{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:F7ITQ424BQQDB5LSO4JXL7W4EJ","short_pith_number":"pith:F7ITQ424","schema_version":"1.0","canonical_sha256":"2fd138735c0c2030f572771375fedc2261b28059697e110fe18b8776cc9c7750","source":{"kind":"arxiv","id":"1908.09791","version":5},"attestation_state":"computed","paper":{"title":"Once-for-All: Train One Network and Specialize it for Efficient Deployment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chuang Gan, Han Cai, Song Han, Tianzhe Wang, Zhekai Zhang","submitted_at":"2019-08-26T16:46:23Z","abstract_excerpt":"We address the challenging problem of efficient inference across many devices and resource constraints, especially on edge devices. Conventional approaches either manually design or use neural architecture search (NAS) to find a specialized neural network and train it from scratch for each case, which is computationally prohibitive (causing $CO_2$ emission as much as 5 cars' lifetime) thus unscalable. In this work, we propose to train a once-for-all (OFA) network that supports diverse architectural settings by decoupling training and search, to reduce the cost. We can quickly get a specialized"},"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":"1908.09791","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2019-08-26T16:46:23Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"323f5650fec8c210284e942e38ac2dc00d038ed71d76f6fb53a53314e44c2648","abstract_canon_sha256":"659043f764245514f43e1c3c57cc018b37c61dac6bb5a828702cb26db82c65cb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:59:22.429434Z","signature_b64":"prfJ0vZZBlXo6sDc89pJTbBfmsWeWp5MPtS/oLV2SbSOz0vfpoE6reWcjv4FljGTgT6/tilKUllYqn9rLmvGAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2fd138735c0c2030f572771375fedc2261b28059697e110fe18b8776cc9c7750","last_reissued_at":"2026-07-05T00:59:22.428999Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:59:22.428999Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Once-for-All: Train One Network and Specialize it for Efficient Deployment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chuang Gan, Han Cai, Song Han, Tianzhe Wang, Zhekai Zhang","submitted_at":"2019-08-26T16:46:23Z","abstract_excerpt":"We address the challenging problem of efficient inference across many devices and resource constraints, especially on edge devices. Conventional approaches either manually design or use neural architecture search (NAS) to find a specialized neural network and train it from scratch for each case, which is computationally prohibitive (causing $CO_2$ emission as much as 5 cars' lifetime) thus unscalable. In this work, we propose to train a once-for-all (OFA) network that supports diverse architectural settings by decoupling training and search, to reduce the cost. We can quickly get a specialized"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1908.09791","kind":"arxiv","version":5},"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/1908.09791/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":"1908.09791","created_at":"2026-07-05T00:59:22.429054+00:00"},{"alias_kind":"arxiv_version","alias_value":"1908.09791v5","created_at":"2026-07-05T00:59:22.429054+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1908.09791","created_at":"2026-07-05T00:59:22.429054+00:00"},{"alias_kind":"pith_short_12","alias_value":"F7ITQ424BQQD","created_at":"2026-07-05T00:59:22.429054+00:00"},{"alias_kind":"pith_short_16","alias_value":"F7ITQ424BQQDB5LS","created_at":"2026-07-05T00:59:22.429054+00:00"},{"alias_kind":"pith_short_8","alias_value":"F7ITQ424","created_at":"2026-07-05T00:59:22.429054+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":20,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.14824","citing_title":"Running hardware-aware neural architecture search on embedded devices under 512MB of RAM","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08735","citing_title":"Structure-Conditioned Actor-Critic Branches for Quality-Diversity Reinforcement Learning","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29582","citing_title":"Bilevel Optimization for Neural Architecture Search","ref_index":79,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26636","citing_title":"JetViT: Efficient High-Resolution Vision Transformer with Post-Training Attention Search","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26473","citing_title":"Orion: Enabling Self-adaptive Memory Management for On-device Online Continual Learning","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30019","citing_title":"elasticAI.explorer: Towards a Unified End-to-End Framework for Hardware-Aware Neural Architecture Search","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07621","citing_title":"HASA: Subnet Allocation for Compute-Constrained Model-Heterogeneous Federated Learning","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00535","citing_title":"DREAM-S: Speculative Decoding with Searchable Drafting and Target-Aware Refinement for Multimodal Generation","ref_index":121,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21560","citing_title":"AutoMCU: Feasibility-First MCU Neural Network Customization via LLM-based Multi-Agent Systems","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16138","citing_title":"SNAC-Pack 2.0: Scaled-Out Surrogate Neural Architecture Codesign","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2510.14235","citing_title":"Spiking Neural Network Architecture Search: A Survey","ref_index":132,"is_internal_anchor":false},{"citing_arxiv_id":"2601.15127","citing_title":"DeepFedNAS: Efficient Hardware-Aware Architecture Adaptation for Heterogeneous IoT Federations via Pareto-Guided Supernet Training","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04988","citing_title":"Prune-Quantize-Distill: An Ordered Pipeline for Efficient Neural Network Compression","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12491","citing_title":"Elastic Attention Cores for Scalable Vision Transformers","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11563","citing_title":"TCP-SSM: Efficient Vision State Space Models with Token-Conditioned Poles","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.26181","citing_title":"SWAN: World-Aware Adaptive Multimodal Networks for Runtime Variations","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2207.05608","citing_title":"Inner Monologue: Embodied Reasoning through Planning with Language Models","ref_index":89,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07850","citing_title":"MatryoshkaLoRA: Learning Accurate Hierarchical Low-Rank Representations for LLM Fine-Tuning","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17476","citing_title":"Privatar: Scalable Privacy-preserving Multi-user VR via Secure Offloading","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19958","citing_title":"Equinox: Decentralized Scheduling for Hardware-Aware Orbital Intelligence","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ","json":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ.json","graph_json":"https://pith.science/api/pith-number/F7ITQ424BQQDB5LSO4JXL7W4EJ/graph.json","events_json":"https://pith.science/api/pith-number/F7ITQ424BQQDB5LSO4JXL7W4EJ/events.json","paper":"https://pith.science/paper/F7ITQ424"},"agent_actions":{"view_html":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ","download_json":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ.json","view_paper":"https://pith.science/paper/F7ITQ424","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1908.09791&json=true","fetch_graph":"https://pith.science/api/pith-number/F7ITQ424BQQDB5LSO4JXL7W4EJ/graph.json","fetch_events":"https://pith.science/api/pith-number/F7ITQ424BQQDB5LSO4JXL7W4EJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ/action/storage_attestation","attest_author":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ/action/author_attestation","sign_citation":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ/action/citation_signature","submit_replication":"https://pith.science/pith/F7ITQ424BQQDB5LSO4JXL7W4EJ/action/replication_record"}},"created_at":"2026-07-05T00:59:22.429054+00:00","updated_at":"2026-07-05T00:59:22.429054+00:00"}