import torch
from transformers import BertForSequenceClassification
import gradio as gr
from transformers import BertTokenizer
import torch
from transformers import BertForSequenceClassification, BertTokenizer
import gradio as gr

# Load your BERT model
# Load the model architecture
model = BertForSequenceClassification.from_pretrained('bert-base-uncased', num_labels=2)

# Load the state dict without weights_only
try:
    model.load_state_dict(torch.load('bert_model_complete.pth', map_location=torch.device('cpu')), strict=False)
except Exception as e:
    print(f"Error loading state dict: {e}")
model.eval()  # Set the model to evaluation mode


# Load the tokenizer
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')

def predict(text):
    inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True)
    with torch.no_grad():
        outputs = model(**inputs)
    logits = outputs.logits
    predicted_class = logits.argmax().item()
    return predicted_class

# Set up the Gradio interface
interface = gr.Interface(fn=predict, inputs="text", outputs="label", title="BERT Text Classification")

# Load model and tokenizer
model = BertForSequenceClassification.from_pretrained('bert-base-uncased')
model.load_state_dict(torch.load('bert_model_complete.pth'))
model.eval()

tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')

# Define prediction function
def predict(text):
    inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True)
    with torch.no_grad():
        outputs = model(**inputs)
    logits = outputs.logits
    predicted_class = logits.argmax().item()
    return predicted_class

# Set up Gradio interface
interface = gr.Interface(fn=predict, inputs="text", outputs="label", title="BERT Text Classification")

# Launch the interface
if __name__ == "__main__":
    interface.launch()