app.py

import gradio as gr
from model import SurfinBird
from torchvision import transforms, io
import csv
import torch

with open("birds.csv", "r") as r:
  birds = list(csv.reader(r, delimiter=","))
  birds = birds[0]

config = {"num_channels": 3, "hidden_units": 256, "num_classes": 525, "labels": birds}

model = SurfinBird(config=config)
model = SurfinBird.from_pretrained("ulichovick/birdnet")

titulo = "Gimme da bird!"
desc = "Carga la imágen de un ave y el modelo intentará identificar la especie del ave en la foto(limitado a 525 especies) "


def predice(imagen):
    usr_img_transform = transforms.Compose([
    transforms.Resize(size=(224, 224)),
    ])
    target_image = io.read_image(str(imagen)).type(torch.float32)
    target_image = target_image / 255.
    target_image = usr_img_transform(target_image)
    model.eval()
    with torch.inference_mode():
      target_image = target_image.unsqueeze(dim=0)
      target_image_pred = model(target_image)
    target_image_pred_label = torch.argmax(target_image_pred, dim=1)
    label = config["labels"][target_image_pred_label.item()]
    return str(label)

gr.Interface(
    predice,
    inputs=gr.Image(label="gimme da bird", type="filepath"),
    outputs=["text"],
    title=titulo,
    description=desc,
).launch()