← Glossary
Glossary
CUDA
NVIDIA’s platform for running general-purpose computation on the GPU instead of the CPU — what PyTorch/TensorFlow use under the hood to train and run models fast. Most "out of memory" and device-mismatch errors are CUDA-level.
Related articles
Fix RuntimeError: CUDA error device-side assert triggered in PyTorchFix FAISS_THROW_IF_NOT(assign_index->ntotal != k), IndexError, Dimension MismatchFix FAISS nprobe/nlist Slowdowns, IndexIVFPQ Training Asserts, and StandardGpuResources Alloc FailHow Neural Networks Work: Forward Pass & Backprop in NumPyFix OSError: [Errno 28] No space left on Device — HuggingFace Model CacheFix Hugging Face CUDA Out of Memory When Loading Transformer ModelsFix HuggingFace Tokenizer Padding Errors: No pad_token, Truncation, Left vs RightFix PyTorch RuntimeError: Expected All Tensors to Be on the Same DeviceRecommender Systems That Automate Campaign Execution: NLP and Collaborative Filtering in Practicetorch.cuda.OutOfMemoryError: How to Fix GPU Out of Memory in PyTorchFix PyTorch RuntimeError: Expected Scalar Type Long but Found Float
Related tools