import gradio as gr

models = {
    # "object-detection": "facebook/detr-resnet-50",
    "image-classification": "microsoft/resnet-50",
    "text-to-image": "runwayml/stable-diffusion-v1-5",
    "image-to-text": "nlpconnect/vit-gpt2-image-captioning",
    "audio-classification": "mtg-upf/discogs-maest-30s-pw-73e-ts",
    "audio-to-audio": "speechbrain/mtl-mimic-voicebank",
    "automatic-speech-recognition": "jonatasgrosman/wav2vec2-large-xlsr-53-english",
    "conversational": "microsoft/DialoGPT-medium",
    "feature-extraction": "cambridgeltl/SapBERT-from-PubMedBERT-fulltext",
    "fill-mask": "bert-base-uncased",
    "question-answering": "deepset/roberta-base-squad2",
    "summarization": "facebook/bart-large-cnn",
    "text-classification": "cardiffnlp/twitter-roberta-base-sentiment-latest",
    "text-generation": "gpt2",
    "text2text-generation": "vennify/t5-base-grammar-correction",
    "translation": "Helsinki-NLP/opus-mt-fr-en",
    "zero-shot-classification": "facebook/bart-large-mnli",
    "sentence-similarity": "sentence-transformers/all-mpnet-base-v2",
    "text-to-speech": "facebook/mms-tts-eng",
    "token-classification": "benjamin/wtp-canine-s-1l",
    "document-question-answering": "fxmarty/tiny-doc-qa-vision-encoder-decoder",
    "visual-question-answering": "Salesforce/blip-vqa-capfilt-large",
}

with gr.Blocks() as demo:
    gr.Markdown("## Gradio Pipelines Tasks")
    for k, v in models.items():
        with gr.Tab(k):
            gr.load(v, src="models")
demo.launch()