ens-appraiser / v0_4_metadata.json
quantumly's picture
v0.4 multi-embedding appraiser: 2026-04-26
22f4a87 verified
{
"trained_at": "2026-04-26T09:19:05.178649+00:00",
"data_run_date": "2026-04-25",
"version": "v0.4-multi-embedding",
"description": "v0.3 + bge-base-en-v1.5 alongside mpnet-finetuned",
"parent_version": "v0.3",
"embedders": [
"mpnet-finetuned (from v0.3)",
"BAAI/bge-base-en-v1.5"
],
"splits": {
"train": {
"rows": 265240,
"start": "2022-01-28",
"end": "2023-09-30"
},
"val": {
"rows": 3545,
"start": "2023-10-01",
"end": "2023-12-31"
},
"test": {
"rows": 2744,
"start": "2024-01-01",
"end": "2024-05-04"
}
},
"feature_count": 154,
"feature_cols": [
"pca_000",
"pca_001",
"pca_002",
"pca_003",
"pca_004",
"pca_005",
"pca_006",
"pca_007",
"pca_008",
"pca_009",
"pca_010",
"pca_011",
"pca_012",
"pca_013",
"pca_014",
"pca_015",
"pca_016",
"pca_017",
"pca_018",
"pca_019",
"pca_020",
"pca_021",
"pca_022",
"pca_023",
"pca_024",
"pca_025",
"pca_026",
"pca_027",
"pca_028",
"pca_029",
"pca_030",
"pca_031",
"pca_032",
"pca_033",
"pca_034",
"pca_035",
"pca_036",
"pca_037",
"pca_038",
"pca_039",
"pca_040",
"pca_041",
"pca_042",
"pca_043",
"pca_044",
"pca_045",
"pca_046",
"pca_047",
"pca_048",
"pca_049",
"pca_050",
"pca_051",
"pca_052",
"pca_053",
"pca_054",
"pca_055",
"pca_056",
"pca_057",
"pca_058",
"pca_059",
"pca_060",
"pca_061",
"pca_062",
"pca_063",
"len",
"n_digits",
"n_letters",
"n_special",
"n_lower",
"n_upper",
"is_palindrome",
"is_all_digits",
"is_all_letters",
"is_ascii",
"has_unicode",
"starts_digit",
"ends_digit",
"max_char_run",
"n_unique_chars",
"in_wikipedia",
"in_geonames",
"in_us_firstname",
"in_iso3166",
"in_ticker",
"in_sec_edgar",
"in_wiktionary_en",
"wordlist_hits",
"club__prepunk_full_rankings",
"club__social_handles",
"club__fine_art",
"club__performing_arts",
"club__gamertags_double",
"club__currency_symbols",
"club__logistics",
"club__gamertags",
"club__pokemon_gen1",
"club__common_english",
"club__top_crypto_tickers",
"club__common_animals",
"club__familynames_usa",
"club__natural_wonders",
"club__historic_figures",
"club__conspiracy_theories",
"club__mythical_creatures",
"club__top_nouns",
"club__finance_terms",
"club__personas",
"club__country_codes",
"club__catholicism",
"club__crypto_terms",
"club__sports",
"club__home",
"club__top500_cities_global",
"club__firstnames_usa",
"club__paranormal",
"club__wikidata_top_fantasy_char",
"club__crayola_classic",
"club__top500_cities_usa",
"club__top_crypto_names",
"club__us_government",
"club__pokemon_gen4",
"club__pokemon_gen2",
"club__memes",
"club__currency_names",
"club__luxury",
"club__pokemon_gen3",
"club__gen_alpha",
"club__us_states",
"club__holidays",
"n_clubs",
"trademark_conflict",
"name_age_days",
"prior_transfer_count",
"fg_value",
"eth_tvl_usd",
"eth_stable_mcap",
"eth_dex_volume",
"nft_total_fee_usd",
"knnmp_count",
"knnmp_mean_log",
"knnmp_median_log",
"knnmp_p90_log",
"knnmp_max_sim",
"knnmp_min_sim",
"knnmp_log_max",
"knnmp_log_min",
"knnbg_count",
"knnbg_mean_log",
"knnbg_median_log",
"knnbg_p90_log",
"knnbg_max_sim",
"knnbg_min_sim",
"knnbg_log_max",
"knnbg_log_min"
],
"pca_dim": 64,
"best_iteration": 757,
"xgb_params": {
"objective": "reg:squarederror",
"eval_metric": "rmse",
"tree_method": "hist",
"device": "cuda",
"max_depth": 7,
"learning_rate": 0.04,
"subsample": 0.85,
"colsample_bytree": 0.65,
"min_child_weight": 8,
"reg_alpha": 0.5,
"reg_lambda": 2.0,
"seed": 42
},
"metrics": {
"train": {
"r2_log": 0.830089807510376,
"rmse_log": 0.6656108498573303,
"mae_log": 0.44242581725120544,
"median_ape": 0.2753954231739044,
"bias_log": -0.00027043218142353
},
"val": {
"r2_log": 0.6791324615478516,
"rmse_log": 1.0376933813095093,
"mae_log": 0.7315031886100769,
"median_ape": 0.5113449096679688,
"bias_log": 0.10410919040441513
},
"test": {
"r2_log": 0.31753313541412354,
"rmse_log": 1.5363107919692993,
"mae_log": 1.2583258152008057,
"median_ape": 1.2974694967269897,
"bias_log": 0.7355491518974304
}
},
"top_features": [
{
"name": "knnmp_mean_log",
"gain": 1521.73291015625
},
{
"name": "knnmp_median_log",
"gain": 1161.393798828125
},
{
"name": "is_all_digits",
"gain": 576.3497314453125
},
{
"name": "knnmp_p90_log",
"gain": 492.0875244140625
},
{
"name": "len",
"gain": 380.18292236328125
},
{
"name": "in_wikipedia",
"gain": 365.04254150390625
},
{
"name": "n_clubs",
"gain": 194.0254669189453
},
{
"name": "knnmp_count",
"gain": 154.111572265625
},
{
"name": "ends_digit",
"gain": 137.3547821044922
},
{
"name": "trademark_conflict",
"gain": 134.1524658203125
},
{
"name": "n_letters",
"gain": 129.6276092529297
},
{
"name": "name_age_days",
"gain": 128.65174865722656
},
{
"name": "club__gamertags",
"gain": 128.5192108154297
},
{
"name": "is_palindrome",
"gain": 125.1925048828125
},
{
"name": "n_digits",
"gain": 125.02507019042969
},
{
"name": "club__social_handles",
"gain": 119.58012390136719
},
{
"name": "eth_stable_mcap",
"gain": 114.49725341796875
},
{
"name": "has_unicode",
"gain": 105.6874771118164
},
{
"name": "n_lower",
"gain": 102.0856704711914
},
{
"name": "n_special",
"gain": 89.66917419433594
},
{
"name": "eth_tvl_usd",
"gain": 89.55354309082031
},
{
"name": "club__prepunk_full_rankings",
"gain": 87.96327209472656
},
{
"name": "n_unique_chars",
"gain": 86.47875213623047
},
{
"name": "pca_045",
"gain": 85.28775787353516
},
{
"name": "is_ascii",
"gain": 83.36601257324219
},
{
"name": "starts_digit",
"gain": 73.62259674072266
},
{
"name": "knnbg_count",
"gain": 61.69795608520508
},
{
"name": "pca_001",
"gain": 60.780391693115234
},
{
"name": "fg_value",
"gain": 57.86528778076172
},
{
"name": "pca_026",
"gain": 57.13859939575195
}
],
"knn_gain_split": {
"mpnet_ft": 3443.8879318237305,
"bge": 257.9545383453369
},
"wandb_run": "https://wandb.ai/quantumly-aletheia-research/ens-appraiser/runs/osbosu7j"
}