Raibows/DynamicBatchSampler

Yet another dynamic batch sampler for variable sequence data in PyTorch.

21
/ 100
Experimental

This tool helps machine learning engineers and researchers efficiently train models on variable-length text data, like sentences in natural language processing. It takes in a dataset of text samples and their lengths, then organizes them into batches that intelligently utilize GPU memory. The output is a data loader that feeds optimized batches to your PyTorch model, speeding up training.

No commits in the last 6 months.

Use this if you are training deep learning models on text or other sequence data in PyTorch and want to make your training more efficient, especially when dealing with varying sequence lengths across your dataset.

Not ideal if your dataset consists of fixed-length inputs, as the benefits of dynamic batching for sequence length optimization would not apply.

natural-language-processing deep-learning-training text-classification sequence-modeling pytorch-optimization
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 16 / 25
Community 0 / 25

How are scores calculated?

Stars

13

Forks

Language

Python

License

GPL-3.0

Last pushed

Dec 09, 2021

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/nlp/Raibows/DynamicBatchSampler"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.