{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2017:CP5CROAVFCAIBBAYYIYGPAKOWN","short_pith_number":"pith:CP5CROAV","canonical_record":{"source":{"id":"1705.01462","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-05-02T10:15:21Z","cross_cats_sorted":["cs.NE"],"title_canon_sha256":"7fe33303c91a093e9aa243b3fb7742a975afbd92fac0d6345d21a19523ae3393","abstract_canon_sha256":"68001054b81cd736818a1a4f46be60fbd99059730cd42454f97e62d32af7ef4c"},"schema_version":"1.0"},"canonical_sha256":"13fa28b8152880808418c23067814eb3484604f42cb5927a50be3c813c710e27","source":{"kind":"arxiv","id":"1705.01462","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1705.01462","created_at":"2026-05-18T00:43:25Z"},{"alias_kind":"arxiv_version","alias_value":"1705.01462v3","created_at":"2026-05-18T00:43:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1705.01462","created_at":"2026-05-18T00:43:25Z"},{"alias_kind":"pith_short_12","alias_value":"CP5CROAVFCAI","created_at":"2026-05-18T12:31:10Z"},{"alias_kind":"pith_short_16","alias_value":"CP5CROAVFCAIBBAY","created_at":"2026-05-18T12:31:10Z"},{"alias_kind":"pith_short_8","alias_value":"CP5CROAV","created_at":"2026-05-18T12:31:10Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2017:CP5CROAVFCAIBBAYYIYGPAKOWN","target":"record","payload":{"canonical_record":{"source":{"id":"1705.01462","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-05-02T10:15:21Z","cross_cats_sorted":["cs.NE"],"title_canon_sha256":"7fe33303c91a093e9aa243b3fb7742a975afbd92fac0d6345d21a19523ae3393","abstract_canon_sha256":"68001054b81cd736818a1a4f46be60fbd99059730cd42454f97e62d32af7ef4c"},"schema_version":"1.0"},"canonical_sha256":"13fa28b8152880808418c23067814eb3484604f42cb5927a50be3c813c710e27","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:43:25.233509Z","signature_b64":"kA39Fo5YkqrdpKxR/6tuBPnYCWQCGkLuG/nluHrqBic3w7QtKS+dBQcE17uoJST9WkZk7MSijEkfkL1axb+rBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"13fa28b8152880808418c23067814eb3484604f42cb5927a50be3c813c710e27","last_reissued_at":"2026-05-18T00:43:25.232926Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:43:25.232926Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1705.01462","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:43:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"e0MavY24xRt/cCGwYpwEu0Hng1qpxPM3dbnvF8v4biLZxnD27ui+/nyCrsO+WotXtNrg4NzqHy9f1E/KLKbCAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T08:55:38.959301Z"},"content_sha256":"7cd46fb1d9dd2a59e9eec226cbe565f44b2ae452171d162fe62bed6d475261f8","schema_version":"1.0","event_id":"sha256:7cd46fb1d9dd2a59e9eec226cbe565f44b2ae452171d162fe62bed6d475261f8"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2017:CP5CROAVFCAIBBAYYIYGPAKOWN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Ternary Neural Networks with Fine-Grained Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NE"],"primary_cat":"cs.LG","authors_text":"Abhisek Kundu, Bharat Kaul, Dheevatsa Mudigere, Dipankar Das, Naveen Mellempudi, Pradeep Dubey","submitted_at":"2017-05-02T10:15:21Z","abstract_excerpt":"We propose a novel fine-grained quantization (FGQ) method to ternarize pre-trained full precision models, while also constraining activations to 8 and 4-bits. Using this method, we demonstrate a minimal loss in classification accuracy on state-of-the-art topologies without additional training. We provide an improved theoretical formulation that forms the basis for a higher quality solution using FGQ. Our method involves ternarizing the original weight tensor in groups of $N$ weights. Using $N=4$, we achieve Top-1 accuracy within $3.7\\%$ and $4.2\\%$ of the baseline full precision result for Res"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1705.01462","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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:43:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Uxsjhr2s7UH9TSfYEc8qW+GOJks6PTsMaBAV7pN8vaxwyyxWe4hrQAgeoXN8MkMcbd7qB5EKU+O6SCs2AwGMBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-30T08:55:38.959672Z"},"content_sha256":"096c24fefb5f030caba8f3c401ebd285e2c4458a0fad89d5a502ccd4d7d7718a","schema_version":"1.0","event_id":"sha256:096c24fefb5f030caba8f3c401ebd285e2c4458a0fad89d5a502ccd4d7d7718a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CP5CROAVFCAIBBAYYIYGPAKOWN/bundle.json","state_url":"https://pith.science/pith/CP5CROAVFCAIBBAYYIYGPAKOWN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CP5CROAVFCAIBBAYYIYGPAKOWN/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":"2026-05-30T08:55:38Z","links":{"resolver":"https://pith.science/pith/CP5CROAVFCAIBBAYYIYGPAKOWN","bundle":"https://pith.science/pith/CP5CROAVFCAIBBAYYIYGPAKOWN/bundle.json","state":"https://pith.science/pith/CP5CROAVFCAIBBAYYIYGPAKOWN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CP5CROAVFCAIBBAYYIYGPAKOWN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2017:CP5CROAVFCAIBBAYYIYGPAKOWN","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":"68001054b81cd736818a1a4f46be60fbd99059730cd42454f97e62d32af7ef4c","cross_cats_sorted":["cs.NE"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-05-02T10:15:21Z","title_canon_sha256":"7fe33303c91a093e9aa243b3fb7742a975afbd92fac0d6345d21a19523ae3393"},"schema_version":"1.0","source":{"id":"1705.01462","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1705.01462","created_at":"2026-05-18T00:43:25Z"},{"alias_kind":"arxiv_version","alias_value":"1705.01462v3","created_at":"2026-05-18T00:43:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1705.01462","created_at":"2026-05-18T00:43:25Z"},{"alias_kind":"pith_short_12","alias_value":"CP5CROAVFCAI","created_at":"2026-05-18T12:31:10Z"},{"alias_kind":"pith_short_16","alias_value":"CP5CROAVFCAIBBAY","created_at":"2026-05-18T12:31:10Z"},{"alias_kind":"pith_short_8","alias_value":"CP5CROAV","created_at":"2026-05-18T12:31:10Z"}],"graph_snapshots":[{"event_id":"sha256:096c24fefb5f030caba8f3c401ebd285e2c4458a0fad89d5a502ccd4d7d7718a","target":"graph","created_at":"2026-05-18T00:43:25Z","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"},"paper":{"abstract_excerpt":"We propose a novel fine-grained quantization (FGQ) method to ternarize pre-trained full precision models, while also constraining activations to 8 and 4-bits. Using this method, we demonstrate a minimal loss in classification accuracy on state-of-the-art topologies without additional training. We provide an improved theoretical formulation that forms the basis for a higher quality solution using FGQ. Our method involves ternarizing the original weight tensor in groups of $N$ weights. Using $N=4$, we achieve Top-1 accuracy within $3.7\\%$ and $4.2\\%$ of the baseline full precision result for Res","authors_text":"Abhisek Kundu, Bharat Kaul, Dheevatsa Mudigere, Dipankar Das, Naveen Mellempudi, Pradeep Dubey","cross_cats":["cs.NE"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-05-02T10:15:21Z","title":"Ternary Neural Networks with Fine-Grained Quantization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1705.01462","kind":"arxiv","version":3},"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":"sha256:7cd46fb1d9dd2a59e9eec226cbe565f44b2ae452171d162fe62bed6d475261f8","target":"record","created_at":"2026-05-18T00:43:25Z","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":"68001054b81cd736818a1a4f46be60fbd99059730cd42454f97e62d32af7ef4c","cross_cats_sorted":["cs.NE"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-05-02T10:15:21Z","title_canon_sha256":"7fe33303c91a093e9aa243b3fb7742a975afbd92fac0d6345d21a19523ae3393"},"schema_version":"1.0","source":{"id":"1705.01462","kind":"arxiv","version":3}},"canonical_sha256":"13fa28b8152880808418c23067814eb3484604f42cb5927a50be3c813c710e27","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"13fa28b8152880808418c23067814eb3484604f42cb5927a50be3c813c710e27","first_computed_at":"2026-05-18T00:43:25.232926Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T00:43:25.232926Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"kA39Fo5YkqrdpKxR/6tuBPnYCWQCGkLuG/nluHrqBic3w7QtKS+dBQcE17uoJST9WkZk7MSijEkfkL1axb+rBQ==","signature_status":"signed_v1","signed_at":"2026-05-18T00:43:25.233509Z","signed_message":"canonical_sha256_bytes"},"source_id":"1705.01462","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7cd46fb1d9dd2a59e9eec226cbe565f44b2ae452171d162fe62bed6d475261f8","sha256:096c24fefb5f030caba8f3c401ebd285e2c4458a0fad89d5a502ccd4d7d7718a"],"state_sha256":"6b9b99a409300a6c8ca64f1ab22393cc029985aca1e393c75157933b26bbede3"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WqmDH/w564bGec3iF8QdyBiUC9AT0hFskPz7Q+bLc/iuyXhrFy0H4ogn2fMo5gH0Fy7jEpjpR1bAf3mrzaBtAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-30T08:55:38.961845Z","bundle_sha256":"75ca9d4a6d19de523eff833a63ef7023da87d5eb7539fe63d6d36e14c3278f0a"}}