tinystories-diff-ar / refill_before.json
Asilarkness's picture
Upload refill_before.json with huggingface_hub
bb58814 verified
Raw
History Blame Contribute Delete
6.2 kB
[
{
"scheme": "AR (ceiling)",
"calls": 128,
"sec": 0.3348,
"tok_s": 24466,
"nll": 2.8986,
"dup": 0.0201,
"d2": 0.5169
},
{
"scheme": "DIFF K=2",
"calls": 2,
"sec": 0.0157,
"tok_s": 523289,
"nll": 8.133,
"dup": 0.0506,
"d2": 0.5904
},
{
"scheme": "DIFF K=4",
"calls": 4,
"sec": 0.0311,
"tok_s": 263707,
"nll": 5.7442,
"dup": 0.0561,
"d2": 0.4779
},
{
"scheme": "DIFF K=6",
"calls": 6,
"sec": 0.0464,
"tok_s": 176517,
"nll": 4.9298,
"dup": 0.0563,
"d2": 0.445
},
{
"scheme": "DIFF K=8",
"calls": 8,
"sec": 0.062,
"tok_s": 132049,
"nll": 4.4687,
"dup": 0.056,
"d2": 0.4438
},
{
"scheme": "DIFF K=10",
"calls": 10,
"sec": 0.0773,
"tok_s": 105944,
"nll": 4.1071,
"dup": 0.0465,
"d2": 0.4221
},
{
"scheme": "DIFF K=12",
"calls": 12,
"sec": 0.0931,
"tok_s": 88001,
"nll": 3.9702,
"dup": 0.0471,
"d2": 0.4243
},
{
"scheme": "DIFF K=16",
"calls": 16,
"sec": 0.1235,
"tok_s": 66347,
"nll": 3.4351,
"dup": 0.0429,
"d2": 0.4074
},
{
"scheme": "DIFF K=20",
"calls": 20,
"sec": 0.1546,
"tok_s": 52982,
"nll": 3.263,
"dup": 0.0395,
"d2": 0.4133
},
{
"scheme": "DIFF K=24",
"calls": 24,
"sec": 0.1852,
"tok_s": 44228,
"nll": 3.163,
"dup": 0.038,
"d2": 0.4054
},
{
"scheme": "K=4 p=0.10 R=1",
"calls": 5,
"sec": 0.0362,
"tok_s": 226225,
"nll": 4.9697,
"dup": 0.055,
"d2": 0.4266
},
{
"scheme": "K=4 p=0.10 R=2",
"calls": 6,
"sec": 0.0416,
"tok_s": 196999,
"nll": 4.7276,
"dup": 0.0528,
"d2": 0.4206
},
{
"scheme": "K=4 p=0.10 R=4",
"calls": 8,
"sec": 0.0536,
"tok_s": 152814,
"nll": 4.4119,
"dup": 0.0523,
"d2": 0.4216
},
{
"scheme": "K=4 p=0.20 R=1",
"calls": 5,
"sec": 0.0369,
"tok_s": 222019,
"nll": 4.7627,
"dup": 0.0517,
"d2": 0.4381
},
{
"scheme": "K=4 p=0.20 R=2",
"calls": 6,
"sec": 0.0422,
"tok_s": 193955,
"nll": 4.3189,
"dup": 0.0527,
"d2": 0.4269
},
{
"scheme": "K=4 p=0.20 R=4",
"calls": 8,
"sec": 0.0536,
"tok_s": 152843,
"nll": 3.8024,
"dup": 0.0434,
"d2": 0.4211
},
{
"scheme": "K=4 p=0.30 R=1",
"calls": 5,
"sec": 0.037,
"tok_s": 221551,
"nll": 4.9045,
"dup": 0.0538,
"d2": 0.443
},
{
"scheme": "K=4 p=0.30 R=2",
"calls": 6,
"sec": 0.0428,
"tok_s": 191389,
"nll": 4.363,
"dup": 0.0444,
"d2": 0.4392
},
{
"scheme": "K=4 p=0.30 R=4",
"calls": 8,
"sec": 0.0547,
"tok_s": 149880,
"nll": 3.6176,
"dup": 0.0394,
"d2": 0.4323
},
{
"scheme": "K=4 p=0.50 R=1",
"calls": 5,
"sec": 0.0378,
"tok_s": 216889,
"nll": 4.8723,
"dup": 0.0485,
"d2": 0.4713
},
{
"scheme": "K=4 p=0.50 R=2",
"calls": 6,
"sec": 0.0438,
"tok_s": 187073,
"nll": 4.5311,
"dup": 0.0437,
"d2": 0.4734
},
{
"scheme": "K=4 p=0.50 R=4",
"calls": 8,
"sec": 0.0562,
"tok_s": 145658,
"nll": 4.8553,
"dup": 0.0418,
"d2": 0.4957
},
{
"scheme": "K=8 p=0.10 R=1",
"calls": 9,
"sec": 0.0673,
"tok_s": 121746,
"nll": 3.861,
"dup": 0.0493,
"d2": 0.4026
},
{
"scheme": "K=8 p=0.10 R=2",
"calls": 10,
"sec": 0.0728,
"tok_s": 112487,
"nll": 3.7104,
"dup": 0.0493,
"d2": 0.399
},
{
"scheme": "K=8 p=0.10 R=4",
"calls": 12,
"sec": 0.0849,
"tok_s": 96435,
"nll": 3.4375,
"dup": 0.0404,
"d2": 0.4183
},
{
"scheme": "K=8 p=0.20 R=1",
"calls": 9,
"sec": 0.0675,
"tok_s": 121308,
"nll": 3.944,
"dup": 0.0463,
"d2": 0.4186
},
{
"scheme": "K=8 p=0.20 R=2",
"calls": 10,
"sec": 0.073,
"tok_s": 112265,
"nll": 3.5781,
"dup": 0.0389,
"d2": 0.4203
},
{
"scheme": "K=8 p=0.20 R=4",
"calls": 12,
"sec": 0.0858,
"tok_s": 95522,
"nll": 3.0231,
"dup": 0.0358,
"d2": 0.4088
},
{
"scheme": "K=8 p=0.30 R=1",
"calls": 9,
"sec": 0.0671,
"tok_s": 122176,
"nll": 3.9184,
"dup": 0.0437,
"d2": 0.4215
},
{
"scheme": "K=8 p=0.30 R=2",
"calls": 10,
"sec": 0.0734,
"tok_s": 111555,
"nll": 3.5091,
"dup": 0.042,
"d2": 0.4306
},
{
"scheme": "K=8 p=0.30 R=4",
"calls": 12,
"sec": 0.0853,
"tok_s": 96012,
"nll": 3.1743,
"dup": 0.0348,
"d2": 0.4213
},
{
"scheme": "K=8 p=0.50 R=1",
"calls": 9,
"sec": 0.0675,
"tok_s": 121402,
"nll": 4.1907,
"dup": 0.0396,
"d2": 0.46
},
{
"scheme": "K=8 p=0.50 R=2",
"calls": 10,
"sec": 0.0738,
"tok_s": 110947,
"nll": 4.1945,
"dup": 0.0322,
"d2": 0.4787
},
{
"scheme": "K=8 p=0.50 R=4",
"calls": 12,
"sec": 0.0874,
"tok_s": 93724,
"nll": 4.6549,
"dup": 0.0386,
"d2": 0.4943
},
{
"scheme": "K=12 p=0.10 R=1",
"calls": 13,
"sec": 0.0968,
"tok_s": 84625,
"nll": 3.6755,
"dup": 0.0455,
"d2": 0.4082
},
{
"scheme": "K=12 p=0.10 R=2",
"calls": 14,
"sec": 0.1025,
"tok_s": 79925,
"nll": 3.3101,
"dup": 0.0402,
"d2": 0.4221
},
{
"scheme": "K=12 p=0.10 R=4",
"calls": 16,
"sec": 0.1138,
"tok_s": 71993,
"nll": 3.1429,
"dup": 0.0397,
"d2": 0.4126
},
{
"scheme": "K=12 p=0.20 R=1",
"calls": 13,
"sec": 0.0978,
"tok_s": 83796,
"nll": 3.562,
"dup": 0.0429,
"d2": 0.4163
},
{
"scheme": "K=12 p=0.20 R=2",
"calls": 14,
"sec": 0.1052,
"tok_s": 77834,
"nll": 3.1824,
"dup": 0.0356,
"d2": 0.422
},
{
"scheme": "K=12 p=0.20 R=4",
"calls": 16,
"sec": 0.1139,
"tok_s": 71914,
"nll": 2.8586,
"dup": 0.0288,
"d2": 0.4162
},
{
"scheme": "K=12 p=0.30 R=1",
"calls": 13,
"sec": 0.0972,
"tok_s": 84300,
"nll": 3.5599,
"dup": 0.0378,
"d2": 0.4332
},
{
"scheme": "K=12 p=0.30 R=2",
"calls": 14,
"sec": 0.1041,
"tok_s": 78684,
"nll": 3.2104,
"dup": 0.0338,
"d2": 0.4418
},
{
"scheme": "K=12 p=0.30 R=4",
"calls": 16,
"sec": 0.1148,
"tok_s": 71389,
"nll": 3.172,
"dup": 0.0325,
"d2": 0.4182
},
{
"scheme": "K=12 p=0.50 R=1",
"calls": 13,
"sec": 0.098,
"tok_s": 83601,
"nll": 4.0057,
"dup": 0.0375,
"d2": 0.4654
},
{
"scheme": "K=12 p=0.50 R=2",
"calls": 14,
"sec": 0.1043,
"tok_s": 78556,
"nll": 4.1454,
"dup": 0.032,
"d2": 0.4737
},
{
"scheme": "K=12 p=0.50 R=4",
"calls": 16,
"sec": 0.1171,
"tok_s": 69979,
"nll": 4.5204,
"dup": 0.0357,
"d2": 0.4918
}
]