Upload spellchecker evaluation
Browse files- spellcheck_report.json +219 -0
spellcheck_report.json
ADDED
|
@@ -0,0 +1,219 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"config": {
|
| 3 |
+
"model": "LiquidAI/LFM2.5-Encoder-350M-Spellchecker",
|
| 4 |
+
"output_dir": "/tmp/minima-spellcheck",
|
| 5 |
+
"output_repo": "ProCreations/minima-spellcheck",
|
| 6 |
+
"dataset": "HuggingFaceFW/fineweb-edu",
|
| 7 |
+
"dataset_config": "sample-10BT",
|
| 8 |
+
"sequence_length": 128,
|
| 9 |
+
"batch_size": 2,
|
| 10 |
+
"gradient_accumulation": 4,
|
| 11 |
+
"steps": 1000,
|
| 12 |
+
"learning_rate": 5e-05,
|
| 13 |
+
"group_size": 32,
|
| 14 |
+
"recovery_rank": 128,
|
| 15 |
+
"topk": 16,
|
| 16 |
+
"eval_batches": 8,
|
| 17 |
+
"seed": 77,
|
| 18 |
+
"log_every": 50
|
| 19 |
+
},
|
| 20 |
+
"duration_seconds": 394.26561307907104,
|
| 21 |
+
"trainable_parameters": 57020424,
|
| 22 |
+
"tagger_label_top1_candidate_agreement": 0.9951171875,
|
| 23 |
+
"detection_top1_agreement": 0.994140625,
|
| 24 |
+
"correction_exact_agreement": 0.7,
|
| 25 |
+
"correction_evaluation_examples": 20,
|
| 26 |
+
"example_exact_agreement": 1.0,
|
| 27 |
+
"examples": [
|
| 28 |
+
{
|
| 29 |
+
"input": "She go to school every day .",
|
| 30 |
+
"teacher": "She goes to school every day .",
|
| 31 |
+
"minima": "She goes to school every day ."
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"input": "I has went to the stor yesterday .",
|
| 35 |
+
"teacher": "I went to the store yesterday .",
|
| 36 |
+
"minima": "I went to the store yesterday ."
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"input": "Their are many reason to study hard .",
|
| 40 |
+
"teacher": "There are many reasons to study hard .",
|
| 41 |
+
"minima": "There are many reasons to study hard ."
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"input": "That 's a fair point , let 's discuss it tomorrow .",
|
| 45 |
+
"teacher": "That 's a fair point , let 's discuss it tomorrow .",
|
| 46 |
+
"minima": "That 's a fair point , let 's discuss it tomorrow ."
|
| 47 |
+
}
|
| 48 |
+
],
|
| 49 |
+
"history": [
|
| 50 |
+
{
|
| 51 |
+
"step": 1,
|
| 52 |
+
"elapsed_s": 9.728960990905762,
|
| 53 |
+
"loss": 2.4879839420318604,
|
| 54 |
+
"label": 1.3699460625648499,
|
| 55 |
+
"detect": 0.3631512150168419,
|
| 56 |
+
"hidden": 0.4682311415672302
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"step": 51,
|
| 60 |
+
"elapsed_s": 47.39888334274292,
|
| 61 |
+
"loss": 0.5646287947893143,
|
| 62 |
+
"label": 0.21763740107417107,
|
| 63 |
+
"detect": 0.027677168138325214,
|
| 64 |
+
"hidden": 0.16657640412449837
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"step": 101,
|
| 68 |
+
"elapsed_s": 64.47661709785461,
|
| 69 |
+
"loss": 0.433330237865448,
|
| 70 |
+
"label": 0.14443234354257584,
|
| 71 |
+
"detect": 0.012719596037641168,
|
| 72 |
+
"hidden": 0.1412690468132496
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"step": 151,
|
| 76 |
+
"elapsed_s": 81.66340661048889,
|
| 77 |
+
"loss": 0.374489426612854,
|
| 78 |
+
"label": 0.12132998369634151,
|
| 79 |
+
"detect": 0.0135470035020262,
|
| 80 |
+
"hidden": 0.12319297343492508
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"step": 201,
|
| 84 |
+
"elapsed_s": 98.94163203239441,
|
| 85 |
+
"loss": 0.38375477492809296,
|
| 86 |
+
"label": 0.12871305830776691,
|
| 87 |
+
"detect": 0.015309339622035623,
|
| 88 |
+
"hidden": 0.12369352579116821
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"step": 251,
|
| 92 |
+
"elapsed_s": 116.117196559906,
|
| 93 |
+
"loss": 0.4129961282014847,
|
| 94 |
+
"label": 0.16921869292855263,
|
| 95 |
+
"detect": 0.018067472614347935,
|
| 96 |
+
"hidden": 0.11737184971570969
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"step": 301,
|
| 100 |
+
"elapsed_s": 133.41278839111328,
|
| 101 |
+
"loss": 0.3421344459056854,
|
| 102 |
+
"label": 0.12530853413045406,
|
| 103 |
+
"detect": 0.008865324780344963,
|
| 104 |
+
"hidden": 0.10619662329554558
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"step": 351,
|
| 108 |
+
"elapsed_s": 150.6108753681183,
|
| 109 |
+
"loss": 0.4015732929110527,
|
| 110 |
+
"label": 0.15202416479587555,
|
| 111 |
+
"detect": 0.021410188637673855,
|
| 112 |
+
"hidden": 0.11942201666533947
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"step": 401,
|
| 116 |
+
"elapsed_s": 167.74786019325256,
|
| 117 |
+
"loss": 0.3250713273882866,
|
| 118 |
+
"label": 0.11101343296468258,
|
| 119 |
+
"detect": 0.014453576179221272,
|
| 120 |
+
"hidden": 0.10341555811464787
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"step": 451,
|
| 124 |
+
"elapsed_s": 184.91804909706116,
|
| 125 |
+
"loss": 0.30887674540281296,
|
| 126 |
+
"label": 0.1138718631118536,
|
| 127 |
+
"detect": 0.009780762484297156,
|
| 128 |
+
"hidden": 0.09505725093185902
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"step": 501,
|
| 132 |
+
"elapsed_s": 201.9468400478363,
|
| 133 |
+
"loss": 0.2911076098680496,
|
| 134 |
+
"label": 0.093521011993289,
|
| 135 |
+
"detect": 0.011887337197549641,
|
| 136 |
+
"hidden": 0.09582146629691124
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"step": 551,
|
| 140 |
+
"elapsed_s": 218.95755100250244,
|
| 141 |
+
"loss": 0.3279508873820305,
|
| 142 |
+
"label": 0.12233076430857182,
|
| 143 |
+
"detect": 0.012748322915285826,
|
| 144 |
+
"hidden": 0.09962298162281513
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"step": 601,
|
| 148 |
+
"elapsed_s": 235.95491790771484,
|
| 149 |
+
"loss": 0.2837226167321205,
|
| 150 |
+
"label": 0.08802595920860767,
|
| 151 |
+
"detect": 0.006528681959025562,
|
| 152 |
+
"hidden": 0.09621615894138813
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"step": 651,
|
| 156 |
+
"elapsed_s": 252.9777970314026,
|
| 157 |
+
"loss": 0.29664454981684685,
|
| 158 |
+
"label": 0.1072889119386673,
|
| 159 |
+
"detect": 0.010023394948802888,
|
| 160 |
+
"hidden": 0.09217197075486183
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"step": 701,
|
| 164 |
+
"elapsed_s": 270.00525283813477,
|
| 165 |
+
"loss": 0.31047341972589493,
|
| 166 |
+
"label": 0.11355133540928364,
|
| 167 |
+
"detect": 0.01577813387848437,
|
| 168 |
+
"hidden": 0.09451650828123093
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"step": 751,
|
| 172 |
+
"elapsed_s": 287.11339259147644,
|
| 173 |
+
"loss": 0.2924303859472275,
|
| 174 |
+
"label": 0.09383961372077465,
|
| 175 |
+
"detect": 0.008933465695008636,
|
| 176 |
+
"hidden": 0.09706202149391174
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"step": 801,
|
| 180 |
+
"elapsed_s": 304.38399958610535,
|
| 181 |
+
"loss": 0.26952020823955536,
|
| 182 |
+
"label": 0.08447664976119995,
|
| 183 |
+
"detect": 0.01343391730915755,
|
| 184 |
+
"hidden": 0.08916330337524414
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"step": 851,
|
| 188 |
+
"elapsed_s": 321.6072781085968,
|
| 189 |
+
"loss": 0.27088215947151184,
|
| 190 |
+
"label": 0.08796967100352049,
|
| 191 |
+
"detect": 0.008289208286441863,
|
| 192 |
+
"hidden": 0.08938394114375114
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"step": 901,
|
| 196 |
+
"elapsed_s": 338.75015807151794,
|
| 197 |
+
"loss": 0.2826335057616234,
|
| 198 |
+
"label": 0.09538542665541172,
|
| 199 |
+
"detect": 0.010135660180822015,
|
| 200 |
+
"hidden": 0.0910901241004467
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"step": 951,
|
| 204 |
+
"elapsed_s": 355.75725746154785,
|
| 205 |
+
"loss": 0.2881084829568863,
|
| 206 |
+
"label": 0.09700636379420757,
|
| 207 |
+
"detect": 0.014985204674303532,
|
| 208 |
+
"hidden": 0.0918047558516264
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"step": 1000,
|
| 212 |
+
"elapsed_s": 372.5106551647186,
|
| 213 |
+
"loss": 0.3113468959927559,
|
| 214 |
+
"label": 0.1258483398705721,
|
| 215 |
+
"detect": 0.012814029469154775,
|
| 216 |
+
"hidden": 0.08954576589167118
|
| 217 |
+
}
|
| 218 |
+
]
|
| 219 |
+
}
|