{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:ZPXWVRZVOURYMXCA2VCAIIFRTS","short_pith_number":"pith:ZPXWVRZV","canonical_record":{"source":{"id":"2502.00965","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-02-03T00:04:50Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"b72a96d5f832f9399d85e359b9ae99a5a40a85459e57b5f2942f117791e023f1","abstract_canon_sha256":"db6fa1e2ea539cf8270b40682d288ee033f6083003d7d82a5b1f05e807d5f89e"},"schema_version":"1.0"},"canonical_sha256":"cbef6ac7357523865c40d5440420b19cada8edd21f1ea115d982757fba4ef237","source":{"kind":"arxiv","id":"2502.00965","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.00965","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"arxiv_version","alias_value":"2502.00965v2","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.00965","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"pith_short_12","alias_value":"ZPXWVRZVOURY","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"pith_short_16","alias_value":"ZPXWVRZVOURYMXCA","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"pith_short_8","alias_value":"ZPXWVRZV","created_at":"2026-07-05T11:09:07Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:ZPXWVRZVOURYMXCA2VCAIIFRTS","target":"record","payload":{"canonical_record":{"source":{"id":"2502.00965","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-02-03T00:04:50Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"b72a96d5f832f9399d85e359b9ae99a5a40a85459e57b5f2942f117791e023f1","abstract_canon_sha256":"db6fa1e2ea539cf8270b40682d288ee033f6083003d7d82a5b1f05e807d5f89e"},"schema_version":"1.0"},"canonical_sha256":"cbef6ac7357523865c40d5440420b19cada8edd21f1ea115d982757fba4ef237","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:09:07.476873Z","signature_b64":"7dHOTZioNLcnS10xHSd9fiqEb+GgKOmvzNfFH1WL3GRjEPzqtf3BwkpQxTVKYktMKWNegjNsZ4t4Y0s37E2ABA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cbef6ac7357523865c40d5440420b19cada8edd21f1ea115d982757fba4ef237","last_reissued_at":"2026-07-05T11:09:07.476405Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:09:07.476405Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.00965","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:09:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"75ElDUcTDcoSCTjyaGZxvD8NHGKEPDa6uyG4b4hBNfVxPkcNDy3BadzDzA6uRAXjSyyBZw+Sn5IV7zXx0IkGDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T09:18:30.049857Z"},"content_sha256":"d63eefbea1f26e5ea91a2c6dc2df9202f58fe528cb04a2d023ab845ab632a696","schema_version":"1.0","event_id":"sha256:d63eefbea1f26e5ea91a2c6dc2df9202f58fe528cb04a2d023ab845ab632a696"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:ZPXWVRZVOURYMXCA2VCAIIFRTS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"CLIP-UP: A Simple and Efficient Mixture-of-Experts CLIP Training Recipe with Sparse Upcycling","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Aditya Pal, Bowen Zhang, Chen Chen, Hong-You Chen, Xiangxin Zhu, Xianzhi Du, Xinze Wang, Yinfei Yang","submitted_at":"2025-02-03T00:04:50Z","abstract_excerpt":"Mixture-of-Experts (MoE) models are crucial for scaling model capacity while controlling inference costs. While integrating MoE into multimodal models like CLIP improves performance, training these models is notoriously challenging and expensive. We propose CLIP-Upcycling (CLIP-UP), an efficient alternative training strategy that converts a pre-trained dense CLIP model into a sparse MoE architecture. Through extensive experimentation with various settings and auxiliary losses, we demonstrate that CLIP-UP significantly reduces training complexity and cost. Remarkably, our sparse CLIP B/16 model"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.00965","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/2502.00965/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:09:07Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8PhfUuVA4DwcPFcP+9hbrRAispIQ1evPqvHJQ6KURFxjK4lR9SPQIKjDetn0s45T/KDOOcWlKCHN3LuaPyb2Bw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T09:18:30.050348Z"},"content_sha256":"76937f004b35a9ce091ffc081a054cd93047c6da531cb0623f1eb80a57ddb31c","schema_version":"1.0","event_id":"sha256:76937f004b35a9ce091ffc081a054cd93047c6da531cb0623f1eb80a57ddb31c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZPXWVRZVOURYMXCA2VCAIIFRTS/bundle.json","state_url":"https://pith.science/pith/ZPXWVRZVOURYMXCA2VCAIIFRTS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZPXWVRZVOURYMXCA2VCAIIFRTS/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-10T09:18:30Z","links":{"resolver":"https://pith.science/pith/ZPXWVRZVOURYMXCA2VCAIIFRTS","bundle":"https://pith.science/pith/ZPXWVRZVOURYMXCA2VCAIIFRTS/bundle.json","state":"https://pith.science/pith/ZPXWVRZVOURYMXCA2VCAIIFRTS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZPXWVRZVOURYMXCA2VCAIIFRTS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:ZPXWVRZVOURYMXCA2VCAIIFRTS","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"db6fa1e2ea539cf8270b40682d288ee033f6083003d7d82a5b1f05e807d5f89e","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-02-03T00:04:50Z","title_canon_sha256":"b72a96d5f832f9399d85e359b9ae99a5a40a85459e57b5f2942f117791e023f1"},"schema_version":"1.0","source":{"id":"2502.00965","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.00965","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"arxiv_version","alias_value":"2502.00965v2","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.00965","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"pith_short_12","alias_value":"ZPXWVRZVOURY","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"pith_short_16","alias_value":"ZPXWVRZVOURYMXCA","created_at":"2026-07-05T11:09:07Z"},{"alias_kind":"pith_short_8","alias_value":"ZPXWVRZV","created_at":"2026-07-05T11:09:07Z"}],"graph_snapshots":[{"event_id":"sha256:76937f004b35a9ce091ffc081a054cd93047c6da531cb0623f1eb80a57ddb31c","target":"graph","created_at":"2026-07-05T11:09:07Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2502.00965/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Mixture-of-Experts (MoE) models are crucial for scaling model capacity while controlling inference costs. While integrating MoE into multimodal models like CLIP improves performance, training these models is notoriously challenging and expensive. We propose CLIP-Upcycling (CLIP-UP), an efficient alternative training strategy that converts a pre-trained dense CLIP model into a sparse MoE architecture. Through extensive experimentation with various settings and auxiliary losses, we demonstrate that CLIP-UP significantly reduces training complexity and cost. Remarkably, our sparse CLIP B/16 model","authors_text":"Aditya Pal, Bowen Zhang, Chen Chen, Hong-You Chen, Xiangxin Zhu, Xianzhi Du, Xinze Wang, Yinfei Yang","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-02-03T00:04:50Z","title":"CLIP-UP: A Simple and Efficient Mixture-of-Experts CLIP Training Recipe with Sparse Upcycling"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.00965","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:d63eefbea1f26e5ea91a2c6dc2df9202f58fe528cb04a2d023ab845ab632a696","target":"record","created_at":"2026-07-05T11:09:07Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"db6fa1e2ea539cf8270b40682d288ee033f6083003d7d82a5b1f05e807d5f89e","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-02-03T00:04:50Z","title_canon_sha256":"b72a96d5f832f9399d85e359b9ae99a5a40a85459e57b5f2942f117791e023f1"},"schema_version":"1.0","source":{"id":"2502.00965","kind":"arxiv","version":2}},"canonical_sha256":"cbef6ac7357523865c40d5440420b19cada8edd21f1ea115d982757fba4ef237","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"cbef6ac7357523865c40d5440420b19cada8edd21f1ea115d982757fba4ef237","first_computed_at":"2026-07-05T11:09:07.476405Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:09:07.476405Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"7dHOTZioNLcnS10xHSd9fiqEb+GgKOmvzNfFH1WL3GRjEPzqtf3BwkpQxTVKYktMKWNegjNsZ4t4Y0s37E2ABA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:09:07.476873Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.00965","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:d63eefbea1f26e5ea91a2c6dc2df9202f58fe528cb04a2d023ab845ab632a696","sha256:76937f004b35a9ce091ffc081a054cd93047c6da531cb0623f1eb80a57ddb31c"],"state_sha256":"905088c811a6c7d7d6bafa25921b0b1983d6dbbeede947f9270341ae27b7286c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"R8MqJSKiy5gNUfW6f8ahk5n/RblKUNx81Pen+LZilGNNistuQk9qNrXXZ33Qd/NJxtektX2XSHlbsOQZjUtiAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T09:18:30.054336Z","bundle_sha256":"0198b5b314ff42ddc9b48ec86093595ff61e7efcb0d90b157323814458cf6c4f"}}