Learn to say popular machine learning compiler and kernel tool names correctly.
0 / 5 completed
1 / 5
How is TVM (Apache's open-source deep learning compiler stack for optimizing models across hardware) correctly pronounced?
TVM is pronounced 'TEE-VEE-EM' — every letter spoken individually, T-V-M, short for Tensor Virtual Machine. In a technical interview: "TVM compiled the model down to a kernel tuned specifically for our edge device."
2 / 5
How is XLA (Google's Accelerated Linear Algebra compiler used to speed up machine learning workloads) correctly pronounced?
XLA is pronounced 'EKS-EL-AY' — every letter spoken individually, X-L-A. In a technical interview: "XLA fused a dozen small operations into a single kernel and cut our training step time noticeably."
3 / 5
How is MLIR (Multi-Level Intermediate Representation compiler infrastructure used across many ML frameworks) correctly pronounced?
MLIR is pronounced 'EM-EL-EYE-AR' — every letter spoken individually, M-L-I-R. In a technical interview: "MLIR let the compiler represent the model at several levels of abstraction in one unified pipeline."
4 / 5
How is ONNX Runtime (cross-platform inference engine for running models exported in the ONNX format) correctly pronounced?
ONNX Runtime is pronounced 'AHN-eks RUN-tym' — 'ONNX' rhymes with 'lonnix', plus 'runtime'. In a technical interview: "ONNX Runtime ran the same exported model on both our server and our mobile app."
5 / 5
How is Glow (Facebook's machine learning compiler for optimizing neural network execution on hardware) correctly pronounced?
Glow (the ML compiler) is pronounced 'GLOH' — exactly like the everyday word for a soft light, one syllable. In a technical interview: "Glow lowered the graph into hardware-specific instructions the accelerator could execute directly."