{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:LHRIBK5ZDLHG7LZK7Q3625IP64","short_pith_number":"pith:LHRIBK5Z","canonical_record":{"source":{"id":"2605.25451","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:01:45Z","cross_cats_sorted":[],"title_canon_sha256":"67968da5dfad49f03de281069cbe828fa2b1d288a1e0108ed626f7da9893306e","abstract_canon_sha256":"26fed20ecffba92c77157f0c1a3ab2ce5aee9f96925bd735350d232d6c2ca76b"},"schema_version":"1.0"},"canonical_sha256":"59e280abb91ace6faf2afc37ed750ff718627d1872a5905a7a2530a8652e6a08","source":{"kind":"arxiv","id":"2605.25451","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.25451","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.25451v1","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25451","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_12","alias_value":"LHRIBK5ZDLHG","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_16","alias_value":"LHRIBK5ZDLHG7LZK","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_8","alias_value":"LHRIBK5Z","created_at":"2026-05-26T02:04:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:LHRIBK5ZDLHG7LZK7Q3625IP64","target":"record","payload":{"canonical_record":{"source":{"id":"2605.25451","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:01:45Z","cross_cats_sorted":[],"title_canon_sha256":"67968da5dfad49f03de281069cbe828fa2b1d288a1e0108ed626f7da9893306e","abstract_canon_sha256":"26fed20ecffba92c77157f0c1a3ab2ce5aee9f96925bd735350d232d6c2ca76b"},"schema_version":"1.0"},"canonical_sha256":"59e280abb91ace6faf2afc37ed750ff718627d1872a5905a7a2530a8652e6a08","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T02:04:36.406791Z","signature_b64":"xmjZ6VxSz5+7gmuPb3BsoczwHod9ds8FYFWxHi29xB8u38zwX4xrAqKrOVc7DCOGz98zMCUwmWh9O971QQvDBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"59e280abb91ace6faf2afc37ed750ff718627d1872a5905a7a2530a8652e6a08","last_reissued_at":"2026-05-26T02:04:36.405922Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T02:04:36.405922Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.25451","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-26T02:04:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4HWoC2sKpat2K13PCXp9E2bdWGMSLqVNJes3c8usFiLHoIo3u1qzKjyV0NvkOZssgbF0eiw59dQS0ACdnJtTBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T21:21:27.285800Z"},"content_sha256":"ec3f9698ad458c51409297140ebd81cdb90aa5d2cdece07f8f3bb9813039d4d3","schema_version":"1.0","event_id":"sha256:ec3f9698ad458c51409297140ebd81cdb90aa5d2cdece07f8f3bb9813039d4d3"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:LHRIBK5ZDLHG7LZK7Q3625IP64","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"BigMac: Breaking the Pareto Frontier of Compute and Memory in Multimodal LLM Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chao Jin, Chengxu Yang, Chenyu Wang, Shenglong Zhang, Tuo Dai, Xin Jin, Yufan Zhang, Yuhong Ge, Yuliang Liu, Zhouyang Li, Zili Zhang","submitted_at":"2026-05-25T06:01:45Z","abstract_excerpt":"Training multimodal large language models (MLLMs) is challenged by both model and data heterogeneity. Existing systems redesign the training pipeline to address these challenges, but remain bound by a Pareto frontier between compute and memory efficiency, improving one only at the expense of the other. We present BigMac, a new training pipeline for multimodal LLMs. The core idea of BigMac is to elegantly nest the encoder and generator computation into the original LLM pipeline, forming a dependency-safe nested pipeline structure. With this design, BigMac reduces the activation memory complexit"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25451","kind":"arxiv","version":1},"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/2605.25451/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":"2026-05-26T02:04:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kdn/OpyXdU86c8+LP/ca+KP/x3HhQkWFp0+hx7TbW0RyjzBD5pFhRTbhOFEqcwWplDHqU8cJE5CVSmZOTva3BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T21:21:27.286264Z"},"content_sha256":"510a2b7ac99db2a403e8961e9cae302d989435aee0ae49da4be8563491602cf5","schema_version":"1.0","event_id":"sha256:510a2b7ac99db2a403e8961e9cae302d989435aee0ae49da4be8563491602cf5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LHRIBK5ZDLHG7LZK7Q3625IP64/bundle.json","state_url":"https://pith.science/pith/LHRIBK5ZDLHG7LZK7Q3625IP64/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LHRIBK5ZDLHG7LZK7Q3625IP64/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":"2026-05-27T21:21:27Z","links":{"resolver":"https://pith.science/pith/LHRIBK5ZDLHG7LZK7Q3625IP64","bundle":"https://pith.science/pith/LHRIBK5ZDLHG7LZK7Q3625IP64/bundle.json","state":"https://pith.science/pith/LHRIBK5ZDLHG7LZK7Q3625IP64/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LHRIBK5ZDLHG7LZK7Q3625IP64/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LHRIBK5ZDLHG7LZK7Q3625IP64","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":"26fed20ecffba92c77157f0c1a3ab2ce5aee9f96925bd735350d232d6c2ca76b","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:01:45Z","title_canon_sha256":"67968da5dfad49f03de281069cbe828fa2b1d288a1e0108ed626f7da9893306e"},"schema_version":"1.0","source":{"id":"2605.25451","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.25451","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"arxiv_version","alias_value":"2605.25451v1","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.25451","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_12","alias_value":"LHRIBK5ZDLHG","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_16","alias_value":"LHRIBK5ZDLHG7LZK","created_at":"2026-05-26T02:04:36Z"},{"alias_kind":"pith_short_8","alias_value":"LHRIBK5Z","created_at":"2026-05-26T02:04:36Z"}],"graph_snapshots":[{"event_id":"sha256:510a2b7ac99db2a403e8961e9cae302d989435aee0ae49da4be8563491602cf5","target":"graph","created_at":"2026-05-26T02:04:36Z","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/2605.25451/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Training multimodal large language models (MLLMs) is challenged by both model and data heterogeneity. Existing systems redesign the training pipeline to address these challenges, but remain bound by a Pareto frontier between compute and memory efficiency, improving one only at the expense of the other. We present BigMac, a new training pipeline for multimodal LLMs. The core idea of BigMac is to elegantly nest the encoder and generator computation into the original LLM pipeline, forming a dependency-safe nested pipeline structure. With this design, BigMac reduces the activation memory complexit","authors_text":"Chao Jin, Chengxu Yang, Chenyu Wang, Shenglong Zhang, Tuo Dai, Xin Jin, Yufan Zhang, Yuhong Ge, Yuliang Liu, Zhouyang Li, Zili Zhang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:01:45Z","title":"BigMac: Breaking the Pareto Frontier of Compute and Memory in Multimodal LLM Training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.25451","kind":"arxiv","version":1},"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:ec3f9698ad458c51409297140ebd81cdb90aa5d2cdece07f8f3bb9813039d4d3","target":"record","created_at":"2026-05-26T02:04:36Z","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":"26fed20ecffba92c77157f0c1a3ab2ce5aee9f96925bd735350d232d6c2ca76b","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-25T06:01:45Z","title_canon_sha256":"67968da5dfad49f03de281069cbe828fa2b1d288a1e0108ed626f7da9893306e"},"schema_version":"1.0","source":{"id":"2605.25451","kind":"arxiv","version":1}},"canonical_sha256":"59e280abb91ace6faf2afc37ed750ff718627d1872a5905a7a2530a8652e6a08","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"59e280abb91ace6faf2afc37ed750ff718627d1872a5905a7a2530a8652e6a08","first_computed_at":"2026-05-26T02:04:36.405922Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T02:04:36.405922Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"xmjZ6VxSz5+7gmuPb3BsoczwHod9ds8FYFWxHi29xB8u38zwX4xrAqKrOVc7DCOGz98zMCUwmWh9O971QQvDBg==","signature_status":"signed_v1","signed_at":"2026-05-26T02:04:36.406791Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.25451","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ec3f9698ad458c51409297140ebd81cdb90aa5d2cdece07f8f3bb9813039d4d3","sha256:510a2b7ac99db2a403e8961e9cae302d989435aee0ae49da4be8563491602cf5"],"state_sha256":"1ad6d3492086d08eaa97199f44228fe2b2c0d617e7979623fbe086f5c1126b3c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Qf4h4jGYA/Ds8mhnVweIzwhQZGQVqXDY81FOAkzRr3BiO++tJ1G7g+ODNA9w8KFwSyhpsbmOOjmEvFsAmcwpAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-27T21:21:27.288952Z","bundle_sha256":"d1cc6c4359f4046b14adad24cc9e6c2286d9839777eae3696c7685fd5a3227fd"}}