Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
149 tokens/sec
GPT-4o
9 tokens/sec
Gemini 2.5 Pro Pro
47 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

TesseraQ: Ultra Low-Bit LLM Post-Training Quantization with Block Reconstruction (2410.19103v1)

Published 24 Oct 2024 in cs.LG

Abstract: LLMs have revolutionized natural language processing, albeit at the cost of immense memory and computation requirements. Post-training quantization (PTQ) is becoming the de facto method to reduce the memory footprint and improve the inference throughput of LLMs. In this work, we aim to push the upper limit of LLM PTQ by optimizing the weight rounding parameters with the block reconstruction technique, a predominant method in previous vision models. We propose TesseraQ, a new state-of-the-art PTQ technique, to quantize the weights of LLMs to ultra-low bits. To effectively optimize the rounding in LLMs and stabilize the reconstruction process, we introduce progressive adaptive rounding. This approach iteratively transits the soft rounding variables to hard variables during the reconstruction process. Additionally, we optimize the dequantization scale parameters to fully leverage the block reconstruction technique. We demonstrate that TesseraQ can be seamlessly integrated with existing scaling or clipping-based PTQ algorithms such as AWQ and OmniQuant, significantly enhancing their performance and establishing a new state-of-the-art. For instance, when compared to AWQ, TesseraQ improves the wikitext2 perplexity from 14.65 to 6.82 and average downstream accuracy from 50.52 to 59.27 with 2-bit weight-only quantization of LLaMA-2-7B. Across a range of quantization schemes, including W2A16, W3A16, W3A3, and W4A4, TesseraQ consistently exhibits superior performance.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (40)
  1. Quarot: Outlier-free 4-bit inference in rotated llms. arXiv preprint arXiv:2404.00456, 2024.
  2. Piqa: Reasoning about physical commonsense in natural language. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pp.  7432–7439, 2020.
  3. Sparks of artificial general intelligence: Early experiments with gpt-4. arXiv preprint arXiv:2303.12712, 2023.
  4. Quip: 2-bit quantization of large language models with guarantees. arXiv preprint arXiv:2307.13304, 2023.
  5. Optimize weight rounding via signed gradient descent for the quantization of llms. arXiv preprint arXiv:2309.05516, 2023.
  6. Think you have solved question answering? try arc, the ai2 reasoning challenge. arXiv preprint arXiv:1803.05457, 2018.
  7. Llm. int8 (): 8-bit matrix multiplication for transformers at scale. arXiv preprint arXiv:2208.07339, 2022.
  8. Qlora: Efficient finetuning of quantized llms. arXiv preprint arXiv:2305.14314, 2023.
  9. Extreme compression of large language models via additive quantization. arXiv preprint arXiv:2401.06118, 2024.
  10. Gptq: Accurate post-training quantization for generative pre-trained transformers. arXiv preprint arXiv:2210.17323, 2022.
  11. A survey of quantization methods for efficient neural network inference. In Low-Power Computer Vision, pp.  291–326. Chapman and Hall/CRC, 2022.
  12. Llm-qbench: A benchmark towards the best practice for post-training quantization of large language models. arXiv preprint arXiv:2405.06001, 2024.
  13. Billm: Pushing the limit of post-training quantization for llms. arXiv preprint arXiv:2402.04291, 2024a.
  14. Slim-llm: Salience-driven mixed-precision quantization for large language models. arXiv preprint arXiv:2405.14917, 2024b.
  15. An empirical study of llama3 quantization: From llms to mllms, 2024c. URL https://arxiv.org/abs/2404.14047.
  16. Improving post training neural quantization: Layer-wise calibration and integer programming. arXiv preprint arXiv:2006.10518, 2020.
  17. Mistral 7b. arXiv preprint arXiv:2310.06825, 2023.
  18. JonathanSalwan. Jonathansalwan/triton: Triton is a dynamic binary analysis library. build your own program analysis tools, automate your reverse engineering, perform software verification or just emulate code. URL https://github.com/jonathansalwan/Triton.
  19. Diederik P Kingma. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2014.
  20. Brecq: Pushing the limit of post-training quantization by block reconstruction. arXiv preprint arXiv:2102.05426, 2021.
  21. Awq: Activation-aware weight quantization for llm compression and acceleration. arXiv preprint arXiv:2306.00978, 2023.
  22. Qllm: Accurate and efficient low-bitwidth quantization for large language models. arXiv preprint arXiv:2310.08041, 2023a.
  23. Llm-qat: Data-free quantization aware training for large language models. arXiv preprint arXiv:2305.17888, 2023b.
  24. Spinquant–llm quantization with learned rotations. arXiv preprint arXiv:2405.16406, 2024.
  25. The era of 1-bit llms: All large language models are in 1.58 bits. arXiv preprint arXiv:2402.17764, 2024.
  26. Pointer sentinel mixture models. arXiv preprint arXiv:1609.07843, 2016.
  27. Meta. Introducing llama 3.1: Our most capable models to date, 2024. URL https://ai.meta.com/blog/meta-llama-3-1/.
  28. Up or down? adaptive rounding for post-training quantization. In International Conference on Machine Learning, pp.  7197–7206. PMLR, 2020.
  29. A white paper on neural network quantization. arXiv preprint arXiv:2106.08295, 2021.
  30. Exploring the limits of transfer learning with a unified text-to-text transformer. The Journal of Machine Learning Research, 21(1):5485–5551, 2020.
  31. Winogrande: An adversarial winograd schema challenge at scale. Communications of the ACM, 64(9):99–106, 2021.
  32. Omniquant: Omnidirectionally calibrated quantization for large language models. arXiv preprint arXiv:2308.13137, 2023.
  33. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971, 2023a.
  34. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288, 2023b.
  35. Gptvq: The blessing of dimensionality for llm quantization. arXiv preprint arXiv:2402.15319, 2024.
  36. Bitnet: Scaling 1-bit transformers for large language models. arXiv preprint arXiv:2310.11453, 2023.
  37. Outlier suppression: Pushing the limit of low-bit transformer language models. Advances in Neural Information Processing Systems, 35:17402–17414, 2022.
  38. Outlier suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling. arXiv preprint arXiv:2304.09145, 2023.
  39. Hellaswag: Can a machine really finish your sentence? arXiv preprint arXiv:1905.07830, 2019.
  40. A survey on efficient inference for large language models. arXiv preprint arXiv:2404.14294, 2024.

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now