tinystories-diff-ar / refill_after.json
Asilarkness's picture
Upload refill_after.json with huggingface_hub
637bcd6 verified
Raw History Blame Contribute Delete
6.21 kB
[
{
"scheme": "AR (ceiling)",
"calls": 128,
"sec": 1.0452,
"tok_s": 7838,
"nll": 3.1424,
"dup": 0.0178,
"d2": 0.5274
},
{
"scheme": "DIFF K=2",
"calls": 2,
"sec": 0.0157,
"tok_s": 523421,
"nll": 10.0407,
"dup": 0.0527,
"d2": 0.572
},
{
"scheme": "DIFF K=4",
"calls": 4,
"sec": 0.0313,
"tok_s": 262010,
"nll": 8.7909,
"dup": 0.0812,
"d2": 0.4678
},
{
"scheme": "DIFF K=6",
"calls": 6,
"sec": 0.0465,
"tok_s": 176192,
"nll": 8.0376,
"dup": 0.0939,
"d2": 0.4289
},
{
"scheme": "DIFF K=8",
"calls": 8,
"sec": 0.0618,
"tok_s": 132584,
"nll": 7.2231,
"dup": 0.109,
"d2": 0.4155
},
{
"scheme": "DIFF K=10",
"calls": 10,
"sec": 0.0774,
"tok_s": 105821,
"nll": 6.7987,
"dup": 0.1099,
"d2": 0.4094
},
{
"scheme": "DIFF K=12",
"calls": 12,
"sec": 0.093,
"tok_s": 88091,
"nll": 6.4784,
"dup": 0.1078,
"d2": 0.4
},
{
"scheme": "DIFF K=16",
"calls": 16,
"sec": 0.1238,
"tok_s": 66169,
"nll": 5.6207,
"dup": 0.1022,
"d2": 0.4022
},
{
"scheme": "DIFF K=20",
"calls": 20,
"sec": 0.1544,
"tok_s": 53064,
"nll": 5.155,
"dup": 0.0926,
"d2": 0.3997
},
{
"scheme": "DIFF K=24",
"calls": 24,
"sec": 0.1853,
"tok_s": 44211,
"nll": 4.8638,
"dup": 0.0926,
"d2": 0.3845
},
{
"scheme": "K=4 p=0.10 R=1",
"calls": 5,
"sec": 0.0366,
"tok_s": 223586,
"nll": 8.5231,
"dup": 0.0865,
"d2": 0.3947
},
{
"scheme": "K=4 p=0.10 R=2",
"calls": 6,
"sec": 0.0426,
"tok_s": 192379,
"nll": 8.7187,
"dup": 0.1022,
"d2": 0.3377
},
{
"scheme": "K=4 p=0.10 R=4",
"calls": 8,
"sec": 0.0533,
"tok_s": 153642,
"nll": 8.514,
"dup": 0.1136,
"d2": 0.3132
},
{
"scheme": "K=4 p=0.20 R=1",
"calls": 5,
"sec": 0.0369,
"tok_s": 222143,
"nll": 8.5185,
"dup": 0.0965,
"d2": 0.3776
},
{
"scheme": "K=4 p=0.20 R=2",
"calls": 6,
"sec": 0.0426,
"tok_s": 192330,
"nll": 8.5694,
"dup": 0.1086,
"d2": 0.3488
},
{
"scheme": "K=4 p=0.20 R=4",
"calls": 8,
"sec": 0.0541,
"tok_s": 151502,
"nll": 8.6685,
"dup": 0.1394,
"d2": 0.3301
},
{
"scheme": "K=4 p=0.30 R=1",
"calls": 5,
"sec": 0.0369,
"tok_s": 221748,
"nll": 8.6454,
"dup": 0.1013,
"d2": 0.4038
},
{
"scheme": "K=4 p=0.30 R=2",
"calls": 6,
"sec": 0.0428,
"tok_s": 191300,
"nll": 8.7204,
"dup": 0.1121,
"d2": 0.3882
},
{
"scheme": "K=4 p=0.30 R=4",
"calls": 8,
"sec": 0.0547,
"tok_s": 149869,
"nll": 8.3218,
"dup": 0.1307,
"d2": 0.3942
},
{
"scheme": "K=4 p=0.50 R=1",
"calls": 5,
"sec": 0.0375,
"tok_s": 218543,
"nll": 8.9768,
"dup": 0.0934,
"d2": 0.4953
},
{
"scheme": "K=4 p=0.50 R=2",
"calls": 6,
"sec": 0.0439,
"tok_s": 186456,
"nll": 8.7903,
"dup": 0.0977,
"d2": 0.501
},
{
"scheme": "K=4 p=0.50 R=4",
"calls": 8,
"sec": 0.0569,
"tok_s": 143908,
"nll": 8.3455,
"dup": 0.1042,
"d2": 0.5014
},
{
"scheme": "K=8 p=0.10 R=1",
"calls": 9,
"sec": 0.0674,
"tok_s": 121568,
"nll": 6.9983,
"dup": 0.1026,
"d2": 0.3852
},
{
"scheme": "K=8 p=0.10 R=2",
"calls": 10,
"sec": 0.0729,
"tok_s": 112327,
"nll": 6.9742,
"dup": 0.1116,
"d2": 0.3607
},
{
"scheme": "K=8 p=0.10 R=4",
"calls": 12,
"sec": 0.0846,
"tok_s": 96874,
"nll": 7.1224,
"dup": 0.1453,
"d2": 0.3292
},
{
"scheme": "K=8 p=0.20 R=1",
"calls": 9,
"sec": 0.0673,
"tok_s": 121716,
"nll": 7.091,
"dup": 0.1207,
"d2": 0.3962
},
{
"scheme": "K=8 p=0.20 R=2",
"calls": 10,
"sec": 0.0734,
"tok_s": 111652,
"nll": 7.0672,
"dup": 0.116,
"d2": 0.3931
},
{
"scheme": "K=8 p=0.20 R=4",
"calls": 12,
"sec": 0.0849,
"tok_s": 96541,
"nll": 6.7981,
"dup": 0.1341,
"d2": 0.4005
},
{
"scheme": "K=8 p=0.30 R=1",
"calls": 9,
"sec": 0.0679,
"tok_s": 120619,
"nll": 7.3276,
"dup": 0.1115,
"d2": 0.4438
},
{
"scheme": "K=8 p=0.30 R=2",
"calls": 10,
"sec": 0.0736,
"tok_s": 111314,
"nll": 7.0889,
"dup": 0.1121,
"d2": 0.4508
},
{
"scheme": "K=8 p=0.30 R=4",
"calls": 12,
"sec": 0.0855,
"tok_s": 95868,
"nll": 6.7887,
"dup": 0.1107,
"d2": 0.4434
},
{
"scheme": "K=8 p=0.50 R=1",
"calls": 9,
"sec": 0.0686,
"tok_s": 119442,
"nll": 7.4452,
"dup": 0.0839,
"d2": 0.5279
},
{
"scheme": "K=8 p=0.50 R=2",
"calls": 10,
"sec": 0.0743,
"tok_s": 110253,
"nll": 7.3156,
"dup": 0.0866,
"d2": 0.5268
},
{
"scheme": "K=8 p=0.50 R=4",
"calls": 12,
"sec": 0.0877,
"tok_s": 93429,
"nll": 6.3702,
"dup": 0.065,
"d2": 0.5353
},
{
"scheme": "K=12 p=0.10 R=1",
"calls": 13,
"sec": 0.0979,
"tok_s": 83712,
"nll": 5.9573,
"dup": 0.1061,
"d2": 0.3799
},
{
"scheme": "K=12 p=0.10 R=2",
"calls": 14,
"sec": 0.1035,
"tok_s": 79123,
"nll": 6.101,
"dup": 0.1193,
"d2": 0.3604
},
{
"scheme": "K=12 p=0.10 R=4",
"calls": 16,
"sec": 0.1147,
"tok_s": 71408,
"nll": 6.2246,
"dup": 0.1292,
"d2": 0.3685
},
{
"scheme": "K=12 p=0.20 R=1",
"calls": 13,
"sec": 0.0977,
"tok_s": 83816,
"nll": 6.1959,
"dup": 0.111,
"d2": 0.4209
},
{
"scheme": "K=12 p=0.20 R=2",
"calls": 14,
"sec": 0.1035,
"tok_s": 79145,
"nll": 6.4067,
"dup": 0.1164,
"d2": 0.4234
},
{
"scheme": "K=12 p=0.20 R=4",
"calls": 16,
"sec": 0.1161,
"tok_s": 70548,
"nll": 6.126,
"dup": 0.1169,
"d2": 0.4194
},
{
"scheme": "K=12 p=0.30 R=1",
"calls": 13,
"sec": 0.0982,
"tok_s": 83443,
"nll": 6.7185,
"dup": 0.1059,
"d2": 0.4403
},
{
"scheme": "K=12 p=0.30 R=2",
"calls": 14,
"sec": 0.1029,
"tok_s": 79578,
"nll": 6.4778,
"dup": 0.0961,
"d2": 0.4703
},
{
"scheme": "K=12 p=0.30 R=4",
"calls": 16,
"sec": 0.1146,
"tok_s": 71475,
"nll": 6.0415,
"dup": 0.0966,
"d2": 0.4603
},
{
"scheme": "K=12 p=0.50 R=1",
"calls": 13,
"sec": 0.0993,
"tok_s": 82477,
"nll": 6.8466,
"dup": 0.064,
"d2": 0.545
},
{
"scheme": "K=12 p=0.50 R=2",
"calls": 14,
"sec": 0.1041,
"tok_s": 78689,
"nll": 6.3497,
"dup": 0.0611,
"d2": 0.5411
},
{
"scheme": "K=12 p=0.50 R=4",
"calls": 16,
"sec": 0.1166,
"tok_s": 70236,
"nll": 5.9313,
"dup": 0.057,
"d2": 0.5479
}
]