Bert Base Uncased Squadv1 X2.44 F87.7 D26 Hybrid Filled V1

🚀 BERT-base uncased model fine-tuned on SQuAD v1

This model is a fine-tuned BERT-base uncased model on SQuAD v1, which uses pruning techniques to reduce weights and speed up inference while maintaining high accuracy.

🚀 Quick Start

To use this model, first install the nn_pruning library:

pip install nn_pruning

Then, you can use the transformers library as follows:

from transformers import pipeline
from nn_pruning.inference_model_patcher import optimize_model

qa_pipeline = pipeline(
    "question-answering",
    model="madlag/bert-base-uncased-squadv1-x2.44-f87.7-d26-hybrid-filled-v1",
    tokenizer="madlag/bert-base-uncased-squadv1-x2.44-f87.7-d26-hybrid-filled-v1"
)

print("/home/lagunas/devel/hf/nn_pruning/nn_pruning/analysis/tmp_finetune parameters: 189.0M")
print(f"Parameters count (includes only head pruning, not feed forward pruning)={int(qa_pipeline.model.num_parameters() / 1E6)}M")
qa_pipeline.model = optimize_model(qa_pipeline.model, "dense")

print(f"Parameters count after complete optimization={int(qa_pipeline.model.num_parameters() / 1E6)}M")
predictions = qa_pipeline({
    'context': "Frédéric François Chopin, born Fryderyk Franciszek Chopin (1 March 1810 – 17 October 1849), was a Polish composer and virtuoso pianist of the Romantic era who wrote primarily for solo piano.",
    'question': "Who is Frederic Chopin?",
})
print("Predictions", predictions)

✨ Features

• Pruned Weights: The linear layers contain 26.0% of the original weights, and the model contains 42.0% of the original weights overall.
• Faster Inference: With a simple resizing of the linear matrices, it runs 2.44x as fast as the original model on the evaluation.
• High Accuracy: Its F1 is 87.71, with an F1 drop of only 0.79 compared to the original model.

🔧 Technical Details

Model Creation

This model was created using the nn_pruning python library. The pruning method leads to structured matrices, which can be visualized by hovering on the plot below.

Fine-Pruning Details

This model was fine-tuned from the HuggingFace model checkpoint on SQuAD1.1, and distilled from the model csarron/bert-base-uncased-squad-v1. It is case-insensitive.

A side-effect of the block pruning is that some of the attention heads are completely removed: 80 heads were removed on a total of 144 (55.6%). Here is a detailed view on how the remaining heads are distributed in the network after pruning.

Dataset Details

Fine-tuning Environment

• Python: 3.8.5
• Machine specs:

Memory: 64 GiB
GPUs: 1 GeForce GTX 3090, with 24GiB memory
GPU driver: 455.23.05, CUDA: 11.1

Results

Pytorch model file size: 355MB (original BERT: 420MB)

📄 License

This model is released under the MIT license.