{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2020:KRR6ON3MGCWSVK3CIW45EEK4CY","short_pith_number":"pith:KRR6ON3M","canonical_record":{"source":{"id":"2002.10957","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-25T15:21:10Z","cross_cats_sorted":[],"title_canon_sha256":"2e56a006344a3e51c13cea4a583852eb3647d9be192ac97808dcf61964c3feb9","abstract_canon_sha256":"e4ee673f036f941f83ef43180f9730ed3f34232b430acbb62faf1765b2b1a3ed"},"schema_version":"1.0"},"canonical_sha256":"5463e7376c30ad2aab6245b9d2115c162d3299095e2056d59a81ff11efacc134","source":{"kind":"arxiv","id":"2002.10957","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2002.10957","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"arxiv_version","alias_value":"2002.10957v2","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.10957","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"pith_short_12","alias_value":"KRR6ON3MGCWS","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"pith_short_16","alias_value":"KRR6ON3MGCWSVK3C","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"pith_short_8","alias_value":"KRR6ON3M","created_at":"2026-07-05T00:52:52Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2020:KRR6ON3MGCWSVK3CIW45EEK4CY","target":"record","payload":{"canonical_record":{"source":{"id":"2002.10957","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-25T15:21:10Z","cross_cats_sorted":[],"title_canon_sha256":"2e56a006344a3e51c13cea4a583852eb3647d9be192ac97808dcf61964c3feb9","abstract_canon_sha256":"e4ee673f036f941f83ef43180f9730ed3f34232b430acbb62faf1765b2b1a3ed"},"schema_version":"1.0"},"canonical_sha256":"5463e7376c30ad2aab6245b9d2115c162d3299095e2056d59a81ff11efacc134","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:52:52.633456Z","signature_b64":"SUHJlE+3sRqFTqgZaU63lT8aQM2DqJqlWiHG4rg3vb36b3SMd5oOiGMRvmUqdBEedMqHUFB9AiQ3WyYAJrpHDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5463e7376c30ad2aab6245b9d2115c162d3299095e2056d59a81ff11efacc134","last_reissued_at":"2026-07-05T00:52:52.632968Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:52:52.632968Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2002.10957","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":"2026-07-05T00:52:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"zDfg2y+TbrAHB0dMrcSik4s0eFK3O320BHJEwhSbmy7LnTBX3raDJbbiPfyws6xYj9vHdTKfdiAuj4XZzobKCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T08:20:13.261311Z"},"content_sha256":"7834ddfc81cb8ef5ff9562346193301d22e7d54b9712568f97156b93bf983309","schema_version":"1.0","event_id":"sha256:7834ddfc81cb8ef5ff9562346193301d22e7d54b9712568f97156b93bf983309"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2020:KRR6ON3MGCWSVK3CIW45EEK4CY","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Furu Wei, Hangbo Bao, Li Dong, Ming Zhou, Nan Yang, Wenhui Wang","submitted_at":"2020-02-25T15:21:10Z","abstract_excerpt":"Pre-trained language models (e.g., BERT (Devlin et al., 2018) and its variants) have achieved remarkable success in varieties of NLP tasks. However, these models usually consist of hundreds of millions of parameters which brings challenges for fine-tuning and online serving in real-life applications due to latency and capacity constraints. In this work, we present a simple and effective approach to compress large Transformer (Vaswani et al., 2017) based pre-trained models, termed as deep self-attention distillation. The small model (student) is trained by deeply mimicking the self-attention mo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2002.10957","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/2002.10957/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":"2026-07-05T00:52:52Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CGIxHZXff/2GxXFdfRwR6vEjP7VW+rxdD6abK7TbnB2DZP3TcLOKHvmJn1GH7wmaAp0LdfpQKWsuDVuv0GLCDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T08:20:13.261699Z"},"content_sha256":"0d11be3ea2539c54e15d651ffc888d95f87e1e1d4935f3662c2d4334b69d96a3","schema_version":"1.0","event_id":"sha256:0d11be3ea2539c54e15d651ffc888d95f87e1e1d4935f3662c2d4334b69d96a3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KRR6ON3MGCWSVK3CIW45EEK4CY/bundle.json","state_url":"https://pith.science/pith/KRR6ON3MGCWSVK3CIW45EEK4CY/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KRR6ON3MGCWSVK3CIW45EEK4CY/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":"2026-07-07T08:20:13Z","links":{"resolver":"https://pith.science/pith/KRR6ON3MGCWSVK3CIW45EEK4CY","bundle":"https://pith.science/pith/KRR6ON3MGCWSVK3CIW45EEK4CY/bundle.json","state":"https://pith.science/pith/KRR6ON3MGCWSVK3CIW45EEK4CY/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KRR6ON3MGCWSVK3CIW45EEK4CY/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:KRR6ON3MGCWSVK3CIW45EEK4CY","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":"e4ee673f036f941f83ef43180f9730ed3f34232b430acbb62faf1765b2b1a3ed","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-25T15:21:10Z","title_canon_sha256":"2e56a006344a3e51c13cea4a583852eb3647d9be192ac97808dcf61964c3feb9"},"schema_version":"1.0","source":{"id":"2002.10957","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2002.10957","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"arxiv_version","alias_value":"2002.10957v2","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.10957","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"pith_short_12","alias_value":"KRR6ON3MGCWS","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"pith_short_16","alias_value":"KRR6ON3MGCWSVK3C","created_at":"2026-07-05T00:52:52Z"},{"alias_kind":"pith_short_8","alias_value":"KRR6ON3M","created_at":"2026-07-05T00:52:52Z"}],"graph_snapshots":[{"event_id":"sha256:0d11be3ea2539c54e15d651ffc888d95f87e1e1d4935f3662c2d4334b69d96a3","target":"graph","created_at":"2026-07-05T00:52:52Z","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/2002.10957/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pre-trained language models (e.g., BERT (Devlin et al., 2018) and its variants) have achieved remarkable success in varieties of NLP tasks. However, these models usually consist of hundreds of millions of parameters which brings challenges for fine-tuning and online serving in real-life applications due to latency and capacity constraints. In this work, we present a simple and effective approach to compress large Transformer (Vaswani et al., 2017) based pre-trained models, termed as deep self-attention distillation. The small model (student) is trained by deeply mimicking the self-attention mo","authors_text":"Furu Wei, Hangbo Bao, Li Dong, Ming Zhou, Nan Yang, Wenhui Wang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-25T15:21:10Z","title":"MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2002.10957","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":"sha256:7834ddfc81cb8ef5ff9562346193301d22e7d54b9712568f97156b93bf983309","target":"record","created_at":"2026-07-05T00:52:52Z","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":"e4ee673f036f941f83ef43180f9730ed3f34232b430acbb62faf1765b2b1a3ed","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-02-25T15:21:10Z","title_canon_sha256":"2e56a006344a3e51c13cea4a583852eb3647d9be192ac97808dcf61964c3feb9"},"schema_version":"1.0","source":{"id":"2002.10957","kind":"arxiv","version":2}},"canonical_sha256":"5463e7376c30ad2aab6245b9d2115c162d3299095e2056d59a81ff11efacc134","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5463e7376c30ad2aab6245b9d2115c162d3299095e2056d59a81ff11efacc134","first_computed_at":"2026-07-05T00:52:52.632968Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T00:52:52.632968Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SUHJlE+3sRqFTqgZaU63lT8aQM2DqJqlWiHG4rg3vb36b3SMd5oOiGMRvmUqdBEedMqHUFB9AiQ3WyYAJrpHDA==","signature_status":"signed_v1","signed_at":"2026-07-05T00:52:52.633456Z","signed_message":"canonical_sha256_bytes"},"source_id":"2002.10957","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7834ddfc81cb8ef5ff9562346193301d22e7d54b9712568f97156b93bf983309","sha256:0d11be3ea2539c54e15d651ffc888d95f87e1e1d4935f3662c2d4334b69d96a3"],"state_sha256":"bdcc2f5645c23042f6164b2213c9f304a6b444698317cab63e3d748f4ac39481"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"bP5/uARPHBDJ3nvDeIKE6bEc1I9mu7Dt00U+h5jbIY1LSdrP2mgBribdfzf9pnEWk9XvxMjU5fIfScUz3qclCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T08:20:13.263646Z","bundle_sha256":"bfaf08133861bca964927eedf991c672c4ecc7cf3000d50f2d401ebb9fa60d41"}}