Pith. sign in

REVIEW 5 major objections 5 minor 28 references

BAQ: Efficient Bit Allocation Quantization for Large Language Models

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper claims that allocating bits by the closed-form rule $R^*_{ij}=\frac{1}{2}\log_2(c_{ij}/G)+R_{\text{sum}}/(MN)$ provably minimizes loss under a Hessian-weighted model, and that per-column implementation in GPTQ lowers perplexity…

desk verdict Textbook water-filling cleverly repackaged for LLM quantization, with real empirical promise but a load-bearing approximation that needs validation and a bit-budget mismatch in the main table. read the letter →

arxiv 2506.05664 v1 pith:AALQEICB submitted 2025-06-06 cs.LG cs.CL

Chao Zhang , Li Wang , Samson Lasaulce , Merouane Debbah This is my paper · ORCID
classification cs.LGcs.CL
keywords post-trainingquantizationbitallocationLLMcompressionHessiansensitivityequal-lossprinciplemixed-precisionGPTQconvexoptimization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that the right way to spend a fixed quantization bit budget across a language model's weights is to allocate precision according to each weight's sensitivity, quantified through a Hessian proxy. It models per-weight quantization loss with the standard high-resolution uniform-quantizer distortion formula, turns bit allocation into a convex optimization problem, and solves it in closed form. The optimal solution obeys an equal-loss principle, and the paper packages it as a lightweight algorithm, BAQ, that drops into GPTQ-style pipelines by assigning one bitwidth per column. If the argument is right, it explains why uniform 2-bit quantization fails and gives a principled replacement that improves perplexity at the same bit budget.

What carries the argument

The load-bearing object is the sensitivity coefficient $c_{ij}=\frac{(w^{\max}_{ij}-w^{\min}_{ij})^2}{12\,[H^{-1}_F]_{n_{ij}n_{ij}}}$, which combines the quantizer's step-size range with the diagonal of the inverse Hessian proxy. The paper feeds these coefficients into the convex problem (OP-A), $\min\sum_{i,j}c_{ij}2^{-2R_{ij}}$ subject to $\sum_{i,j}R_{ij}\le R_{\text{sum}}$, whose closed-form solution carries the entire argument: it produces the bit-allocation rule, the equal-loss identity, and the geometric-mean-over-arithmetic-mean loss ratio. The algorithm then approximates this per-weight rule at column granularity by defining column coefficients $C_j=\sum_i c_{ij}$ and setting each column's bitwidth from $C_j2^{-2R_j}=L_{\text{ref}}$.

What would settle it

Take a pretrained OPT layer, quantize each column at 2 bits, and compare the measured contribution to the loss with the model's prediction $C_j2^{-2R_j}$. If the predicted ordering of columns disagrees with the measured losses at low bitwidths, the optimality result no longer explains the reported gains.

Watch

Extended reading notes

Core claim

The central discovery is that, under a Hessian-weighted squared-error loss model, the optimal bitwidth for each weight is not uniform but is $R^*_{ij}=\frac{1}{2}\log_2(c_{ij}/G)+\frac{R_{\text{sum}}}{MN}$, where $c_{ij}$ is a sensitivity coefficient built from the weight range and the inverse-Hessian diagonal, and $G$ is the geometric mean of all $c_{ij}$. This solution makes every term $c_{ij}2^{-2R^*_{ij}}$ equal, which the paper calls the equal-loss principle, and it implies that the ratio of optimal-allocation loss to uniform-allocation loss is the geometric mean divided by the arithmetic mean of the $c_{ij}$ values. The paper therefore claims that the benefit of adaptive bit allocation grows exactly with the dispersion of Hessian-derived sensitivities, and it reports consistent perplexity and accuracy gains when BAQ replaces GPTQ's uniform bitwidths on OPT models from 125M to 30B parameters.

Load-bearing premise

Everything rests on the standard shortcut that each weight's quantization error equals its quantization step squared divided by 12, an approximation the paper asserts but does not separately validate for 2-bit or non-uniform weight distributions.

Editorial extensions

If this is right

  • GPTQ-type pipelines can replace fixed bitwidths with closed-form per-column bitwidths at negligible overhead, improving perplexity without changing the total bit budget.
  • The equal-loss principle gives a diagnostic: layers with a low geometric-to-arithmetic mean ratio of column sensitivities are exactly where adaptive bit allocation buys the most.
  • The gain over uniform allocation is predicted to grow with the spread of Hessian diagonal entries and weight ranges, so models with more heterogeneous sensitivities should benefit more.
  • Because the loss model is written as $c_{ij}2^{-2R_{ij}}$, the allocation rule generalizes to any compression method whose per-component loss can be expressed that way, including pruning or low-rank factor precision.
  • Transformations such as QuIP's incoherence processing homogenize the sensitivity coefficients, which shrinks the benefit of adaptive allocation; BAQ's role is therefore complementary to such preprocessing.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the geometric-to-arithmetic ratio of column sensitivities could be used before quantization as a cheap predictor of whether mixed precision will pay off, a diagnostic the paper illustrates but does not turn into a standalone tool.
  • Editorial inference: the equal-loss property is proved for continuous bitwidths, while the algorithm rounds to integers; a dedicated analysis of rounding tie-breaking rules could show whether near-optimality is preserved and how much precision is lost.
  • Editorial inference: BAQ's gains should concentrate in layers where the Hessian's eigenvalue spread is largest, so per-layer Ratio_C could be used to choose which layers to quantize adaptively and which to leave at uniform precision.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. The paper proposes BAQ, a mixed-precision weight-quantization framework for large language models. It models per-weight quantization loss by combining the GPTQ/OBS Hessian weighting with a high-resolution uniform-quantizer distortion formula (Eq. 2), then solves a convex bit-allocation problem (OP-A) to obtain a closed-form allocation rule (Eq. 6), an equal-loss property (Eq. 7), and a geometric-vs-arithmetic-mean loss ratio (Eq. 9). The practical algorithm assigns a shared bitwidth per column with an adjustable reference loss (Algorithms 1–3) and is demonstrated as a plugin for GPTQ on OPT-125M to OPT-30B and LLaMA2-13B, reporting large perplexity reductions over GPTQ at 2-bit weight-only quantization. Appendices include proofs, LLaMA results, and a study of integration with QuIP-style transformations.

Significance. If the central cost model and the empirical claims hold, BAQ offers a simple, theoretically grounded, and low-overhead way to turn a fixed-bit PTQ method into a mixed-precision one, with a closed-form solution that can be plugged into existing pipelines. The derivation of the allocation rule is a correct application of standard rate-distortion optimization, and the equal-loss interpretation is a useful design principle. The paper also provides a falsifiable diagnostic (Ratio_C) that connects the dispersion of sensitivity coefficients to the expected benefit of bit allocation. However, the current evidence is not conclusive: the cost model is asserted to be valid at 2 bits without dedicated validation, the main experiments compare only against uniform-bit GPTQ without matched bit budgets, and the overhead claim is contradicted by the reported runtime. These are addressable with additional experiments and analysis, so the work is a solid candidate for major revision.

major comments (5)
  1. [Section 3, Eq. (2)] The high-resolution uniform-quantizer distortion formula Δ²/12 is the load-bearing premise for the entire optimization: the closed-form allocation (Eq. 6), the equal-loss property (Eq. 7), and the gain ratio (Eq. 9) all follow from minimizing Σ c_ij 2^{-2R_ij}. The paper asserts that "as all our simulations have shown, this approximation remains relevant even when the bitwidth is intermediate or low" (Section 3), but no simulation, figure, or table validating Eq. (2) at 2 bits is provided. At 2 bits, with weight distributions that are concentrated near zero and heavy-tailed, the asymptotic uniform-density justification of Δ²/12 need not hold. Please add a direct validation: compare the predicted per-column loss Σ_i c_ij 2^{-2R} with the actual quantization loss for representative GPTQ-quantized OPT layers at R=2, and report the distribution of weights or the approximation error.
  2. [Table 1 and Abstract] The comparison is not at the same bit budget: BAQ's reported average bits are 2.05, 2.08, 1.92, 2.07, and 1.95 for the OPT models, whereas GPTQ uses exactly 2.00. For several models BAQ uses more bits on average, so part of the perplexity gain may be due to a larger budget rather than to the allocation rule. Please report results with exactly matched total bits (or per-average-bitwidth curves) and specify what "at the same bitwidth" means in the abstract. In addition, the abstract's "up to 56× lower perplexity" cannot be verified from Table 1 because the table formatting is broken (entries such as "500.729582521" and "33.6452.7170.84" are concatenated without separators). Please fix the table and state which model/dataset gives the 56× figure.
  3. [Section 5, Main results] The main experimental comparison is only against GPTQ with uniform 2-bit quantization. Since BAQ is a mixed-precision method, the reported improvements may largely reflect the transition from fixed-bit to mixed-bit allocation rather than the specific optimality of the BAQ rule. To support the claim that BAQ is broadly effective, please compare against existing mixed-precision or low-bit PTQ baselines under the same average bit budget, such as AWQ, SqueezeLLM, SpQR, and QuIP (the paper already contains QuIP integration results in Appendix C, but those are not in the main table). At minimum, state clearly that the main claim is about improving GPTQ, not about outperforming the current state of the art.
  4. [Section 4.5 and Table 2] The paper states in Section 4.5 that BAQ brings "negligible additional computational complexity" and "minimal overhead," but Table 2 shows BAQ is about 1.54× slower than GPTQ on OPT-2.7B (797.07 s vs 517.08 s) and similar ratios on other models. A 54% slowdown is not negligible in the context of one-shot PTQ. Please either revise the overhead claim to specify that the overhead is small relative to inference or calibration, or provide a more careful accounting of where the extra time comes from (e.g., the per-column GPTQ passes at mixed bitwidths may be less efficient than a single uniform pass).
  5. [Section 4.2 and Algorithm 1] The theoretical optimality is derived for the per-weight relaxed problem (OP-A), but the implemented algorithm allocates a shared bitwidth per column and rounds to integer bitwidths. The paper does not analyze how much loss is incurred by these two restrictions: the column-wise structure and the rounding step. At 2-bit precision, rounding to the nearest integer can change the average bitwidth by up to 0.5 bits per column, which is substantial relative to the budget, and the equal-loss property is broken after rounding. Please provide an analysis or empirical study of the rounding and column-structure losses; otherwise the "provably optimal" claim applies only to the relaxed per-weight problem, not to the algorithm that is actually evaluated.
minor comments (5)
  1. [Eq. (2)] The quantities w_max_ij and w_min_ij are referred to as the "maximum and minimum bounds of the quantizer Q_ij," but it is not specified how these bounds are estimated per weight (e.g., min/max over a calibration set, or per-column range). Please define them precisely.
  2. [Section 4.1] The statement that λ is a normalization factor proportional to the geometric mean of {c_ij} is imprecise; the exact expression is given in Appendix A.1. Use a consistent notation for λ and λ′ to avoid confusion.
  3. [Table 3, Appendix B] The sentence "BAQ reduces WikiText2 perplexity from 254.44 to 54.44" appears to compare the C4 column (253) with the WikiText2 column (54.44) in the table; please align the text with the column headers.
  4. [References] Reference [19] is listed as "Thomas M Cover. Elements of information theory" but the standard citation includes the co-author Joy A. Thomas; please correct the reference.
  5. [Figures] Figures 3 and 4 are mentioned in the text but are not visible in the provided manuscript text (Figure 3 appears to be missing entirely). Please ensure all figures are included and properly referenced in the final version.

Circularity Check

1 steps flagged · score 2.0 of 10

Bit-allocation derivation is self-contained; the only circular element is the layer-wise loss analysis, which restates Eq. (9) by construction rather than independently confirming BAQ.

  1. self definitional [Section 5, 'Layer-wise loss analysis' (Fig. 3), using Eq. (9)]
    "Quantization loss is approximated by P j Cj2−2Rj , where uniform bitwidth yields loss scaling with the arithmetic mean, while optimal allocation achieves scaling with the geometric mean. ... Fig. 3 reveals that layers with low Ratio_C, which indicate diverse sensitivities on weights, show greater improvement under BAQ (lower Ratio_L). This confirms that bit allocation is especially effective in layers with heterogeneous sensitivity."

    The 'confirmation' is not an independent check: Ratio_L is defined as the ratio of BAQ loss to GPTQ loss under the same approximate model Σ_j C_j 2^{-2R_j}, and Eq. (9) already derives that this ratio equals GM/AM of {C_j} (up to bitwidth correction) whenever BAQ is the optimizer and GPTQ is uniform. Hence the observed correlation between low Ratio_C and low Ratio_L is an algebraic consequence of the definitions, not evidence that the allocation principle is correct. Actual perplexity and accuracy results in Table 1 are independent; the layer-wise analysis merely restates the model's optimum.

full rationale

The main optimization chain (Eq. 2 -> OP-A -> Eq. 6 -> Eq. 7 -> Eq. 8 -> Eq. 9) is a self-contained convex-analysis derivation from stated assumptions; it does not fit or rename any evaluation quantity. The high-resolution distortion approximation in Eq. (2) is a standard external result (refs [18],[19]) and is an assumption about the loss, not a circular input. Algorithm 2's L_init is calibrated to a target average bitwidth, not to the reported perplexity or accuracy metrics, so the empirical gains in Table 1 are independent evidence. No load-bearing self-citation chain exists. The only circular element is the layer-wise loss analysis (Fig. 3): Ratio_L is computed from the same approximate loss model whose optimal-vs-uniform ratio Eq. (9) already evaluates as GM/AM; hence the reported correlation between Ratio_L and Ratio_C confirms Eq. (9) by construction rather than providing independent validation. Separately, Table 1 shows BAQ average bits (1.92-2.08) not exactly 2.00, a fairness confound for 'same bitwidth' claims, but this is a correctness or experimental-design concern, not circularity.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new physical or computational entities. It relies on standard rate-distortion mathematics, a Hessian proxy from the prior GPTQ framework, and one calibration parameter L_init that controls the bit budget. The core derivation is self-contained given the cost model, but the cost model itself is an unvalidated low-bitrate approximation.

free parameters (1)
  • L_init (initial reference loss) = Not specified; selected per layer in Algorithm 2
    Used to set L_ref = L_init·2^{2(Rinit-Rref)}; the choice affects the resulting bitwidth distribution, though it is a calibration knob rather than a fit to target performance.
assumptions (4)
  • domain assumption The layer-wise quantization loss is the sum of independent per-weight losses L_ij (Eq. 1), and OP-A is the correct objective to minimize.
    Assumes weights quantize independently and that minimizing the layer-wise Hessian-weighted distortion is the relevant surrogate for final model loss; inherited from GPTQ/OBS.
  • domain assumption The high-resolution uniform quantizer distortion approximation Delta^2/12 (Eq. 2) remains accurate at the low bitwidths used.
    The authors explicitly state 'this approximation remains relevant even when the bitwidth is intermediate or low' and cite image compression, but give no dedicated experiment at 2-bit weight quantization.
  • domain assumption The proxy Hessian H_F = 2 X X^T over calibration samples captures weight sensitivity through [H^{-1}]_nij nij.
    This is the standard GPTQ/OBS assumption; the paper reuses it without new validation.
  • domain assumption Continuous bitwidth relaxation and rounding to integers does not materially degrade the optimality.
    Algorithm 1 rounds R_j to the nearest integer; the paper does not analyze the rounding error.

how reviews work

0 comments
Cite this review

Pith. "Pith review of BAQ: Efficient Bit Allocation Quantization for Large Language Models." pith.science (2026). https://pith.science/paper/AALQEICB

@misc{pith2026250605664,
  author       = {Pith},
  title        = {Pith review of: BAQ: Efficient Bit Allocation Quantization for Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/AALQEICB}},
  note         = {Machine review of arXiv:2506.05664}
}
abstract

Post-training model quantization is a widely adopted technique for reducing the memory and computational costs of large language models (LLMs). However, most existing methods rely on uniform or heuristic bitwidth assignments, failing to account for the nonuniform sensitivity of weights to quantization noise. In this paper, we propose a novel framework for allocating quantization bitwidths based on sensitivity metrics derived from a Hessian proxy. We make key assumptions, which allow the layer/component-wise loss function to be expressed as an explicit function of the bitwidths. This enables a neat formulation of the bit allocation problem as a convex optimization task, whose closed-form solution adapts precision across weights to minimize the layer-wise quantization loss. Inspecting the solution provides several insights (such as the equal-loss structure), which are then exploited to design the proposed \textbf{BAQ} (Bit Allocation Quantization) algorithm. The proposed algorithm achieves a good trade-off between loss minimization and complexity and allows BAQ to be integrated into standard quantization pipelines with minimal overhead. Experimental results show that BAQ consistently outperforms GPTQ, achieving up to 56$\times$ lower perplexity at the same bitwidth on large language models ranging from 125M to 30B parameters. Leveraging our analytical results derived from solving the optimal bit allocation problem, we also provide a theoretical explanation for the observed gains. All codes of this paper are available at https://github.com/CSU-ModelCompression/BAQ.

Figures

Figures reproduced from arXiv: 2506.05664 by the authors.

Figure 1
Figure 1. Comparison between BAQ-2bit and several baselines (GPTQ-2bit, GPTQ-3bit, RTN-16bit) [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Distribution of assigned bitwidths Rj in OPT-2.7B under BAQ. While most weights are quantized to 2 bits, BAQ adaptively allocates higher or lower precision to match weight sensitivity. Layers with higher variance in sensitivity (as measured by {Cj}) exhibit broader bitwidth distribu￾tions, reflecting structural adaptivity. Layer-wise loss analysis [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Layer-wise comparison of quantization efficiency with OPT-2.7B model. [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Histograms of Ratio_C = GM({Cj})/AM({Cj}) across different transformation strate￾gies. A lower value of Ratio_C indicates greater dispersion in the sensitivity coefficients {Cj}, which corresponds to higher potential gains from bit allocation. In mild and moderate tran…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 10 canonical work pages

  1. [1]

    Introducing ChatGPT.OpenAI Blog

    OpenAI. Introducing ChatGPT.OpenAI Blog

    work page

  2. [2]

    Qlora: Efficient finetuning of quantized llms.Advances in neural information processing systems, 36:10088– 10115, 2023

    Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. Qlora: Efficient finetuning of quantized llms.Advances in neural information processing systems, 36:10088– 10115, 2023

    2023

  3. [3]

    OPTQ: Accurate quantization for generative pre-trained transformers

    Elias Frantar, Saleh Ashkboos, Torsten Hoefler, and Dan Alistarh. OPTQ: Accurate quantization for generative pre-trained transformers. InThe Eleventh International Conference on Learning Representations, 2023

    2023

  4. [4]

    Bitnet: Scaling 1-bit transformers for large language models.arXiv preprint arXiv:2310.11453, 2023

    Hongyu Wang, Shuming Ma, Li Dong, Shaohan Huang, Huaijie Wang, Lingxiao Ma, Fan Yang, Ruiping Wang, Yi Wu, and Furu Wei. Bitnet: Scaling 1-bit transformers for large language models.arXiv preprint arXiv:2310.11453, 2023

    arXiv 2023

  5. [5]

    QuIP: 2-bit quantization of large language models with guarantees

    Jerry Chee, Yaohui Cai, V olodymyr Kuleshov, and Christopher De Sa. QuIP: 2-bit quantization of large language models with guarantees. InThirty-seventh Conference on Neural Information Processing Systems, 2023

    work page 2023

  6. [6]

    Smoothquant: Accurate and efficient post-training quantization for large language models

    Guangxuan Xiao, Ji Lin, Mickael Seznec, Hao Wu, Julien Demouth, and Song Han. Smoothquant: Accurate and efficient post-training quantization for large language models. InInternational Conference on Machine Learning, pages 38087–38099. PMLR, 2023

    2023

  7. [7]

    Omniquant: Omnidirectionally calibrated quan- tization for large language models

    Wenqi Shao, Mengzhao Chen, Zhaoyang Zhang, Peng Xu, Lirui Zhao, Zhiqian Li, Kaipeng Zhang, Peng Gao, Yu Qiao, and Ping Luo. Omniquant: Omnidirectionally calibrated quan- tization for large language models. InThe Twelfth International Conference on Learning Representations, 2024

    2024

  8. [8]

    Kvquant: Towards 10 million context length llm inference with kv cache quantization.Advances in Neural Information Processing Systems, 37:1270–1303, 2024

    Coleman Hooper, Sehoon Kim, Hiva Mohammadzadeh, Michael W Mahoney, Sophia Shao, Kurt Keutzer, and Amir Gholami. Kvquant: Towards 10 million context length llm inference with kv cache quantization.Advances in Neural Information Processing Systems, 37:1270–1303, 2024

    2024

Show all 28 references
  1. [9]

    Awq: Activation-aware weight quantization for on-device llm compression and acceleration.Proceedings of Machine Learning and Systems, 6:87–100, 2024

    Ji Lin, Jiaming Tang, Haotian Tang, Shang Yang, Wei-Ming Chen, Wei-Chen Wang, Guangxuan Xiao, Xingyu Dang, Chuang Gan, and Song Han. Awq: Activation-aware weight quantization for on-device llm compression and acceleration.Proceedings of Machine Learning and Systems, 6:87–100, 2024

    2024

  2. [10]

    Owq: Outlier- aware weight quantization for efficient fine-tuning and inference of large language models

    Changhun Lee, Jungyu Jin, Taesu Kim, Hyungjun Kim, and Eunhyeok Park. Owq: Outlier- aware weight quantization for efficient fine-tuning and inference of large language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, pages 13355–13364, 2024

    2024

  3. [11]

    Squeezellm: Dense-and-sparse quantization.arXiv preprint arXiv:2306.07629, 2023

    Sehoon Kim, Coleman Hooper, Amir Gholami, Zhen Dong, Xiuyu Li, Sheng Shen, Michael W Mahoney, and Kurt Keutzer. Squeezellm: Dense-and-sparse quantization.arXiv preprint arXiv:2306.07629, 2023

    2023 arXiv

  4. [12]

    Svirschevski, Vage Egiazarian, Denis Kuznedelev, Elias Frantar, Saleh Ashkboos, Alexander Borzunov, Torsten Hoefler, and Dan Alistarh

    Tim Dettmers, Ruslan A. Svirschevski, Vage Egiazarian, Denis Kuznedelev, Elias Frantar, Saleh Ashkboos, Alexander Borzunov, Torsten Hoefler, and Dan Alistarh. SpQR: A sparse-quantized representation for near-lossless LLM weight compression. InThe Twelfth International Confer- ...

    2024

  5. [13]

    Optimal brain surgeon and general network pruning

    Babak Hassibi, David G Stork, and Gregory J Wolff. Optimal brain surgeon and general network pruning. InIEEE International Conference on Neural Networks, 1993

    1993

  6. [14]

    Optimal brain damage.Advances in neural information processing systems, 2, 1989

    Yann LeCun, John Denker, and Sara Solla. Optimal brain damage.Advances in neural information processing systems, 2, 1989

    1989

  7. [15]

    Sparsegpt: Massive language models can be accurately pruned in one-shot

    Elias Frantar and Dan Alistarh. Sparsegpt: Massive language models can be accurately pruned in one-shot. InInternational Conference on Machine Learning, pages 10323–10337. PMLR, 2023. 11

    2023

  8. [16]

    Optimal brain compression: A framework for accurate post- training quantization and pruning.Advances in Neural Information Processing Systems, 35:4475– 4488, 2022

    Elias Frantar and Dan Alistarh. Optimal brain compression: A framework for accurate post- training quantization and pruning.Advances in Neural Information Processing Systems, 35:4475– 4488, 2022

    2022

  9. [17]

    BiLLM: Pushing the limit of post-training quantization for LLMs

    Wei Huang, Yangdong Liu, Haotong Qin, Ying Li, Shiming Zhang, Xianglong Liu, Michele Magno, and Xiaojuan Qi. BiLLM: Pushing the limit of post-training quantization for LLMs. arXiv preprint arXiv:2402.04291, 2024

    2024 arXiv

  10. [18]

    Gray and David L

    Robert M. Gray and David L. Neuhoff. Quantization.IEEE Transactions on Information Theory, 44(6):2325–2383, 1998

    1998

  11. [19]

    John Wiley & Sons, 1999

    Thomas M Cover.Elements of information theory. John Wiley & Sons, 1999

    1999

  12. [20]

    Jpeg2000: Image compression fundamentals, standards and practice.Journal of Electronic Imaging, 11(2):286–287, 2002

    David S Taubman, Michael W Marcellin, and Majid Rabbani. Jpeg2000: Image compression fundamentals, standards and practice.Journal of Electronic Imaging, 11(2):286–287, 2002

    2002

  13. [21]

    Springer Science & Business Media, 2012

    Allen Gersho and Robert M Gray.Vector quantization and signal compression, volume 159. Springer Science & Business Media, 2012

    2012

  14. [22]

    Diab, Xian Li, Xi Victoria Lin, Todor Mihaylov, Myle Ott, Sam Shleifer, Kurt Shuster, Daniel Simig, Punit Singh Koura, Anjali Sridhar, Tianlu Wang, and Luke Zettlemoyer

    Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen, Christopher Dewan, Mona T. Diab, Xian Li, Xi Victoria Lin, Todor Mihaylov, Myle Ott, Sam Shleifer, Kurt Shuster, Daniel Simig, Punit Singh Koura, Anjali Sridhar, Tianlu Wang, and Luke Zettlemoyer...

    2022 arXiv

  15. [23]

    Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu. Exploring the limits of transfer learning with a unified text-to-text transformer.Journal of Machine Learning Research, 21(140):1–67, 2020

    2020

  16. [24]

    Pointer sentinel mixture models.arXiv preprint arXiv:1609.07843, 2016

    Stephen Merity, Caiming Xiong, James Bradbury, and Richard Socher. Pointer sentinel mixture models.arXiv preprint arXiv:1609.07843, 2016

    2016 arXiv

  17. [25]

    The Penn Treebank: Annotating predicate argument structure

    Mitchell Marcus, Grace Kim, Mary Ann Marcinkiewicz, Robert MacIntyre, Ann Bies, Mark Ferguson, Karen Katz, and Britta Schasberger. The Penn Treebank: Annotating predicate argument structure. InHuman Language Technology: Proceedings of a Workshop held at Plainsboro, New Jersey,...

    1994

  18. [26]

    The LAMBADA dataset: Word prediction requiring a broad discourse context

    Denis Paperno, Germán Kruszewski, Angeliki Lazaridou, Ngoc Quan Pham, Raffaella Bernardi, Sandro Pezzelle, Marco Baroni, Gemma Boleda, and Raquel Fernández. The LAMBADA dataset: Word prediction requiring a broad discourse context. InProceedings of the 54th Annual Meeting of th...

    2016

  19. [27]

    Piqa: An algebra for querying protein data sets

    Sandeep Tata and Jignesh M Patel. Piqa: An algebra for querying protein data sets. In International Conference on Scientific and Statistical Database Management, 2003

    2003

  20. [28]

    A systematic classification of knowl- edge, reasoning, and context within the ARC dataset

    Michael Boratko, Harshit Padigela, Divyendra Mikkilineni, Pritish Yuvraj, Rajarshi Das, An- drew McCallum, Maria Chang, Achille Fokoue-Nkoutche, Pavan Kapanipathi, Nicholas Mattei, Ryan Musa, Kartik Talamadupula, and Michael Witbrock. A systematic classification of knowl- edge...

    2018

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.