app.py

from transformers.utils import logging
from language_directions import *
logging.set_verbosity_error()

import warnings
warnings.filterwarnings("ignore", 
                        message="Using the model-agnostic default `max_length`")

import os
import gradio as gr
from transformers import pipeline

pipe = pipeline("image-to-text",
                model="Salesforce/blip-image-captioning-base")


def translate(input_text, source, target):
    source_readable = source
    if source == "Auto Detect" or source.startswith("Detected"):
      source, _ = auto_detect_language_code(input_text)
    if source in source_lang_dict.keys():
      source = source_lang_dict[source]
    target_lang_dict, _ = get_target_languages(source)
    try:
      target = target_lang_dict[target]
      # Helsinki-NLP/opus-mt-en-sq
      model = f"Helsinki-NLP/opus-mt-{source}-{target}"
      pipe = pipeline("translation", model=model)
      translation = pipe(input_text)
      return translation[0]['translation_text'], ""
    except KeyError:
      return "", f"Error: Translation direction {source_readable} to {target} is not supported by Helsinki Translation Models"

def launch(input):
    out = pipe(input)
    context_str = out[0]['generated_text']
    translate_str = translate(context_str, 'en', 'sq')
    return translate_str

iface = gr.Interface(launch,
                     inputs=gr.Image(type='pil'),
                     outputs="text")

iface.launch(share=True)
# iface.launch(share=True, 
#              server_port=int(os.environ['PORT1']))

iface.close()