MLIR (software)

MLIR was publicly released as part of LLVM in 2019

MLIR, a compiler infrastructure project, was integrated into the LLVM project in 2019, enhancing its capabilities and reach.

MLIR's integration into LLVM allowed it to leverage the existing infrastructure and tools provided by LLVM, which is known for its robust compiler infrastructure.

Example

TensorFlow utilizes MLIR to improve its compilation process for modern workloads.

Remember this

MLIR's release as part of LLVM in 2019 signifies a significant milestone in its development and adoption, particularly for projects like TensorFlow that benefit from its advanced features.

Text adapted from Wikipedia, licensed under CC BY-SA 4.0.

Related concepts

the ONNX format does: standardizes model representation for cross-framework deployment

ONNX standardizes machine learning model representation

384-dim all-MiniLM-L6-v2 optimizes: fast sentence similarity with 6 layers

All-MiniLM-L6-v2 optimizes fast sentence similarity with 6 layers

768-dim BERT embeddings capture: bidirectional context from masked language modeling

768-dim BERT embeddings capture bidirectional context from masked language modeling

[CLS] pooling does: uses the first token's embedding as the sentence representation

CLS pooling: uses the first token's embedding as the sentence representation

XLA does for TensorFlow/JAX: compiles computation graphs for TPU/GPU execution

XLA accelerates TensorFlow models

Alex Lora Cercos

Alex Lora is a Spanish film director

Swipe through 100 ML concepts daily

Open Pocket Polymath