torch.nn.functional.softmax(logits, dim=-1)
+    predicted_class = torch.argmax(probabilities, dim=-1).item()
+    label_mapping = {0: "lack of self-confidence", 1: "self-confident"}
+    return {
+        "label": label_mapping[predicted_class],
+        "confidence_score": round(probabilities[0][predicted_class].item(), 4)
+    }
+
+# Define root route
+@app.get("/")
+def read_root():
+    return {"message": "Welcome to the Confidence Statement API!"}
+
+# Define prediction route
+@app.post("/predict")
+def predict(input_text: InputText):
+    return predict_statement(input_text.statement)