Portrait de Rahma Nouaji

Rahma Nouaji

Doctorat - McGill
Superviseur⋅e principal⋅e
Oana Balmau
Sujets de recherche
Systèmes d'apprentissage automatique
Systèmes distribués
Site web

Publications

MinatoLoader: Accelerating Machine Learning Training Through Efficient Data Preprocessing
Rahma Nouaji
Stella Bitchebe
Ricardo Macedo
Oana Balmau
Machine learning (ML) frameworks, such as PyTorch and TensorFlow, rely on data loaders to preprocess data before feeding it to accelerators.… (voir plus) When preprocessing is inefficiently pipelined, GPUs can remain idle over long periods of time, leading to substantial training delays. For example, PyTorch's default data loaders can cause up to 76% GPU idleness. A key bottleneck is the variability in preprocessing time across samples within the same dataset. Existing data loaders are oblivious to this variability, training all samples uniformly. In this case, a single slow sample can stall the entire batch, causing head-of-line blocking.
2026-04-25
European Conference on Computer Systems (publié)
doi.org
arxiv.org