{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2511-04063","title":"DartQuant: Efficient Rotational Distribution Calibration for LLM Quantization","arxiv_id":"2511.04063","date":"2025-11-06","proceeding":"NeurIPS","authors":["Yuantian Shao","Yuanteng Chen","Peisong Wang","Jianlin Yu","Jing Lin","Yiwu Yao","Zhihui Wei","Jian Cheng"],"abstract":"Quantization plays a crucial role in accelerating the inference of large-scale models, and rotational matrices have been shown to effectively improve quantization performance by smoothing outliers. However, end-to-end fine-tuning of rotational optimization algorithms incurs high computational costs and is prone to overfitting. To address this challenge, we propose an efficient distribution-aware rotational calibration method, DartQuant, which reduces the complexity of rotational optimization by constraining the distribution of the activations after rotation. This approach also effectively reduces reliance on task-specific losses, thereby mitigating the risk of overfitting. Additionally, we introduce the QR-Orth optimization scheme, which replaces expensive alternating optimization with a more efficient solution. In a variety of model quantization experiments, DartQuant demonstrates superior performance. Compared to existing methods, it achieves 47$\\times$ acceleration and 10$\\times$ memory savings for rotational optimization on a 70B model. Furthermore, it is the first to successfully complete rotational calibration for a 70B model on a single 3090 GPU, making quantization of large language models feasible in resource-constrained environments. Code is available at https://github.com/CAS-CLab/DartQuant.git.","url_abs":"https://arxiv.org/abs/2511.04063","url_pdf":"https://arxiv.org/pdf/2511.04063","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2511.04063","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2511.04063"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/CAS-CLab/DartQuant","reach":{"status":"ok"}}],"summary":{"ran":3,"ran_honours":1,"unverified":12},"by_repo_kind":{"found_in_text":{"samples":16,"ran":4,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":16,"samples":[{"code_sha256_prefix":"e649cf99e1985c23","entry":"add_wrapper_after_function_call_in_method","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/monkeypatch.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/monkeypatch.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e649cf99e1985c23"}},{"code_sha256_prefix":"0225f1f47f1347ae","entry":"asym_dequant","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/quant_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/quant_utils.py","link_basis":"plan_row","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0225f1f47f1347ae"}},{"code_sha256_prefix":"76b345bcf88c8668","entry":"copy_func_with_new_globals","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/monkeypatch.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/monkeypatch.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"76b345bcf88c8668"}},{"code_sha256_prefix":"8005fc23ccbe61fb","entry":"get_rope_function_name","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/model_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/model_utils.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8005fc23ccbe61fb"}},{"code_sha256_prefix":"0979b095b9a8e5d5","entry":"asym_quant","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/quant_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/quant_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0979b095b9a8e5d5"}},{"code_sha256_prefix":"9b105523e5cc293f","entry":"get_c4_new","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/data_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/data_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"9b105523e5cc293f"}},{"code_sha256_prefix":"eb59d0a73734e449","entry":"get_layers","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/model_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/model_utils.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"eb59d0a73734e449"}},{"code_sha256_prefix":"56fe9b504cdc9689","entry":"get_minq_maxq","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/quant_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/quant_utils.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"56fe9b504cdc9689"}},{"code_sha256_prefix":"738315ace19a8a3e","entry":"get_mutil_head_scale","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/calibrater/get_train_data.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/calibrater/get_train_data.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"738315ace19a8a3e"}},{"code_sha256_prefix":"b8790746f232537a","entry":"get_per_channel_scale","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/calibrater/get_train_data.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/calibrater/get_train_data.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b8790746f232537a"}},{"code_sha256_prefix":"3af6f58d0be0329f","entry":"get_ptb_new","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/data_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/data_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3af6f58d0be0329f"}},{"code_sha256_prefix":"8715ffbc2b65ebe5","entry":"get_wikitext2","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/data_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/data_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8715ffbc2b65ebe5"}},{"code_sha256_prefix":"8c91192eb1f9602e","entry":"llama_2_70b_forward","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/eval_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/eval_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"8c91192eb1f9602e"}},{"code_sha256_prefix":"cdd805f7fc7a0842","entry":"model_type_extractor","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/model_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/model_utils.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"cdd805f7fc7a0842"}},{"code_sha256_prefix":"60abacfad210e6fb","entry":"optimize_npu","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/calibrater/r1_base_qr.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/calibrater/r1_base_qr.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"60abacfad210e6fb"}},{"code_sha256_prefix":"9b07900fe82b07cb","entry":"ppl_evaluator","repo":"CAS-CLab/DartQuant","repo_kind":"found_in_text","path":"NPU_DartQuant/fake_quant/eval_utils.py","file_url":"https://github.com/CAS-CLab/DartQuant/blob/HEAD/NPU_DartQuant/fake_quant/eval_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"9b07900fe82b07cb"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_api"},"syntology_extracted_results":null}