{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ABD62YPSKO6IOLOA7XYU4DSG7R","short_pith_number":"pith:ABD62YPS","schema_version":"1.0","canonical_sha256":"0047ed61f253bc872dc0fdf14e0e46fc693e47b8069afe20cddb8a2cea5fdc5e","source":{"kind":"arxiv","id":"2505.02390","version":2},"attestation_state":"computed","paper":{"title":"Quantitative Analysis of Performance Drop in DeepSeek Model Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Enbo Zhao, Jian Zhang, Jieyun Huang, Kai Wang, Ning Wang, Shiguo Lian, Shuming Shi, Siqi Xiao, Yi Shen, Zhihao Chen","submitted_at":"2025-05-05T06:25:20Z","abstract_excerpt":"Recently, there is a high demand for deploying DeepSeek-R1 and V3 locally, possibly because the official service often suffers from being busy and some organizations have data privacy concerns. While single-machine deployment offers infrastructure simplicity, the models' 671B FP8 parameter configuration exceeds the practical memory limits of a standard 8-GPU machine. Quantization is a widely used technique that helps reduce model memory consumption. However, it is unclear what the performance of DeepSeek-R1 and V3 will be after being quantized. This technical report presents the first quantita"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2505.02390","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-05T06:25:20Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"18130539ee1c40babfe4b4eab200cf89922eafbd2e3b154602c1076f60577934","abstract_canon_sha256":"e3713431e19af9c735ee4a7744fd84278f2bd7e0a5894dffcfca7234657e5357"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:20:53.752996Z","signature_b64":"DEvXMk1yQNo5YKbVN6n5y+DfG0dQAufKVXgrVCAnaW5vUXugicQA8CZ5hin27wNZCXWpL4+uF3hEQho2o9sZCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0047ed61f253bc872dc0fdf14e0e46fc693e47b8069afe20cddb8a2cea5fdc5e","last_reissued_at":"2026-07-05T11:20:53.752501Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:20:53.752501Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Quantitative Analysis of Performance Drop in DeepSeek Model Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Enbo Zhao, Jian Zhang, Jieyun Huang, Kai Wang, Ning Wang, Shiguo Lian, Shuming Shi, Siqi Xiao, Yi Shen, Zhihao Chen","submitted_at":"2025-05-05T06:25:20Z","abstract_excerpt":"Recently, there is a high demand for deploying DeepSeek-R1 and V3 locally, possibly because the official service often suffers from being busy and some organizations have data privacy concerns. While single-machine deployment offers infrastructure simplicity, the models' 671B FP8 parameter configuration exceeds the practical memory limits of a standard 8-GPU machine. Quantization is a widely used technique that helps reduce model memory consumption. However, it is unclear what the performance of DeepSeek-R1 and V3 will be after being quantized. This technical report presents the first quantita"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.02390","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2505.02390/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2505.02390","created_at":"2026-07-05T11:20:53.752561+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.02390v2","created_at":"2026-07-05T11:20:53.752561+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.02390","created_at":"2026-07-05T11:20:53.752561+00:00"},{"alias_kind":"pith_short_12","alias_value":"ABD62YPSKO6I","created_at":"2026-07-05T11:20:53.752561+00:00"},{"alias_kind":"pith_short_16","alias_value":"ABD62YPSKO6IOLOA","created_at":"2026-07-05T11:20:53.752561+00:00"},{"alias_kind":"pith_short_8","alias_value":"ABD62YPS","created_at":"2026-07-05T11:20:53.752561+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R","json":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R.json","graph_json":"https://pith.science/api/pith-number/ABD62YPSKO6IOLOA7XYU4DSG7R/graph.json","events_json":"https://pith.science/api/pith-number/ABD62YPSKO6IOLOA7XYU4DSG7R/events.json","paper":"https://pith.science/paper/ABD62YPS"},"agent_actions":{"view_html":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R","download_json":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R.json","view_paper":"https://pith.science/paper/ABD62YPS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.02390&json=true","fetch_graph":"https://pith.science/api/pith-number/ABD62YPSKO6IOLOA7XYU4DSG7R/graph.json","fetch_events":"https://pith.science/api/pith-number/ABD62YPSKO6IOLOA7XYU4DSG7R/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R/action/storage_attestation","attest_author":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R/action/author_attestation","sign_citation":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R/action/citation_signature","submit_replication":"https://pith.science/pith/ABD62YPSKO6IOLOA7XYU4DSG7R/action/replication_record"}},"created_at":"2026-07-05T11:20:53.752561+00:00","updated_at":"2026-07-05T11:20:53.752561+00:00"}