Prune-then-Quantize or Quantize-then-Prune? Understanding the Impact of Compression Order in Joint Model Compression

Fuente: arXiv

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	Kim, Minjun, Choi, Jaehyeon, Yang, Hyunwoo, Kim, Jongjin, Song, Jinho, Kang, U
Format:	Preprint
Veröffentlicht:	2026
Schlagworte:	Artificial Intelligence
Online-Zugang:	Acceder al recurso
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

Ähnliche Einträge

SynQ: Accurate Zero-shot Quantization by Synthesis-aware Fine-tuning
von: Kim, Minjun, et al.
Veröffentlicht: (2026)

Automatic Joint Structured Pruning and Quantization for Efficient Neural Network Training and Compression
von: Qu, Xiaoyi, et al.
Veröffentlicht: (2025)

Prune-Quantize-Distill: An Ordered Pipeline for Efficient Neural Network Compression
von: Zhou, Longsheng, et al.
Veröffentlicht: (2026)

The Impact of Quantization and Pruning on Deep Reinforcement Learning Models
von: Lu, Heng, et al.
Veröffentlicht: (2024)

A Comprehensive Survey of Compression Algorithms for Language Models
von: Park, Seungcheol, et al.
Veröffentlicht: (2024)

Locality-Aware Redundancy Pruning for LLM Depth Compression
von: Yun, Vincent-Daniel, et al.
Veröffentlicht: (2026)

Zero-shot Quantization: A Comprehensive Survey
von: Kim, Minjun, et al.
Veröffentlicht: (2025)

Focus on the Core: Efficient Attention via Pruned Token Compression for Document Classification
von: Yun, Jungmin, et al.
Veröffentlicht: (2024)

Singular Value Scaling: Efficient Generative Model Compression via Pruned Weights Refinement
von: Kim, Hyeonjin, et al.
Veröffentlicht: (2024)

DYCP: Dynamic Context Pruning for Long-Form Dialogue with LLMs
von: Choi, Nayoung, et al.
Veröffentlicht: (2026)

Model Compression using Progressive Channel Pruning
von: Guo, Jinyang, et al.
Veröffentlicht: (2025)

Quantization-Aware and Tensor-Compressed Training of Transformers for Natural Language Understanding
von: Yang, Zi, et al.
Veröffentlicht: (2023)

On the Compressibility of Quantized Large Language Models
von: Mao, Yu, et al.
Veröffentlicht: (2024)

LampQ: Towards Accurate Layer-wise Mixed Precision Quantization for Vision Transformers
von: Kim, Minjun, et al.
Veröffentlicht: (2025)

Soft Label Pruning and Quantization for Large-Scale Dataset Distillation
von: Lingao, Xiao, et al.
Veröffentlicht: (2026)

FibQuant: Universal Vector Quantization for Random-Access KV-Cache Compression
von: Lee, Namyoon, et al.
Veröffentlicht: (2026)

Large Multimodal Model Compression via Efficient Pruning and Distillation at AntGroup
von: Wang, Maolin, et al.
Veröffentlicht: (2023)

Binary Quadratic Quantization: Beyond First-Order Quantization for Real-Valued Matrix Compression
von: Kuroki, Kyo, et al.
Veröffentlicht: (2025)

Frequency Matters: Fast Model-Agnostic Data Curation for Pruning and Quantization
von: Monaco, Francesco Pio, et al.
Veröffentlicht: (2026)

One Shot vs. Iterative: Rethinking Pruning Strategies for Model Compression
von: Janusz, Mikołaj, et al.
Veröffentlicht: (2025)

GETA-3DGS: Automatic Joint Structured Pruning and Quantization for 3D Gaussian Splatting
von: Zhang, Baobing, et al.
Veröffentlicht: (2026)

QAPruner: Quantization-Aware Vision Token Pruning for Multimodal Large Language Models
von: Wang, Xinhao, et al.
Veröffentlicht: (2026)

Rotation Invariant Quantization for Model Compression
von: Kampeas, Joseph, et al.
Veröffentlicht: (2023)

SQS: Bayesian DNN Compression through Sparse Quantized Sub-distributions
von: Wang, Ziyi, et al.
Veröffentlicht: (2025)

EPSD: Early Pruning with Self-Distillation for Efficient Model Compression
von: Chen, Dong, et al.
Veröffentlicht: (2024)

Adaptive Dataset Quantization: A New Direction for Dataset Pruning
von: Yu, Chenyue, et al.
Veröffentlicht: (2025)

Structured Pruning and Quantization for Learned Image Compression
von: Hossain, Md Adnan Faisal, et al.
Veröffentlicht: (2025)

HierarchicalPrune: Position-Aware Compression for Large-Scale Diffusion Models
von: Kwon, Young D., et al.
Veröffentlicht: (2025)

ROSAQ: Rotation-based Saliency-Aware Weight Quantization for Efficiently Compressing Large Language Models
von: Yoon, Junho, et al.
Veröffentlicht: (2025)

DAQ: Delta-Aware Quantization for Post-Training LLM Weight Compression
von: Yu, Xiaoming, et al.
Veröffentlicht: (2026)

PPC-GPT: Federated Task-Specific Compression of Large Language Models via Pruning and Chain-of-Thought Distillation
von: Fan, Tao, et al.
Veröffentlicht: (2025)

Hurwitz Quaternion Multiplicative Quantization for KV Cache Compression
von: Swain, Kabir, et al.
Veröffentlicht: (2026)

No Token Left Behind: Reliable KV Cache Compression via Importance-Aware Mixed Precision Quantization
von: Yang, June Yong, et al.
Veröffentlicht: (2024)

SwiftPrune: Hessian-Free Weight Pruning for Large Language Models
von: Kang, Yuhan, et al.
Veröffentlicht: (2025)

CommVQ: Commutative Vector Quantization for KV Cache Compression
von: Li, Junyan, et al.
Veröffentlicht: (2025)

RAP: KV-Cache Compression via RoPE-Aligned Pruning
von: Xin, Jihao, et al.
Veröffentlicht: (2026)

LLMs can Compress LLMs: Adaptive Pruning by Agents
von: Kodathala, Sai Varun, et al.
Veröffentlicht: (2026)

A Free Lunch in LLM Compression: Revisiting Retraining after Pruning
von: Wagner, Moritz, et al.
Veröffentlicht: (2025)

Sensitivity-Guided Framework for Pruned and Quantized Reservoir Computing Accelerators
von: Jafari, Atousa, et al.
Veröffentlicht: (2026)

Multi-View Node Pruning for Accurate Graph Representation
von: Kim, Hanjin, et al.
Veröffentlicht: (2025)

if (filemtime($filename) + $maxlifetime < time()) { unlink($filename); $count++; } } return $count; } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ protected function saveSession($sessId, $data): bool { $sessFile = $this->path . '/sess_' . $sessId; if ($handle = fopen($sessFile, 'w')) { $return = false; // Lock the file for exclusive access to avoid issues with multiple // processes writing session simultaneously: if (flock($handle, LOCK_EX)) { $return = fwrite($handle, $data); // Make sure that there's no trailing data by truncating the file to // the correct length: ftruncate($handle, strlen($data)); fflush($handle); flock($handle, LOCK_UN); } fclose($handle); if ($return !== false) { return true; } } // If we got this far, something went wrong with the file output... // It is tempting to throw an exception here, but this code is called // outside of the context of exception handling, so all we can do is // echo a message.

// Anecdotal testing Today and Yesterday seems to indicate destroy() // is called by the garbage collector and everything is good. // Something to keep in mind though. return 0; } /** * Write function that is called when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ public function write($sessId, $data): bool { if ($this->writesDisabled) { return true; } return $this->saveSession($sessId, $data); } /** * A function that is called internally when session data is to be saved. * * @param string $sessId The current session ID * @param string $data The session data to write * * @return bool */ abstract protected function saveSession($sessId, $data): bool; }

* * @return void */ public function writeClose() { // The assumption is that we're using PHP's ext/session. // session_write_close() will actually overwrite $_SESSION with an // empty array on completion -- which leads to a mismatch between what // is in the storage object and $_SESSION. To get around this, we // temporarily reset $_SESSION to an array, and then re-link it to // the storage object. // // Additionally, while you _can_ write to $_SESSION following a // session_write_close() operation, no changes made to it will be // flushed to the session handler. As such, we now mark the storage // object isImmutable. $storage = $this->getStorage(); if (! $storage->isImmutable()) { $_SESSION = $storage->toArray(true); session_write_close(); $storage->fromArray($_SESSION); $storage->markImmutable(); } } /** * Attempt to set the session name * * If the session has already been started, or if the name provided fails * validation, an exception will be raised. * * @param string $name * @return SessionManager * @throws Exception\InvalidArgumentException */ public function setName($name) { if ($this->sessionExists()) { throw new Exception\InvalidArgumentException( 'Cannot set session name after a session has already started'

Environment & details:

GET Data empty

POST Data empty

Files empty

Key	Value
ui	"standard"
VUFIND_SESSION	"483cdd3c7beb746b0e55297779baedcc"
language	"de"

Session

Key	Value
__Laminas	array:2 [`"_REQUEST_ACCESS_TIME" => 1790979698.2875 "_VALID" => array:1 [ "Laminas\Session\Validator\Id" => "483cdd3c7beb746b0e55297779baedcc" ]`]
SessionState	Laminas\Stdlib\ArrayObject {#938}
FlashMessenger	Laminas\Stdlib\ArrayObject {#1433}
SessionHelper	Laminas\Stdlib\ArrayObject {#1759}

Server/Request Data

Key	Value
REDIRECT_VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
REDIRECT_HTTPS	"on"
REDIRECT_SSL_TLS_SNI	"info.umar.mx"
REDIRECT_STATUS	"200"
VUFIND_LOCAL_DIR	"/usr/local/vufind/local"
HTTPS	"on"
SSL_TLS_SNI	"info.umar.mx"
HTTP_ACCEPT	"/"
HTTP_USER_AGENT	"Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)"
HTTP_COOKIE	"ui=standard; VUFIND_SESSION=483cdd3c7beb746b0e55297779baedcc; language=de"
HTTP_ACCEPT_ENCODING	"gzip, br, zstd, deflate"
HTTP_HOST	"info.umar.mx"
PATH	"/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/snap/bin"
SERVER_SIGNATURE	"<address>Apache/2.4.66 (Ubuntu) Server at info.umar.mx Port 443</address>\n"
SERVER_SOFTWARE	"Apache/2.4.66 (Ubuntu)"
SERVER_NAME	"info.umar.mx"
SERVER_ADDR	"200.23.222.19"
SERVER_PORT	"443"
REMOTE_ADDR	"216.73.217.142"
DOCUMENT_ROOT	"/usr/local/vufind/public"
REQUEST_SCHEME	"https"
CONTEXT_PREFIX	""
CONTEXT_DOCUMENT_ROOT	"/usr/local/vufind/public"
SERVER_ADMIN	"admin@localhost"
SCRIPT_FILENAME	"/usr/local/vufind/public/index.php"
REMOTE_PORT	"14391"
REDIRECT_URL	""" /Record/\n arxiv_https___arxiv_org_abs_2603_18426/Similar """
GATEWAY_INTERFACE	"CGI/1.1"
SERVER_PROTOCOL	"HTTP/1.1"
REQUEST_METHOD	"GET"
QUERY_STRING	""
REQUEST_URI	"/Record/%0Aarxiv_https___arxiv_org_abs_2603_18426/Similar"
SCRIPT_NAME	"/index.php"
PHP_SELF	"/index.php"
REQUEST_TIME_FLOAT	1790979698.2704
REQUEST_TIME	1790979698

Environment Variables empty

Registered Handlers

0. Whoops\Handler\PrettyPageHandler