mlsanityData to 7 Oct 2026

MCTensor: A High-Precision Deep Learning Library with Multi-Component Floating-Point

arXiv 2207.08867 ↗First posted 18 Jul 2022cs.LG cs.MScited by 3

Not on the map. Why?

Abstract (arXiv)

In this paper, we introduce MCTensor, a library based on PyTorch for providing general-purpose and high-precision arithmetic for DL training. MCTensor is used in the same way as PyTorch Tensor: we implement multiple basic, matrix-level computation operators and NN modules for MCTensor with identical PyTorch interface. Our algorithms achieve high precision computation and also benefits from heavily-optimized PyTorch floating-point arithmetic. We evaluate MCTensor arithmetic against PyTorch native arithmetic for a series of tasks, where models using MCTensor in float16 would match or outperform the PyTorch model with float32 or float64 precision.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 15 of its 35 references that are in the map. First posted on arXiv 18 Jul 2022.

Builds on

15 of 35 references.

References · 15

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  2. 2017Poincar'e Embeddings for Learning Hierarchical Representationscited by 674
  3. 2017In-Datacenter Performance Analysis of a Tensor Processing Unitcited by 610
  4. 2016Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activationscited by 413
  5. 2019Hyperbolic Graph Convolutional Neural Networkscited by 400
  6. 2018Hyperbolic Neural Networkscited by 393
  7. 2014Caffe: Convolutional Architecture for Fast Feature Embeddingcited by 287
  8. 2018Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometrycited by 261
  9. 2019Hyperbolic Graph Neural Networkscited by 244
  10. 2018Representation Tradeoffs for Hyperbolic Embeddingscited by 204
  11. 2015MXNet: A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systemscited by 195
  12. 2018Training and Inference with Integers in Deep Neural Networkscited by 84
3 more
  1. 2019QPyTorch: A Low-Precision Arithmetic Simulation Frameworkcited by 21
  2. 2018High-Accuracy Low-Precision Trainingcited by 20
  3. 2017High performance ultra-low-precision convolutions on mobile devicescited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 1; 2026: 1.

3 papers.

  1. 2024Collage: Light-Weight Low-Precision Strategy for LLM Trainingcited by 9
  2. 2025Low-distortion and GPU-compatible Tree Embeddings in Hyperbolic Spacecited by 6
  3. 2026Rethinking the Fully Hyperbolic Vision Transformer in Polar Coordinatescited by 0