Spaces:
Sleeping
Sleeping
from fastapi import FastAPI | |
import models | |
from schema import Prediction | |
from sentence_transformers import util | |
app = FastAPI() | |
def home_page(): | |
return {"Home": "Welcome to prediction hub"} | |
def display_embedding(message : str = "Hello guys enter a text to get embeddings"): | |
try: | |
embedding = models.get_embedding(message) | |
dimension = len(embedding) | |
return {"Dimension" : {dimension : embedding.tolist()}} | |
except Exception as e: | |
return {f"Unable to fetch the embeddings. Error :{e}" } | |
def display_prediction(prediction : Prediction): | |
message = prediction.message | |
embedding = models.get_embedding([message]) | |
loaded_model = models.load_model('log_reg_model.pkl') | |
result = loaded_model.predict(embedding).tolist() | |
return {"Prediction": f"{message} is a {result}"} | |
def display_cosine_similarity(prediction : Prediction): | |
message = prediction.message | |
message_1 = prediction.message_1 | |
embendding = models.get_embedding([message,message_1]) | |
similarity = util.cos_sim(embendding[0], embendding[1]).item() | |
return {f"Cosine Similarity between {message} and {message_1} is" : round(similarity, 4)} | |