{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:5JGNZFZKC4XW4MRZPJD7RKSDBB","short_pith_number":"pith:5JGNZFZK","schema_version":"1.0","canonical_sha256":"ea4cdc972a172f6e32397a47f8aa43084a53a81883069bef0d09c4abefa42f28","source":{"kind":"arxiv","id":"1902.04103","version":3},"attestation_state":"computed","paper":{"title":"Bag of Freebies for Training Object Detection Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hang Zhang, Junyuan Xie, Mu Li, Tong He, Zhi Zhang, Zhongyue Zhang","submitted_at":"2019-02-11T19:20:09Z","abstract_excerpt":"Training heuristics greatly improve various image classification model accuracies~\\cite{he2018bag}. Object detection models, however, have more complex neural network structures and optimization targets. The training strategies and pipelines dramatically vary among different models. In this works, we explore training tweaks that apply to various models including Faster R-CNN and YOLOv3. These tweaks do not change the model architectures, therefore, the inference costs remain the same. Our empirical results demonstrate that, however, these freebies can improve up to 5% absolute precision compar"},"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":"1902.04103","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2019-02-11T19:20:09Z","cross_cats_sorted":[],"title_canon_sha256":"c1d9a8c2988fdbaa52414e19008d4f86af228e67a9b5574ca56cb34f2c0f3d8e","abstract_canon_sha256":"314eca9d0fc99b10fb15d84bd47f14adbc92dc97311c509afe58fdbfa3e440b4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:48:41.630298Z","signature_b64":"8/uubfS/BCmtC1FOrQczrM8GtS6YNrGbJ9n83Di3BQi1hRAGT+ozqHS1U8vCZpAehzI7VlNOzZs55Be5egL8BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ea4cdc972a172f6e32397a47f8aa43084a53a81883069bef0d09c4abefa42f28","last_reissued_at":"2026-05-17T23:48:41.629729Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:48:41.629729Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Bag of Freebies for Training Object Detection Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hang Zhang, Junyuan Xie, Mu Li, Tong He, Zhi Zhang, Zhongyue Zhang","submitted_at":"2019-02-11T19:20:09Z","abstract_excerpt":"Training heuristics greatly improve various image classification model accuracies~\\cite{he2018bag}. Object detection models, however, have more complex neural network structures and optimization targets. The training strategies and pipelines dramatically vary among different models. In this works, we explore training tweaks that apply to various models including Faster R-CNN and YOLOv3. These tweaks do not change the model architectures, therefore, the inference costs remain the same. Our empirical results demonstrate that, however, these freebies can improve up to 5% absolute precision compar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1902.04103","kind":"arxiv","version":3},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"1902.04103","created_at":"2026-05-17T23:48:41.629817+00:00"},{"alias_kind":"arxiv_version","alias_value":"1902.04103v3","created_at":"2026-05-17T23:48:41.629817+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1902.04103","created_at":"2026-05-17T23:48:41.629817+00:00"},{"alias_kind":"pith_short_12","alias_value":"5JGNZFZKC4XW","created_at":"2026-05-18T12:33:10.108867+00:00"},{"alias_kind":"pith_short_16","alias_value":"5JGNZFZKC4XW4MRZ","created_at":"2026-05-18T12:33:10.108867+00:00"},{"alias_kind":"pith_short_8","alias_value":"5JGNZFZK","created_at":"2026-05-18T12:33:10.108867+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"1907.03465","citing_title":"A unified neural network for object detection, multiple object tracking and vehicle re-identification","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2507.06321","citing_title":"Centralized Copy-Paste: Enhanced Data Augmentation Strategy for Wildland Fire Semantic Segmentation","ref_index":36,"is_internal_anchor":true},{"citing_arxiv_id":"2107.08430","citing_title":"YOLOX: Exceeding YOLO Series in 2021","ref_index":38,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB","json":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB.json","graph_json":"https://pith.science/api/pith-number/5JGNZFZKC4XW4MRZPJD7RKSDBB/graph.json","events_json":"https://pith.science/api/pith-number/5JGNZFZKC4XW4MRZPJD7RKSDBB/events.json","paper":"https://pith.science/paper/5JGNZFZK"},"agent_actions":{"view_html":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB","download_json":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB.json","view_paper":"https://pith.science/paper/5JGNZFZK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1902.04103&json=true","fetch_graph":"https://pith.science/api/pith-number/5JGNZFZKC4XW4MRZPJD7RKSDBB/graph.json","fetch_events":"https://pith.science/api/pith-number/5JGNZFZKC4XW4MRZPJD7RKSDBB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB/action/storage_attestation","attest_author":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB/action/author_attestation","sign_citation":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB/action/citation_signature","submit_replication":"https://pith.science/pith/5JGNZFZKC4XW4MRZPJD7RKSDBB/action/replication_record"}},"created_at":"2026-05-17T23:48:41.629817+00:00","updated_at":"2026-05-17T23:48:41.629817+00:00"}