Datasets:
Download baseline_study/verification.json from tim-shu/drug-affinity-multisource: direct link, hf CLI and curl.
- Browser
- Download file 8.23 kB
-
https://huggingface.co/datasets/tim-shu/drug-affinity-multisource/resolve/main/baseline_study/verification.json
- Command line
-
hf download hf://datasets/tim-shu/drug-affinity-multisource/baseline_study/verification.json
-
curl -L -o verification.json https://huggingface.co/datasets/tim-shu/drug-affinity-multisource/resolve/main/baseline_study/verification.json
8.23 kB
| { | |
| "verified_at_utc": "2026-11-10T19:50:47.387450+00:00", | |
| "checks": [ | |
| "common cohort matches full-structure inputs exactly", | |
| "unique unambiguous labelled pairs", | |
| "complete structures and selected inputs do not cross partitions", | |
| "every label is the median of observations of that full structure only", | |
| "main curated data has no component selection fields", | |
| "recorded components are preserved in source structures", | |
| "regression fixture shares an experimental selected component", | |
| "different complete forms never collapse into a median label", | |
| "scaffolds and molecules do not cross partitions", | |
| "prepared inputs have verified checksums", | |
| "source inputs have verified checksums", | |
| "ChEMBL source counts and percentages", | |
| "BindingDB source counts and percentages", | |
| "DrugCentral source counts and percentages", | |
| "feature indexes match structures", | |
| "count and binary fingerprints encode the same presence bits", | |
| "descriptor matrices are finite and aligned", | |
| "sample molecular properties and counts recomputed from SMILES", | |
| "radius ablation includes radii 1 and 3", | |
| "radius 1 vectors have the same width and binary encoding", | |
| "radius 1 sample vectors independently recomputed", | |
| "RF_full_radius1 changes only radius", | |
| "RF_full_radius1 has no additional features", | |
| "RF_full_descriptors_radius1 changes only radius", | |
| "radius 3 vectors have the same width and binary encoding", | |
| "radius 3 sample vectors independently recomputed", | |
| "RF_full_radius3 changes only radius", | |
| "RF_full_radius3 has no additional features", | |
| "RF_full_descriptors_radius3 changes only radius", | |
| "radius variants use the intended matrices and identical added properties", | |
| "width ablation includes 1024 and 4096 bits", | |
| "1024-bit vectors have the expected shape and binary encoding", | |
| "1024-bit sample vectors independently recomputed", | |
| "RF_full_bits1024 changes width with effective RF settings fixed", | |
| "RF_full_bits1024 has no additional features", | |
| "RF_full_descriptors_bits1024 changes width with effective RF settings fixed", | |
| "4096-bit vectors have the expected shape and binary encoding", | |
| "4096-bit sample vectors independently recomputed", | |
| "RF_full_bits4096 changes width with effective RF settings fixed", | |
| "RF_full_bits4096 has no additional features", | |
| "RF_full_descriptors_bits4096 changes width with effective RF settings fixed", | |
| "width variants use aligned intended matrices and identical added properties", | |
| "frozen recipe records the effective selected fingerprint radius and width", | |
| "trained implementation and configuration are current", | |
| "all candidates are evaluated", | |
| "RF_full_dropout/42 covers validation and test without duplicates", | |
| "RF_full_dropout/43 covers validation and test without duplicates", | |
| "RF_full_dropout/44 covers validation and test without duplicates", | |
| "Tanimoto_5NN/42 covers validation and test without duplicates", | |
| "Tanimoto_1NN/42 covers validation and test without duplicates", | |
| "RF_full_bits1024/42 covers validation and test without duplicates", | |
| "RF_full_bits1024/43 covers validation and test without duplicates", | |
| "RF_full_bits1024/44 covers validation and test without duplicates", | |
| "RF_full_radius1/42 covers validation and test without duplicates", | |
| "RF_full_radius1/43 covers validation and test without duplicates", | |
| "RF_full_radius1/44 covers validation and test without duplicates", | |
| "Median/42 covers validation and test without duplicates", | |
| "RF_full_descriptors_bits4096/42 covers validation and test without duplicates", | |
| "RF_full_descriptors_bits4096/43 covers validation and test without duplicates", | |
| "RF_full_descriptors_bits4096/44 covers validation and test without duplicates", | |
| "RF_parent/42 covers validation and test without duplicates", | |
| "RF_parent/43 covers validation and test without duplicates", | |
| "RF_parent/44 covers validation and test without duplicates", | |
| "RF_full_descriptors/42 covers validation and test without duplicates", | |
| "RF_full_descriptors/43 covers validation and test without duplicates", | |
| "RF_full_descriptors/44 covers validation and test without duplicates", | |
| "RF_full_descriptors_bits1024/42 covers validation and test without duplicates", | |
| "RF_full_descriptors_bits1024/43 covers validation and test without duplicates", | |
| "RF_full_descriptors_bits1024/44 covers validation and test without duplicates", | |
| "RF_full_leaf10/42 covers validation and test without duplicates", | |
| "RF_full_leaf10/43 covers validation and test without duplicates", | |
| "RF_full_leaf10/44 covers validation and test without duplicates", | |
| "RF_full_counts/42 covers validation and test without duplicates", | |
| "RF_full_counts/43 covers validation and test without duplicates", | |
| "RF_full_counts/44 covers validation and test without duplicates", | |
| "Mean/42 covers validation and test without duplicates", | |
| "RF_full_bits4096/42 covers validation and test without duplicates", | |
| "RF_full_bits4096/43 covers validation and test without duplicates", | |
| "RF_full_bits4096/44 covers validation and test without duplicates", | |
| "RF_full_radius3/42 covers validation and test without duplicates", | |
| "RF_full_radius3/43 covers validation and test without duplicates", | |
| "RF_full_radius3/44 covers validation and test without duplicates", | |
| "RF_full/42 covers validation and test without duplicates", | |
| "RF_full/43 covers validation and test without duplicates", | |
| "RF_full/44 covers validation and test without duplicates", | |
| "RF_descriptors_only/42 covers validation and test without duplicates", | |
| "RF_descriptors_only/43 covers validation and test without duplicates", | |
| "RF_descriptors_only/44 covers validation and test without duplicates", | |
| "RF_full_leaf3/42 covers validation and test without duplicates", | |
| "RF_full_leaf3/43 covers validation and test without duplicates", | |
| "RF_full_leaf3/44 covers validation and test without duplicates", | |
| "RF_parent_descriptors/42 covers validation and test without duplicates", | |
| "RF_parent_descriptors/43 covers validation and test without duplicates", | |
| "RF_parent_descriptors/44 covers validation and test without duplicates", | |
| "RF_full_descriptors_radius1/42 covers validation and test without duplicates", | |
| "RF_full_descriptors_radius1/43 covers validation and test without duplicates", | |
| "RF_full_descriptors_radius1/44 covers validation and test without duplicates", | |
| "RF_full_descriptors_radius3/42 covers validation and test without duplicates", | |
| "RF_full_descriptors_radius3/43 covers validation and test without duplicates", | |
| "RF_full_descriptors_radius3/44 covers validation and test without duplicates", | |
| "all validation MAE and RMSE independently recomputed", | |
| "winner selected by validation only and fixed seed is prespecified", | |
| "test predictions have matching labels and identifiers", | |
| "reported test errors recomputed from saved predictions", | |
| "sample heuristic predictions use only same-target train labels", | |
| "saved forest checksums and identities", | |
| "O00408 saved forest has the selected feature width and settings", | |
| "P21728 saved forest has the selected feature width and settings", | |
| "P46098 saved forest has the selected feature width and settings", | |
| "serialized final forests reproduce saved test predictions", | |
| "focused notebook fully executed without errors", | |
| "notebook source table has current counts" | |
| ], | |
| "inputs": { | |
| "pairs.parquet": "0b5c46441b28d2a4667ade4394dc55f02c09c7177b26a83fa2cf2f19be66cd51", | |
| "structures.parquet": "088ed8b7f897bc71642c4fbb240feaa5aa23675d0b925cb7105bf47d6f48ccb8", | |
| "extra_features.npz": "229320e751b04f34146f3a5dd8b52294a91ffd9f278808495b214d44c9f686c7", | |
| "radius_fingerprints.npz": "d2bab9fd7bf84e4a67cf7c346e238302f8b3894dd1c685f5d21734fd1d6a27c4", | |
| "size_fingerprints.npz": "94467794e239ef8bfbc1180a20772a2320e01758194d15c634a6cb1ccb56898c" | |
| }, | |
| "final_baseline_sha256": "358caff4e034b83e825b4d5a3be015f8a12d956125eda069017743053e9883f7", | |
| "notebook_sha256": "702a0d865510d742b795091f7b943543c237f9d4f44812d3e9d1a4b836663303" | |
| } |