Mon Jan 30 2023
Sun Jan 29 2023

SWARM Parallelism: Training Large Models Can Be Surprisingly Communication-Efficient

Deep learning
Model-parallel algorithms
Large-scale training approaches
Training large machine learning models using cheap 'preemptible' instances or pooled existing resources from multiple regions

Proposes a model-parallel training algorithm for large models using cheap 'preemptible' instances or pooled existing resources from multiple regions, designed for poorly connected, heterogeneous, and unreliable devices. Showcases the ability to train a large Transformer language model with 1B shared parameters on preemptible T4 GPUs with less than 200Mb/s network.

Implement SWARM parallelism as an alternative setup for training large models to reduce communication requirements and cost.

Moûsai: Text-to-Music Generation with Long-Context Latent Diffusion

Media synthesis
Music generation
Diffusion models
Generating high-quality stereo music at 48kHz from textual descriptions

Develops a cascading latent diffusion approach for generating high-quality stereo music at 48kHz from textual descriptions. Provides open-source libraries to facilitate future work in the field.

Use Moûsai's cascading latent diffusion approach for text-to-music generation.

Leveraging the Third Dimension in Contrastive Learning

Computer vision
Self-supervised learning methods
RGB+depth input representation
Incorporating depth signals into SSL framework

Shows that incorporating the depth channel into SSL methods leads to an increase in downstream classification accuracy on ImageNette and ImageNet-C datasets.

Incorporate depth signals into SSL methods to improve robustness and generalization.

Thu Jan 26 2023
Wed Jan 25 2023
Tue Jan 24 2023
Mon Jan 23 2023