mlsanityData to 7 Oct 2026

Efficient Post-training Quantization with FP8 Formats

arXiv 2309.14592 ↗First posted 26 Sep 2023cs.LG cs.AI cs.CLcited by 29

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Recent advances in deep learning methods such as LLMs and Diffusion models have created a need for improved quantization methods that can meet the computational demands of these modern architectures while maintaining accuracy. Towards this goal, we study the advantages of FP8 data formats for post-training quantization across 75 unique network architectures covering a wide range of tasks, including machine translation, language modeling, text generation, image classification, generation, and segmentation. We examine three different FP8 representations (E5M2, E4M3, and E3M4) to study the effects of varying degrees of trade-off between dynamic range and precision on model accuracy. Based on our extensive study, we developed a quantization workflow that generalizes across different network architectures. Our empirical results show that FP8 formats outperform INT8 in multiple aspects, including workload coverage (92.64% vs. 65.87%), model accuracy and suitability for a broader range of operations. Furthermore, our findings suggest that E4M3 is better suited for NLP models, whereas E3M4 performs marginally better than E4M3 on computer vision tasks. The code is publicly available on Intel Neural Compressor: https://github.com/intel/neural-compressor.

Key facts. Cited by 29 papers in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 50 of its 59 references that are in the map. First posted on arXiv 26 Sep 2023.

Builds on

50 of 59 references.

References · 50

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  4. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  5. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  6. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  7. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  8. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
  9. 2019DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lightercited by 4,243
  10. 2014Going Deeper with Convolutionscited by 4,231
  11. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,853
  12. 2020Longformer: The Long-Document Transformercited by 3,362
38 more
  1. 2018YOLOv3: An Incremental Improvementcited by 2,652
  2. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,301
  3. 2015Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Codingcited by 2,291
  4. 2015Learning both Weights and Connections for Efficient Neural Networkscited by 1,998
  5. 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,968
  6. 2017Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inferencecited by 1,548
  7. 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,338
  8. 2017ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devicescited by 1,194
  9. 2019PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarizationcited by 1,100
  10. 2022LLM.int8(): 8-bit Matrix Multiplication for Transformers at Scalecited by 881
  11. 2019DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generationcited by 842
  12. 2021A Survey of Quantization Methods for Efficient Neural Network Inferencecited by 798
  13. 2019Learned Step Size Quantizationcited by 604
  14. 2016DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradientscited by 557
  15. 2022BioGPT: Generative Pre-trained Transformer for Biomedical Text Generation and Miningcited by 531
  16. 2018Quantizing deep convolutional networks for efficient inference: A whitepapercited by 507
  17. 2018PACT: Parameterized Clipping Activation for Quantized Neural Networkscited by 456
  18. 2019Deep Learning Recommendation Model for Personalization and Recommendation Systemscited by 327
  19. 2018LQ-Nets: Learned Quantization for Highly Accurate and Compact Deep Neural Networkscited by 272
  20. 2021I-BERT: Integer-only BERT Quantizationcited by 250
  21. 2017Incremental Network Quantization: Towards Lossless CNNs with Low-Precision Weightscited by 211
  22. 2017GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibriumcited by 208
  23. 2018Marian: Fast Neural Machine Translation in C++cited by 202
  24. 2022FP8 Formats for Deep Learningcited by 187
  25. 2019Low-bit Quantization of Neural Networks for Efficient Inferencecited by 185
  26. 2020LSQ+: Improving low-bit quantization through learnable offsets and better initializationcited by 180
  27. 2022Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Modelscited by 176
  28. 2018Training Deep Neural Networks with 8-bit Floating Point Numberscited by 163
  29. 2019Improving Neural Network Quantization without Retraining using Outlier Channel Splittingcited by 150
  30. 2019Additive Powers-of-Two Quantization: An Efficient Non-uniform Discretization for Neural Networkscited by 143
  31. 2017Deep Learning with Low Precision by Half-wave Gaussian Quantizationcited by 115
  32. 2016Convolutional Neural Networks using Logarithmic Data Representationcited by 97
  33. 2022FP8 Quantization: The Power of the Exponentcited by 87
  34. 2019Efficient 8-Bit Quantization of Transformer Neural Machine Language Translation Modelcited by 58
  35. 2015Resiliency of Deep Neural Networks under Quantizationcited by 34
  36. 20228-bit Numerical Formats for Deep Neural Networkscited by 23
  37. 2019Mixed Precision Training With 8-bit Floating Pointcited by 22
  38. 2017Ternary Neural Networks with Fine-Grained Quantizationcited by 15

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 3; 2024: 8; 2025: 10; 2026: 8.

29 papers.

Sort:

By area:

  1. 2023How to Capture Higher-order Correlations? Generalizing Matrix Softmax Attention to Kronecker Computationcited by 44
  2. 2023Efficient LLM Inference on CPUscited by 24
  3. 2025FPSAttention: Training-Aware FP8 and Sparsity Co-Design for Fast Video Diffusioncited by 19
  4. 2024Learning from Students: Applying t-Distributions to Explore Accurate and Efficient Formats for LLMscited by 18
  5. 2024To FP8 and Back Again: Quantifying Reduced Precision Effects on LLM Training Stabilitycited by 9
  6. 2024Designing Large Foundation Models for Efficient Training and Inference: A Surveycited by 7
  7. 2024Hardware and Software Platform Inferencecited by 6
  8. 2025FP4DiT: Towards Effective Floating Point Quantization for Diffusion Transformerscited by 6
  9. 2024Accelerating Communication in Deep Learning Recommendation Model Training with Dual-Level Adaptive Lossy Compressioncited by 4
  10. 2025An Inquiry into Datacenter TCO for LLM Inference with FP8cited by 4
  11. 2025Optimizing Singular Spectrum for Large Language Model Compressioncited by 4
  12. 2026Float8@2bits: Entropy Coding Enables Data-Free Model Compressioncited by 2
17 more
  1. 2023Local Masking Meets Progressive Freezing: Crafting Efficient Vision Transformers for Self-Supervised Learningcited by 1
  2. 2024Towards Federated Learning with On-device Training and Communication in 8-bit Floating Pointcited by 1
  3. 2024On the Impact of White-box Deployment Strategies for Edge AI on Latency and Model Performancecited by 1
  4. 2025Precision Where It Matters: A Novel Spike Aware Mixed-Precision Quantization Strategy for LLaMA-based Language Modelscited by 1
  5. 2025Optimization of embeddings storage for RAG systems using quantization and dimensionality reduction techniquescited by 1
  6. 2026TACO: Efficient Communication Compression of Intermediate Tensors for Scalable Tensor-Parallel LLM Trainingcited by 1
  7. 2024BEExformer: A Fast Inferencing Binarized Transformer with Early Exitscited by 0
  8. 2025Reliable Evaluation Protocol for Low-Precision Retrievalcited by 0
  9. 2025Quantizing Whisper-small: How design choices affect ASR performancecited by 0
  10. 2025BitSnap: Checkpoint Sparsification and Quantization in LLM Trainingcited by 0
  11. 2025Arithmetic-Intensity-Aware Quantizationcited by 0
  12. 2026Lightweight Transformer Architectures for Edge Devices in Real-Time Applicationscited by 0
  13. 2026SNIP: An Adaptive Mixed Precision Framework for Subbyte Large Language Model Trainingcited by 0
  14. 2026RT-Lynx: Putting GEMM Sparsity in the Right Place for Diffusion Modelscited by 0
  15. 2026Towards Scalable Customization and Deployment of Multi-Agent Systems for Enterprise Applicationscited by 0
  16. 2026SparkDiffusion: Mitigating the High-Sparsity Trap --- A Unified Framework for up to 265× Single-GPU Acceleration of Visual Generationcited by 0
  17. 2026DPS: Dual-Mode Precision LLM Serving with Semi-Unified Memorycited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024"Give Me BF16 or Give Me Death"? Accuracy-Performance Trade-Offs in LLM Quantization
  2. 2025A Comprehensive FP8 Training Recipe for Reasoning-Enhanced Language Models
  3. 2023FP8-LM: Training FP8 Large Language Models
  4. 2025Benchmarking Post-Training Quantization in LLMs: Comprehensive Taxonomy, Unified Evaluation, and Comparative Analysis
  5. 2025Mixed-Precision Quantization for Language Models: Techniques and Prospects
  6. 2026Holding the FP8 Quality Ceiling at 8-Bit Weights and Activations: INT8 and GGUF Post-Training Quantization of Ideogram 4.0 for Consumer GPUs
  7. 2024LLMC: Benchmarking Large Language Model Quantization with a Versatile Compression Toolkit
  8. 2026Stable FP4 Training via Transposition-Invariant Block Quantization