import gradio as gr import torch from transformers import BartForConditionalGeneration, BartTokenizer # initialize model + tok variables model = None tok = None # Examples for each models examples = [ ["interview-question-remake", "I have a cat named dolche and he's not very friendly with strangers. I've had him for 9 years now and it has been a pleasure to see him grow closer to us every year."], ["interview-length-tagged","Today's weather was really nice."], ["reverse-interview-question", "There are so many incredible musicians out there and so many really compelling big hits this year that it makes for a really interesting way to recap some of those big events."] ] # Descriptions for each models # descriptions = "Interview question remake is a model that..." # pass in Strings of model choice and input text for context def genQuestion(model_choice, context): # global descriptions if model_choice=="interview-question-remake": model = BartForConditionalGeneration.from_pretrained("hyechanjun/interview-question-remake") tok = BartTokenizer.from_pretrained("hyechanjun/interview-question-remake") # descriptions = "Interview question remake is a model that..." elif model_choice=="interview-length-tagged": model = BartForConditionalGeneration.from_pretrained("hyechanjun/interview-length-tagged") tok = BartTokenizer.from_pretrained("hyechanjun/interview-length-tagged") # descriptions = "Interview question tagged is a model that..." elif model_choice=="reverse-interview-question": model = BartForConditionalGeneration.from_pretrained("hyechanjun/reverse-interview-question") tok = BartTokenizer.from_pretrained("hyechanjun/reverse-interview-question") # descriptions = "Reverse interview question is a model that..." inputs = tok(context, return_tensors="pt") output = model.generate(inputs["input_ids"], num_beams=4, max_length=64, min_length=9, num_return_sequences=4, diversity_penalty =1.0, num_beam_groups=4) final_output = '' for i in range(4): final_output += [tok.decode(beam, skip_special_tokens=True, clean_up_tokenization_spaces=False) for beam in output][i] + "\n" return final_output iface = gr.Interface(fn=genQuestion, inputs=[gr.inputs.Dropdown(["interview-question-remake", "interview-length-tagged", "reverse-interview-question"]), "text"], examples=examples, outputs="text") iface.launch()