app.py

from deepsparse import Pipeline
import time
import gradio as gr

markdownn = '''
# Text Classification Pipeline with DeepSparse
Text Classification involves assigning a label to a given text. For example, sentiment analysis is an example of a text classification use case. 
![Text Classification Pipeline with DeepSparse](https://huggingface.co/spaces/neuralmagic/nlp-text-classification/resolve/main/text-classification.png)
## What is DeepSparse
DeepSparse is an inference runtime offering GPU-class performance on CPUs and APIs to integrate ML into your application. Sparsification is a powerful technique for optimizing models for inference, reducing the compute needed with a limited accuracy tradeoff. DeepSparse is designed to take advantage of model sparsity, enabling you to deploy models with the flexibility and scalability of software on commodity CPUs with the best-in-class performance of hardware accelerators, enabling you to standardize operations and reduce infrastructure costs.
Similar to Hugging Face, DeepSparse provides off-the-shelf pipelines for computer vision and NLP that wrap the model with proper pre- and post-processing to run performantly on CPUs by using sparse models.
The text classification Pipeline, for example, wraps an NLP model with the proper preprocessing and postprocessing pipelines, such as tokenization.
### Inference API Example
Here is sample code for a text classification pipeline: 
```
from deepsparse import Pipeline
pipeline = Pipeline.create(task="zero_shot_text_classification", model_path="zoo:nlp/text_classification/distilbert-none/pytorch/huggingface/mnli/pruned80_quant-none-vnni",model_scheme="mnli",model_config={"hypothesis_template": "This text is related to {}"},)
text = "The senate passed 3 laws today"
inference = pipeline(sequences= text,labels=['politics', 'public health', 'Europe'],)
print(inference)
```
## Use Case Description
Customer review classification is a great example of text classification in action. 
The ability to quickly classify sentiment from customers is an added advantage for any business. 
Therefore, whichever solution you deploy for classifying the customer reviews should deliver results in the shortest time possible. 
By being fast the solution will process more volume, hence cheaper computational resources are utilized.  
When deploying a text classification model, decreasing the model’s latency and increasing its throughput is critical. This is why DeepSparse Pipelines have sparse text classification models. 
[Want to train a sparse model on your data? Checkout the documentation on sparse transfer learning](https://docs.neuralmagic.com/use-cases/natural-language-processing/question-answering)
'''
task = "zero_shot_text_classification"
sparse_classification_pipeline = Pipeline.create(
        task=task,
        model_path="zoo:nlp/text_classification/distilbert-none/pytorch/huggingface/mnli/pruned80_quant-none-vnni",
      model_scheme="mnli",
    model_config={"hypothesis_template": "This text is related to {}"},
    )
def run_pipeline(text):
    sparse_start = time.perf_counter()
    sparse_output = sparse_classification_pipeline(sequences= text,labels=['politics', 'public health', 'Europe'],)
    sparse_result = dict(sparse_output)
    sparse_end = time.perf_counter()
    sparse_duration = (sparse_end - sparse_start) * 1000.0
    dict_r = {sparse_result['labels'][0]:sparse_result['scores'][0],sparse_result['labels'][1]:sparse_result['scores'][1], sparse_result['labels'][2]:sparse_result['scores'][2]}
    return dict_r, sparse_duration


with gr.Blocks() as demo:
    with gr.Row():
        with gr.Column():
            gr.Markdown(markdownn)
            
        with gr.Column():
            gr.Markdown("""
            ### Text classification demo
            """)
            text = gr.Text(label="Text")
            btn = gr.Button("Submit")
            
            sparse_answers = gr.Label(label="Sparse model answers",
                                    num_top_classes=3
                                     )
            sparse_duration = gr.Number(label="Sparse Latency (ms):")
            gr.Examples([["The senate passed 3 laws today"],["Who are you voting for in 2020?"],["Public health is very important"]],inputs=[text],)
    
    btn.click(
        run_pipeline,
        inputs=[text],
        outputs=[sparse_answers,sparse_duration],
    )

if __name__ == "__main__":
    demo.launch()