Pith. sign in

REVIEW 3 major objections 6 minor 27 references

Approximate Attention Weighting for Sustainable FPGA-Based Vision Transformer Inference

T0 review · 3 major / 6 minor · reviewed 2026-07-12 · grok-4.5

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

Pith's one-line read A 16-segment natural-exponential table on LUT fabric alone can run a full ViT attention row on a small FPGA with under 0.2% top-1 loss and no BRAM.

desk verdict Solid Zynq-7020 attention-row core with real post-route/SAIF numbers; the natural-exp PWL move is real but incremental, and the accuracy half is thinner than the hardware half. read the letter →

arxiv 2607.01798 v2 pith:TQP43KE2 submitted 2026-07-02 cs.AR

Muhammad Usman , Muhammad Akmal Shafique , Shujaat Khan , Dorit Merhof This is my paper · ORCID
classification cs.AR
keywords VisionTransformerFPGAsoftmaxapproximationpiecewise-linearexponentialBRAM-freeattentionedgeAIZynq-7020sustainableinference
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

Vision Transformers are useful for edge visual monitoring, but their softmax stage is expensive on small FPGAs because it needs exp and normalization, often via CORDIC or BRAM tables. This paper shows that a fixed 16-segment piecewise-linear approximation of the natural exponential e^x over the max-centered interval [-8,0], stored entirely in distributed LUT fabric, can replace that stage without changing the pre-trained attention temperature. On a Zynq-7020 the complete 197-token attention-row core uses 1,444 LUTs, 77 DSPs and zero BRAM, draws 21 mW dynamic power at 100 MHz, and stays within 0.20% absolute top-1 of exact softmax on ViT-S/B/L under an INT16 matched setting. The design therefore removes a memory and power bottleneck that has limited sustainable ViT deployment on constrained edge platforms.

What carries the argument

The 16-segment uniform PWL table for e^x on [-8,0]: seventeen 16-bit endpoints (272 bits of LUTRAM) that evaluate the natural exponential by linear interpolation inside each half-unit segment, preserving positivity, monotonicity and the original attention temperature.

What would settle it

Replace the exact softmax inside a full ViT-B/16 or ViT-L/16 pipeline with the identical 16-segment natural PWL table, evaluate top-1 on the full ImageNet-1K validation set (or a real photovoltaic-inspection dataset) under the same INT16 protocol, and check whether the absolute accuracy gap stays ≤ 0.20% with no temperature recalibration.

Watch

Extended reading notes

Core claim

A complete attention-row arithmetic core that replaces exact e^{u_j} by a 16-segment uniform piecewise-linear approximation of the natural exponential over the clipped max-centered domain [-8,0] can be implemented with only distributed LUT fabric (zero BRAM), meets 100 MHz timing on a Zynq-7020 at 1,444 LUTs / 77 DSPs / 21 mW dynamic, and keeps absolute top-1 accuracy change ≤ 0.20% versus exact softmax on ViT-family models without any model-specific temperature recalibration.

Load-bearing premise

That clipping max-centered attention scores to [-8,0] and swapping in a fixed 16-segment natural-exp table (maximum absolute error 0.0245) is accurate enough for real edge ViT workloads without fine-tuning or temperature recalibration, even though accuracy is shown only on the small Imagenette split.

Editorial extensions

If this is right

  • Small SoC FPGAs can host a complete 197-token ViT attention row without any BRAM budget for the exponential.
  • Pre-trained ViT weights can be used directly; base-2 conversion and model-specific temperature calibration are unnecessary.
  • Dynamic energy of roughly 1.66 µJ per attention row (601 krows/s/W) becomes available as a building block for multi-node edge monitoring.
  • An illustrative 500-node continuous-monitoring network yields an arithmetic-core energy gap of about 43 MWh/year versus a 10 W embedded-GPU reference.

Reading between the lines

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

  • Because the weight unit itself is only 71 LUTs, the same natural-exp table can be tiled or shared across multiple heads without exhausting the remaining fabric on a Zynq-class device.
  • The same clipped natural-exp PWL primitive could be reused for other Transformer nonlinearities (e.g., GELU tails) that are also evaluated on a compact negative domain.
  • If the 0.20% Imagenette gap holds on larger sets, the design removes the usual trade-off between BRAM pressure and accuracy recovery that forces many edge-ViT accelerators into quantization-aware retraining.
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

3 major / 6 minor

Summary. The paper proposes a BRAM-free approximate attention-weighting unit for FPGA-based Vision Transformer inference. It replaces the natural exponential in max-centered softmax with a 16-segment uniform piecewise-linear approximation of e^x over [−8,0], stored as 17 endpoint values (272 bits) in distributed LUTRAM. The design is implemented as a complete 197-token attention-row core (score MAC, PWL weight, numerator/denominator accumulation, restoring divider) on Xilinx Zynq-7020, reporting 1,444 LUTs, 77 DSPs, 0 BRAM, WNS +1.50 ns at 100 MHz, 21 mW SAIF dynamic / 124 mW total on-chip power, and 1.66 µJ dynamic energy per row. Hardware-accurate emulation on Imagenette (ViT-S/B/L, INT16 matched protocol; also DeiT-S INT8) shows absolute top-1 change ≤0.20% versus exact softmax without model-specific temperature recalibration. An illustrative 500-node energy scenario is used to motivate sustainability for edge monitoring.

Significance. If the hardware and accuracy claims hold under broader evaluation, the work is a useful, concrete contribution to resource-constrained FPGA attention: a zero-BRAM natural-exp PWL weight unit that avoids base-2 temperature rescaling and CORDIC/BRAM tables, with post-route utilization, timing, and SAIF power numbers that are specific and reproducible in principle. Strengths include the explicit error-vs-segment trade-off (Fig. 2), module-level resource/power breakdown (Table 1), and the decision to approximate e^x rather than 2^x so that pre-trained attention scale is preserved without recalibration. The sustainability framing is secondary and illustrative; the primary value is a compact, BRAM-free attention-row building block for small SoC FPGAs.

major comments (3)
  1. §5.1 / Tables 2–3: The load-bearing claim that the fixed 16-segment natural PWL (max |error| 0.0245 near 0; §3.2, Fig. 1) preserves pre-trained attention semantics without recalibration rests only on hardware-accurate software emulation on the 3,550-image Imagenette split. Imagenette is a 10-class easy subset; the paper itself notes Li et al. report larger INT16 deltas on ImageNet-1K and cautions against direct ranking. Without at least one full ImageNet-1K (or application-domain) evaluation under the same matched INT16 protocol, or a distributional analysis of max-centered score mass near the high-curvature / clip boundary, the ≤0.20% top-1 claim does not yet support the “no recalibration / edge deployment” framing.
  2. §3.1–3.2 and Eq. (1): Clipping max-centered scores to [−8,0] is justified as “negligible” contribution, but the manuscript does not report the fraction of attention mass (or rank changes) that falls near −8 or in the highest-curvature segment under real ViT attention maps. Because the approximation error is largest near 0 and the design saturates below −8, a short empirical check of score histograms / mass outside the interval on the evaluated models would make the domain choice load-bearing rather than assumed.
  3. §5.4 / Table 5: The 43.26 MWh/year gap multiplies the 124 mW on-chip arithmetic-kernel power by 500 nodes against a nominal 10 W embedded-GPU module. The text correctly labels this illustrative and excludes memory, DMA, sensors, and the rest of the Transformer, but the abstract and conclusion still use it to motivate “sustainable edge-AI.” Either demote the scenario to a pure power-budget headroom note or add a full-system board-level power bound so the sustainability claim is not carried by the kernel alone.
minor comments (6)
  1. Keywords list “Left-to-right arithmetic, FPGA, adder tree, ultrasound beamforming, dynamic precision, energy efficiency,” which does not match the paper content (ViT attention / PWL softmax). Replace with topic-appropriate keywords.
  2. Table 1: “Natural PWL Weight 71 203” appears to be a formatting/column-alignment error (LUT/DSP/BRAM/Power). Clarify the intended DSP and power entries for that row.
  3. §4.1: Per-head latency is given as 3,782 cycles (score-plus-weight) while full row is 7,920 cycles including two passes and division; a one-line schedule diagram or cycle breakdown would make the two-pass accounting easier to verify.
  4. Table 4 comparison mixes standalone softmax blocks with full attention-row cores; the text already notes the boundary, but a clearer “scope” column or separate sub-tables would reduce apples-to-oranges reading.
  5. Abstract and §1 claim “within a 0.20% absolute top-1 difference … on ViT-family models”; state the dataset (Imagenette) and that results are from hardware-accurate emulation, not on-device end-to-end inference.
  6. Reference list and related-work framing are generally appropriate; ensure consistent naming of base-2 vs natural-exp temperature effects when citing Li et al. and Hirayae et al.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: empirical FPGA design with fixed PWL table and external accuracy measurement, not a derivation that reduces to its inputs.

full rationale

This paper is a hardware implementation study, not a first-principles derivation whose conclusions are forced by construction. The 16-segment uniform PWL of e^x on [−8,0] uses stored boundary values y_i = e^{b_i} (Eqs. 5–6); segment count S=16 and clip range are design choices justified by an error-vs-table-size trade-off (Fig. 2, max |error| 0.0245), not parameters fitted to recover the reported Imagenette top-1 numbers. Accuracy is measured after the fact by hardware-accurate emulation against an exact-softmax reference on held-out images with unmodified pre-trained weights (Tables 2–3), so the ≤0.20% top-1 deltas are external empirical outcomes, not tautologies. Resource/power figures come from post-route Vivado synthesis and SAIF traces. The natural-exp vs base-2 temperature argument is a standard algebraic observation (2^x = e^{x ln 2}), not a self-citation uniqueness claim. No load-bearing self-citation chain, no fitted-input-called-prediction, and no renaming of a known result as a new derivation. Weaknesses (Imagenette-only eval, illustrative 500-node energy scenario) are generalization/scope issues, not circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 1 invented entities

The central claim rests on standard stable-softmax algebra plus a small set of engineering design choices (clip interval, segment count, fixed-point formats, two-pass schedule) rather than new physical entities. Free parameters are discrete hardware knobs chosen by error/area trade-off, not continuous fits to recover accuracy. No invented particles or forces; the ‘entity’ is an IP block defined by its RTL behavior.

free parameters (4)
  • PWL segment count S = 16
    Chosen as S=16 after plotting max/mean absolute error vs S over [−8,0]; balances 0.0245 max error against 272-bit LUTRAM footprint. Directly determines approximation quality and table size.
  • Exponential clip interval = [-8, 0]
    Inputs below −8 saturated to −8; domain restricted to [−8,0] after max-centering. Hand-chosen compact range where exp contribution is argued to be negligible.
  • Fixed-point formats (Q8.8 scores, 16-bit boundary values) = Q8.8 / 16-bit endpoints
    Representation choices that affect quantization error and hardware cost; not derived from a uniqueness theorem.
  • Illustrative deployment constants (500 nodes, 10 W GPU reference, 0.41 kgCO2/kWh) = 500 nodes; 10 W; 0.41 kgCO2/kWh
    Scenario parameters used to produce the 43.26 MWh/year and 17.74 tCO2e/year gap; not measured system quantities.
assumptions (5)
  • standard math Max-centered softmax is numerically stable and equivalent for attention weights: u_j = s_j - max s, p_j = exp(u_j)/sum exp(u_k).
    Invoked from Blanchard et al. and used throughout §2.1 and Eq. (1)–(4) to justify restricting the approximation domain to non-positive inputs.
  • domain assumption Approximating natural exp (not base-2) preserves the pre-trained attention temperature without model-specific recalibration.
    Core design thesis in abstract and §1/§3.1; supported only by Imagenette INT8/INT16 emulation, not by a general proof across all ViT checkpoints and datasets.
  • domain assumption For practical ViT attention rows, scores more negative than −8 contribute negligibly after exp, so hard saturation is acceptable.
    Stated in §1 and §3.1–3.2 as justification for the compact approximation interval.
  • domain assumption Hardware-accurate software emulation of the RTL datapath is a faithful proxy for on-chip numerical behavior for top-1 accuracy reporting.
    Accuracy tables in §5.1 are from emulation, not board-measured end-to-end ViT inference.
  • ad hoc to paper Comparing a 124 mW FPGA arithmetic kernel to a nominal 10 W embedded-GPU module usefully indicates sustainability headroom for edge monitoring fleets.
    §5.4 scenario explicitly excludes sensors, memory, DMA, and remaining Transformer layers; used to motivate the work rather than as a measured system result.
invented entities (1)
  • Natural PWL attention-weighting unit (16-segment LUTRAM exp table + surrounding attention-row core)
    purpose: Replace CORDIC/BRAM softmax weighting with a zero-BRAM fabric implementation that keeps natural-exp softmax form.
    Defined entirely by the paper’s RTL architecture and coefficient table; independent evidence would be third-party reimplementation or board-level full ViT measurements, which are not provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Approximate Attention Weighting for Sustainable FPGA-Based Vision Transformer Inference." pith.science (2026). https://pith.science/paper/TQP43KE2

@misc{pith2026260701798,
  author       = {Pith},
  title        = {Pith review of: Approximate Attention Weighting for Sustainable FPGA-Based Vision Transformer Inference},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TQP43KE2}},
  note         = {Machine review of arXiv:2607.01798}
}
abstract

Vision Transformers have reshaped computer vision by using self-attention to capture global context across image regions. This makes them attractive for edge visual inspection and monitoring in applications such as renewable-energy infrastructure, industrial quality control, medical imaging, and autonomous-system sensing. However, deploying ViTs on small FPGAs remains challenging because the softmax stage in self-attention requires exponential evaluation and normalization, which are costly in hardware. Existing implementations often rely on CORDIC pipelines or BRAM-based look-up tables, increasing area and power consumption. This paper presents a BRAM-free approximate attention-weighting unit for FPGA-based ViT inference. The proposed design approximates the natural exponential in softmax using a 16-segment piecewise-linear function implemented entirely with distributed LUT fabric. Unlike base-2 approximations, the natural-exponential formulation preserves the pre-trained attention temperature and avoids model-specific recalibration. Implemented on a Xilinx Zynq-7020, the complete attention-row core uses 1444 LUTs, 77 DSPs, and no BRAM, while hardware-accurate emulation shows accuracy within a \(0.20\%\) absolute top-1 difference from the exact-softmax reference on ViT-family models. These results demonstrate the potential of the proposed core for energy-efficient ViT inference on resource-constrained edge-AI platforms.

Figures

Figures reproduced from arXiv: 2607.01798 by the authors.

Figure 1
Figure 1. (a) Natural-exponential weighting kernel [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Maximum and mean absolute PWL error vs. number of uniform segments [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 2
Figure 2. Maximum and mean absolute PWL error vs. number of uniform segments [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figures from the paper (2 more)
Figure 3
Figure 3. Figure 3: Five-module attention-row datapath for the conventional DSP-MAC design. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png]
Figure 4
Figure 4. Figure 4: Post-route implementation results on Zynq-7020. (a) Resource utilization by module relative to device [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 1 linked inside Pith

  1. [1]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby, “An image is worth 16x16 words: Transformers for image recognition at scale,” inInternational Conference on Learning Representations (ICLR), 2021

    2021

  2. [2]

    Training data-efficient image trans- formers & distillation through attention,

    H. Touvron, M. Cord, M. Douze, F. Massa, A. Sablayrolles, and H. J ´egou, “Training data-efficient image trans- formers & distillation through attention,” inInternational Conference on Machine Learning (ICML), 2021, pp. 10 347–10 357

    2021

  3. [3]

    Edge intelligence for energy-efficient computation offloading and resource management in IoT-enabled smart grid,

    H. Liao, D. Cao, and L. Nguyen, “Edge intelligence for energy-efficient computation offloading and resource management in IoT-enabled smart grid,”IEEE Transactions on Industrial Informatics, vol. 18, no. 12, pp. 8351– 8362, 2022

    2022

  4. [4]

    Energy and policy considerations for deep learning in NLP,

    E. Strubell, A. Ganesh, and A. McCallum, “Energy and policy considerations for deep learning in NLP,” in Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics (ACL), 2019, pp. 3645–3650

    2019

  5. [5]

    The carbon footprint of machine learning training will plateau, then shrink,

    D. Patterson, J. Gonzalez, U. H ¨olzle, Q. Le, C. Liang, L.-M. Munguia, D. Rothchild, D. R. So, M. Texier, and J. Dean, “The carbon footprint of machine learning training will plateau, then shrink,”Computer, vol. 55, no. 7, pp. 18–28, 2022

    2022

  6. [6]

    Artificial intelligence in sustainable energy industry: Status quo, challenges and opportunities,

    T. Ahmad, D. Zhang, C. Huang, H.-C. Zhang, N. Dai, Y . Song, and H. Chen, “Artificial intelligence in sustainable energy industry: Status quo, challenges and opportunities,”Journal of Cleaner Production, vol. 289, p. 125834, 2021

    2021

  7. [7]

    Pdet: A progressive deformable transformer for photovoltaic panel defect seg- mentation,

    P. Zhou, H. Fang, and G. Wu, “Pdet: A progressive deformable transformer for photovoltaic panel defect seg- mentation,”Sensors, vol. 24, no. 21, p. 6908, 2024

    2024

  8. [8]

    Deep learning model-transformer based wind power forecasting approach,

    S. Huang, C. Yan, and Y . Qu, “Deep learning model-transformer based wind power forecasting approach,”Fron- tiers in Energy Research, vol. 10, p. 1055683, 2023

    2023

Show all 27 references
  1. [9]

    Transformer-based model for electrical load forecasting,

    A. L’Heureux, K. Grolinger, and M. A. M. Capretz, “Transformer-based model for electrical load forecasting,” Energies, vol. 15, no. 14, p. 4993, 2022

    2022

  2. [10]

    Accurately computing the log-sum-exp and softmax functions,

    P. Blanchard, D. J. Higham, and N. J. Higham, “Accurately computing the log-sum-exp and softmax functions,” IMA Journal of Numerical Analysis, vol. 41, no. 4, pp. 2311–2330, 2021

    2021

  3. [11]

    Fq-vit: Post-training quantization for fully quantized vision transformer,

    Y . Lin, T. Zhang, P. Sun, Z. Li, and S. Zhou, “Fq-vit: Post-training quantization for fully quantized vision transformer,”arXiv preprint arXiv:2111.13824, 2021

    2021 arXiv

  4. [12]

    Hardware implementation of multi-rate input softmax activation function,

    M. Wasef and N. Rafla, “Hardware implementation of multi-rate input softmax activation function,” inIEEE International Midwest Symposium on Circuits and Systems, 2021

    2021

  5. [13]

    An empirical evaluation of en- hanced performance softmax function in deep learning,

    S. Mehra, G. Raut, R. Das Purkayastha, S. K. Vishvakarma, and A. Biasizzo, “An empirical evaluation of en- hanced performance softmax function in deep learning,”IEEE Access, vol. 11, pp. 34 912–34 924, 2023

    2023

  6. [14]

    An energy-efficient architecture of approximate softmax functions for Transformer in edge computing,

    S. Li, B. Yin, and H. Zhang, “An energy-efficient architecture of approximate softmax functions for Transformer in edge computing,” inIEEE International Conference on Machine Learning and Applications, 2023, pp. 183– 189

    2023

  7. [15]

    Hardware-oriented and precisely approximated online soft- max for deep learning models,

    S. Hirayae, K. Yoshioka, Y . Tanaka, and H. Tamuko, “Hardware-oriented and precisely approximated online soft- max for deep learning models,” inIEEE International Conference on Electronics, Circuits and Systems (ICECS), 2025

    2025

  8. [16]

    ViTA: A vision transformer inference accelerator for edge applications,

    G. Li, Y . Chen, Z. Li, Z. Chen, Y . Wang, and H. Yang, “ViTA: A vision transformer inference accelerator for edge applications,” inIEEE International Symposium on Circuits and Systems (ISCAS), 2023

    2023

  9. [17]

    A high speed reconfigurable architecture for softmax and GELU in vision transformer,

    T. Li, F. Zhang, G. Xie, X. Fan, Y . Gao, and M. Sun, “A high speed reconfigurable architecture for softmax and GELU in vision transformer,”Electronics Letters, vol. 59, no. 5, p. e12751, 2023

    2023

  10. [18]

    Hyft: A reconfigurable softmax accelerator with hybrid numeric format for both training and inference,

    T. Xia and S. Q. Zhang, “Hyft: A reconfigurable softmax accelerator with hybrid numeric format for both training and inference,” inACM/IEEE International Symposium on Low Power Electronics and Design, 2024

    2024

  11. [19]

    FPGA implementation and analysis on parallel and pipeline approximate softmax for transformer,

    A. Celep, T. Adiono, I. Syafalni, N. Sutisna, N. Ahmadi, and R. Mulyawan, “FPGA implementation and analysis on parallel and pipeline approximate softmax for transformer,” inIEEE Asia Pacific Conference on Circuits and Systems, 2025

    2025

  12. [20]

    I-BERT: Integer-only BERT quantization,

    S. Kim, A. Gholami, Z. Yao, M. W. Mahoney, and K. Keutzer, “I-BERT: Integer-only BERT quantization,” in International Conference on Machine Learning (ICML), 2021, pp. 5506–5518. 9 Approximate Attention Weighting for Sustainable FPGA-Based Vision Transformer InferenceA PREPRINT

    2021

  13. [21]

    ITA: An energy- efficient attention and softmax accelerator for quantized transformers,

    G. Islamoglu, M. Scherer, G. Paulin, T. Fischer, V . J. B. Jung, A. Garofalo, and L. Benini, “ITA: An energy- efficient attention and softmax accelerator for quantized transformers,” inIEEE/ACM International Symposium on Low Power Electronics and Design, 2023

    2023

  14. [22]

    A low power attention and softmax accelerator for large language models inference,

    J.-H. Kim, C.-H. Kim, S.-M. Rho, and K.-S. Chung, “A low power attention and softmax accelerator for large language models inference,” inIEEE International Conference on Big Data and Smart Computing, 2024

    2024

  15. [23]

    Piecewise-linear approximation of self-attention and its accuracy- aware training for area-efficient vision transformer inference accelerator,

    T. Kawamura, Y . Masuda, and T. Ishihara, “Piecewise-linear approximation of self-attention and its accuracy- aware training for area-efficient vision transformer inference accelerator,” inIEEE International Symposium on Quality Electronic Design (ISQED), 2025

    2025

  16. [24]

    Imagenette: A smaller subset of ImageNet,

    J. Howard, “Imagenette: A smaller subset of ImageNet,” https://github.com/fastai/imagenette, 2019

    2019

  17. [25]

    PyTorch image models (timm),

    R. Wightman, “PyTorch image models (timm),” https://github.com/rwightman/pytorch-image-models, 2019, v0.9

    2019

  18. [26]

    Hardware-efficient softmax approximation for self-attention networks,

    N. A. Koca, A. T. Do, and C.-H. Chang, “Hardware-efficient softmax approximation for self-attention networks,” inIEEE International Symposium on Circuits and Systems, 2023

    2023

  19. [27]

    A generalizable low-precision softmax approximation for small-FPGA de- ployment of vision transformers,

    S. Aboagye, L. Zhai, and S. Cui, “A generalizable low-precision softmax approximation for small-FPGA de- ployment of vision transformers,”Electronics, vol. 15, no. 9, p. 1774, 2026. 10

    2026

Pith tools

Reviewed July 12, 2026 · model on record in the stance chip above.