Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +606 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,606 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9992587101556709,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 337,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.014825796886582653,
|
| 13 |
+
"grad_norm": 3.8051848965817707,
|
| 14 |
+
"learning_rate": 2.9411764705882355e-06,
|
| 15 |
+
"loss": 0.956,
|
| 16 |
+
"mean_token_accuracy": 0.7389732335227731,
|
| 17 |
+
"step": 5
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.029651593773165306,
|
| 21 |
+
"grad_norm": 2.3193458934516125,
|
| 22 |
+
"learning_rate": 5.882352941176471e-06,
|
| 23 |
+
"loss": 0.9131,
|
| 24 |
+
"mean_token_accuracy": 0.745571190628745,
|
| 25 |
+
"step": 10
|
| 26 |
+
},
|
| 27 |
+
{
|
| 28 |
+
"epoch": 0.04447739065974796,
|
| 29 |
+
"grad_norm": 1.465130510110654,
|
| 30 |
+
"learning_rate": 8.823529411764707e-06,
|
| 31 |
+
"loss": 0.802,
|
| 32 |
+
"mean_token_accuracy": 0.7669543650793653,
|
| 33 |
+
"step": 15
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.05930318754633061,
|
| 37 |
+
"grad_norm": 1.1649028225710198,
|
| 38 |
+
"learning_rate": 1.1764705882352942e-05,
|
| 39 |
+
"loss": 0.7568,
|
| 40 |
+
"mean_token_accuracy": 0.7719902319902323,
|
| 41 |
+
"step": 20
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.07412898443291327,
|
| 45 |
+
"grad_norm": 0.8613760026054987,
|
| 46 |
+
"learning_rate": 1.4705882352941179e-05,
|
| 47 |
+
"loss": 0.7119,
|
| 48 |
+
"mean_token_accuracy": 0.7831028693528694,
|
| 49 |
+
"step": 25
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 0.08895478131949593,
|
| 53 |
+
"grad_norm": 0.783478820047515,
|
| 54 |
+
"learning_rate": 1.7647058823529414e-05,
|
| 55 |
+
"loss": 0.6953,
|
| 56 |
+
"mean_token_accuracy": 0.7864308608058606,
|
| 57 |
+
"step": 30
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.10378057820607858,
|
| 61 |
+
"grad_norm": 0.845141787895257,
|
| 62 |
+
"learning_rate": 1.9999462497359468e-05,
|
| 63 |
+
"loss": 0.6841,
|
| 64 |
+
"mean_token_accuracy": 0.7876358363858365,
|
| 65 |
+
"step": 35
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.11860637509266123,
|
| 69 |
+
"grad_norm": 0.7223548858937856,
|
| 70 |
+
"learning_rate": 1.9980655971335944e-05,
|
| 71 |
+
"loss": 0.6674,
|
| 72 |
+
"mean_token_accuracy": 0.7906836800335666,
|
| 73 |
+
"step": 40
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.1334321719792439,
|
| 77 |
+
"grad_norm": 0.8020026503983825,
|
| 78 |
+
"learning_rate": 1.993503206718859e-05,
|
| 79 |
+
"loss": 0.6618,
|
| 80 |
+
"mean_token_accuracy": 0.7926633089133089,
|
| 81 |
+
"step": 45
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.14825796886582654,
|
| 85 |
+
"grad_norm": 0.7215498275444733,
|
| 86 |
+
"learning_rate": 1.986271337340182e-05,
|
| 87 |
+
"loss": 0.6421,
|
| 88 |
+
"mean_token_accuracy": 0.797378663003663,
|
| 89 |
+
"step": 50
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 0.16308376575240918,
|
| 93 |
+
"grad_norm": 0.7852660604263051,
|
| 94 |
+
"learning_rate": 1.976389420563607e-05,
|
| 95 |
+
"loss": 0.646,
|
| 96 |
+
"mean_token_accuracy": 0.796246372287669,
|
| 97 |
+
"step": 55
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"epoch": 0.17790956263899185,
|
| 101 |
+
"grad_norm": 0.7068328201097145,
|
| 102 |
+
"learning_rate": 1.9638840084614182e-05,
|
| 103 |
+
"loss": 0.6349,
|
| 104 |
+
"mean_token_accuracy": 0.7993047924297925,
|
| 105 |
+
"step": 60
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.1927353595255745,
|
| 109 |
+
"grad_norm": 0.6758929995809626,
|
| 110 |
+
"learning_rate": 1.9487887022684336e-05,
|
| 111 |
+
"loss": 0.6163,
|
| 112 |
+
"mean_token_accuracy": 0.8042696886446887,
|
| 113 |
+
"step": 65
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.20756115641215717,
|
| 117 |
+
"grad_norm": 0.7716624972063456,
|
| 118 |
+
"learning_rate": 1.9311440620976597e-05,
|
| 119 |
+
"loss": 0.6319,
|
| 120 |
+
"mean_token_accuracy": 0.7993212087850436,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.2223869532987398,
|
| 125 |
+
"grad_norm": 0.6609819990400844,
|
| 126 |
+
"learning_rate": 1.9109974979578852e-05,
|
| 127 |
+
"loss": 0.6356,
|
| 128 |
+
"mean_token_accuracy": 0.798153998778999,
|
| 129 |
+
"step": 75
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.23721275018532245,
|
| 133 |
+
"grad_norm": 0.7028168784009874,
|
| 134 |
+
"learning_rate": 1.8884031423660492e-05,
|
| 135 |
+
"loss": 0.6337,
|
| 136 |
+
"mean_token_accuracy": 0.7985944050496857,
|
| 137 |
+
"step": 80
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.2520385470719051,
|
| 141 |
+
"grad_norm": 0.6683430701808383,
|
| 142 |
+
"learning_rate": 1.8634217048966638e-05,
|
| 143 |
+
"loss": 0.6353,
|
| 144 |
+
"mean_token_accuracy": 0.7978753602810393,
|
| 145 |
+
"step": 85
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 0.2668643439584878,
|
| 149 |
+
"grad_norm": 0.7314922030102382,
|
| 150 |
+
"learning_rate": 1.836120309059107e-05,
|
| 151 |
+
"loss": 0.6115,
|
| 152 |
+
"mean_token_accuracy": 0.8045245726495727,
|
| 153 |
+
"step": 90
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.28169014084507044,
|
| 157 |
+
"grad_norm": 0.6365309024889941,
|
| 158 |
+
"learning_rate": 1.8065723119410885e-05,
|
| 159 |
+
"loss": 0.6232,
|
| 160 |
+
"mean_token_accuracy": 0.8015613553113555,
|
| 161 |
+
"step": 95
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 0.2965159377316531,
|
| 165 |
+
"grad_norm": 0.6672244460635505,
|
| 166 |
+
"learning_rate": 1.77485710710289e-05,
|
| 167 |
+
"loss": 0.6255,
|
| 168 |
+
"mean_token_accuracy": 0.8006891025641025,
|
| 169 |
+
"step": 100
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 0.2965159377316531,
|
| 173 |
+
"eval_loss": 0.6413724422454834,
|
| 174 |
+
"eval_mean_token_accuracy": 0.795018613493309,
|
| 175 |
+
"eval_runtime": 5.4348,
|
| 176 |
+
"eval_samples_per_second": 23.552,
|
| 177 |
+
"eval_steps_per_second": 1.472,
|
| 178 |
+
"step": 100
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.3113417346182357,
|
| 182 |
+
"grad_norm": 0.75128092844917,
|
| 183 |
+
"learning_rate": 1.741059911251997e-05,
|
| 184 |
+
"loss": 0.6223,
|
| 185 |
+
"mean_token_accuracy": 0.8012797619047619,
|
| 186 |
+
"step": 105
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 0.32616753150481836,
|
| 190 |
+
"grad_norm": 0.6526755539743339,
|
| 191 |
+
"learning_rate": 1.7052715352713076e-05,
|
| 192 |
+
"loss": 0.6125,
|
| 193 |
+
"mean_token_accuracy": 0.8042971611721612,
|
| 194 |
+
"step": 110
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.34099332839140106,
|
| 198 |
+
"grad_norm": 0.5958813027219968,
|
| 199 |
+
"learning_rate": 1.667588140216154e-05,
|
| 200 |
+
"loss": 0.6241,
|
| 201 |
+
"mean_token_accuracy": 0.8009714590964588,
|
| 202 |
+
"step": 115
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.3558191252779837,
|
| 206 |
+
"grad_norm": 0.6459938604110328,
|
| 207 |
+
"learning_rate": 1.628110978935756e-05,
|
| 208 |
+
"loss": 0.6214,
|
| 209 |
+
"mean_token_accuracy": 0.8017719780219782,
|
| 210 |
+
"step": 120
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 0.37064492216456635,
|
| 214 |
+
"grad_norm": 0.7336206637425139,
|
| 215 |
+
"learning_rate": 1.586946124013354e-05,
|
| 216 |
+
"loss": 0.6228,
|
| 217 |
+
"mean_token_accuracy": 0.8008722527472527,
|
| 218 |
+
"step": 125
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.385470719051149,
|
| 222 |
+
"grad_norm": 0.6924708455089922,
|
| 223 |
+
"learning_rate": 1.5442041827560274e-05,
|
| 224 |
+
"loss": 0.6093,
|
| 225 |
+
"mean_token_accuracy": 0.8049916056166057,
|
| 226 |
+
"step": 130
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.40029651593773163,
|
| 230 |
+
"grad_norm": 0.6252828305378912,
|
| 231 |
+
"learning_rate": 1.5000000000000002e-05,
|
| 232 |
+
"loss": 0.6119,
|
| 233 |
+
"mean_token_accuracy": 0.8043566849816848,
|
| 234 |
+
"step": 135
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.41512231282431433,
|
| 238 |
+
"grad_norm": 0.5543639102481284,
|
| 239 |
+
"learning_rate": 1.4544523495299843e-05,
|
| 240 |
+
"loss": 0.6184,
|
| 241 |
+
"mean_token_accuracy": 0.8021657509157512,
|
| 242 |
+
"step": 140
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.429948109710897,
|
| 246 |
+
"grad_norm": 0.6721392797389169,
|
| 247 |
+
"learning_rate": 1.4076836149416889e-05,
|
| 248 |
+
"loss": 0.6025,
|
| 249 |
+
"mean_token_accuracy": 0.8070421245421245,
|
| 250 |
+
"step": 145
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.4447739065974796,
|
| 254 |
+
"grad_norm": 0.7061921312824664,
|
| 255 |
+
"learning_rate": 1.3598194608050011e-05,
|
| 256 |
+
"loss": 0.5859,
|
| 257 |
+
"mean_token_accuracy": 0.8116018009768011,
|
| 258 |
+
"step": 150
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 0.45959970348406226,
|
| 262 |
+
"grad_norm": 0.5295415190053663,
|
| 263 |
+
"learning_rate": 1.3109884950114007e-05,
|
| 264 |
+
"loss": 0.6101,
|
| 265 |
+
"mean_token_accuracy": 0.8041674297924299,
|
| 266 |
+
"step": 155
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.4744255003706449,
|
| 270 |
+
"grad_norm": 0.5162022426859912,
|
| 271 |
+
"learning_rate": 1.2613219232128608e-05,
|
| 272 |
+
"loss": 0.5968,
|
| 273 |
+
"mean_token_accuracy": 0.8079342185592185,
|
| 274 |
+
"step": 160
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 0.4892512972572276,
|
| 278 |
+
"grad_norm": 0.5740077528394586,
|
| 279 |
+
"learning_rate": 1.2109531962807333e-05,
|
| 280 |
+
"loss": 0.6052,
|
| 281 |
+
"mean_token_accuracy": 0.8052155664205085,
|
| 282 |
+
"step": 165
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.5040770941438102,
|
| 286 |
+
"grad_norm": 0.5784618789845658,
|
| 287 |
+
"learning_rate": 1.1600176517318742e-05,
|
| 288 |
+
"loss": 0.598,
|
| 289 |
+
"mean_token_accuracy": 0.8077052808302808,
|
| 290 |
+
"step": 170
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.5189028910303929,
|
| 294 |
+
"grad_norm": 0.5622095615626541,
|
| 295 |
+
"learning_rate": 1.1086521500854746e-05,
|
| 296 |
+
"loss": 0.5928,
|
| 297 |
+
"mean_token_accuracy": 0.808961385836386,
|
| 298 |
+
"step": 175
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 0.5337286879169756,
|
| 302 |
+
"grad_norm": 0.555421977804962,
|
| 303 |
+
"learning_rate": 1.0569947071276847e-05,
|
| 304 |
+
"loss": 0.6079,
|
| 305 |
+
"mean_token_accuracy": 0.8038957570207568,
|
| 306 |
+
"step": 180
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 0.5485544848035582,
|
| 310 |
+
"grad_norm": 0.5591254267329707,
|
| 311 |
+
"learning_rate": 1.0051841230721065e-05,
|
| 312 |
+
"loss": 0.5996,
|
| 313 |
+
"mean_token_accuracy": 0.8062667887667889,
|
| 314 |
+
"step": 185
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.5633802816901409,
|
| 318 |
+
"grad_norm": 0.5480628578430322,
|
| 319 |
+
"learning_rate": 9.533596096125826e-06,
|
| 320 |
+
"loss": 0.584,
|
| 321 |
+
"mean_token_accuracy": 0.812239774114774,
|
| 322 |
+
"step": 190
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 0.5782060785767235,
|
| 326 |
+
"grad_norm": 0.5514593108031277,
|
| 327 |
+
"learning_rate": 9.016604158703654e-06,
|
| 328 |
+
"loss": 0.6084,
|
| 329 |
+
"mean_token_accuracy": 0.8041910866910866,
|
| 330 |
+
"step": 195
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 0.5930318754633062,
|
| 334 |
+
"grad_norm": 0.5363039052592674,
|
| 335 |
+
"learning_rate": 8.502254542407186e-06,
|
| 336 |
+
"loss": 0.6039,
|
| 337 |
+
"mean_token_accuracy": 0.8054059829059831,
|
| 338 |
+
"step": 200
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.5930318754633062,
|
| 342 |
+
"eval_loss": 0.6165258288383484,
|
| 343 |
+
"eval_mean_token_accuracy": 0.8018075907910054,
|
| 344 |
+
"eval_runtime": 5.2218,
|
| 345 |
+
"eval_samples_per_second": 24.513,
|
| 346 |
+
"eval_steps_per_second": 1.532,
|
| 347 |
+
"step": 200
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.6078576723498889,
|
| 351 |
+
"grad_norm": 0.5869561564258384,
|
| 352 |
+
"learning_rate": 7.991929271442817e-06,
|
| 353 |
+
"loss": 0.5845,
|
| 354 |
+
"mean_token_accuracy": 0.810370879120879,
|
| 355 |
+
"step": 205
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 0.6226834692364714,
|
| 359 |
+
"grad_norm": 0.5126220770273285,
|
| 360 |
+
"learning_rate": 7.48699955686089e-06,
|
| 361 |
+
"loss": 0.6053,
|
| 362 |
+
"mean_token_accuracy": 0.8049973686472104,
|
| 363 |
+
"step": 210
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.6375092661230541,
|
| 367 |
+
"grad_norm": 0.5283766104356306,
|
| 368 |
+
"learning_rate": 6.988822112200157e-06,
|
| 369 |
+
"loss": 0.5809,
|
| 370 |
+
"mean_token_accuracy": 0.8124908424908428,
|
| 371 |
+
"step": 215
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.6523350630096367,
|
| 375 |
+
"grad_norm": 0.5538487666398877,
|
| 376 |
+
"learning_rate": 6.498735508086094e-06,
|
| 377 |
+
"loss": 0.6041,
|
| 378 |
+
"mean_token_accuracy": 0.8046634615384616,
|
| 379 |
+
"step": 220
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 0.6671608598962194,
|
| 383 |
+
"grad_norm": 0.502142536336529,
|
| 384 |
+
"learning_rate": 6.018056575578075e-06,
|
| 385 |
+
"loss": 0.5831,
|
| 386 |
+
"mean_token_accuracy": 0.8119062881562883,
|
| 387 |
+
"step": 225
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.6819866567828021,
|
| 391 |
+
"grad_norm": 0.5468779254684961,
|
| 392 |
+
"learning_rate": 5.548076867929331e-06,
|
| 393 |
+
"loss": 0.5938,
|
| 394 |
+
"mean_token_accuracy": 0.8081074481074481,
|
| 395 |
+
"step": 230
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.6968124536693847,
|
| 399 |
+
"grad_norm": 0.4985379501273117,
|
| 400 |
+
"learning_rate": 5.090059190266779e-06,
|
| 401 |
+
"loss": 0.5963,
|
| 402 |
+
"mean_token_accuracy": 0.8077899877899876,
|
| 403 |
+
"step": 235
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 0.7116382505559674,
|
| 407 |
+
"grad_norm": 0.5151452293487124,
|
| 408 |
+
"learning_rate": 4.645234206515171e-06,
|
| 409 |
+
"loss": 0.5899,
|
| 410 |
+
"mean_token_accuracy": 0.8092720264856231,
|
| 411 |
+
"step": 240
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.72646404744255,
|
| 415 |
+
"grad_norm": 0.4949458523296326,
|
| 416 |
+
"learning_rate": 4.214797132682597e-06,
|
| 417 |
+
"loss": 0.5921,
|
| 418 |
+
"mean_token_accuracy": 0.8082945650477681,
|
| 419 |
+
"step": 245
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.7412898443291327,
|
| 423 |
+
"grad_norm": 0.47351353980603134,
|
| 424 |
+
"learning_rate": 3.799904525392251e-06,
|
| 425 |
+
"loss": 0.5906,
|
| 426 |
+
"mean_token_accuracy": 0.8085080891330888,
|
| 427 |
+
"step": 250
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"epoch": 0.7561156412157154,
|
| 431 |
+
"grad_norm": 0.48563113738070895,
|
| 432 |
+
"learning_rate": 3.401671174289469e-06,
|
| 433 |
+
"loss": 0.5822,
|
| 434 |
+
"mean_token_accuracy": 0.8119319261486757,
|
| 435 |
+
"step": 255
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.770941438102298,
|
| 439 |
+
"grad_norm": 0.4624658793689265,
|
| 440 |
+
"learning_rate": 3.021167106673928e-06,
|
| 441 |
+
"loss": 0.5895,
|
| 442 |
+
"mean_token_accuracy": 0.8094238400488403,
|
| 443 |
+
"step": 260
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.7857672349888807,
|
| 447 |
+
"grad_norm": 0.4887471221196979,
|
| 448 |
+
"learning_rate": 2.6594147124053983e-06,
|
| 449 |
+
"loss": 0.5933,
|
| 450 |
+
"mean_token_accuracy": 0.8082280219780221,
|
| 451 |
+
"step": 265
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.8005930318754633,
|
| 455 |
+
"grad_norm": 0.4755865942641489,
|
| 456 |
+
"learning_rate": 2.317385996808195e-06,
|
| 457 |
+
"loss": 0.5796,
|
| 458 |
+
"mean_token_accuracy": 0.8127861721611721,
|
| 459 |
+
"step": 270
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.815418828762046,
|
| 463 |
+
"grad_norm": 0.47035132329171114,
|
| 464 |
+
"learning_rate": 1.9959999689556407e-06,
|
| 465 |
+
"loss": 0.5967,
|
| 466 |
+
"mean_token_accuracy": 0.8072924297924298,
|
| 467 |
+
"step": 275
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 0.8302446256486287,
|
| 471 |
+
"grad_norm": 0.472860512590234,
|
| 472 |
+
"learning_rate": 1.6961201723520248e-06,
|
| 473 |
+
"loss": 0.578,
|
| 474 |
+
"mean_token_accuracy": 0.8127960927960929,
|
| 475 |
+
"step": 280
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 0.8450704225352113,
|
| 479 |
+
"grad_norm": 0.4614371078147505,
|
| 480 |
+
"learning_rate": 1.4185523646469822e-06,
|
| 481 |
+
"loss": 0.5901,
|
| 482 |
+
"mean_token_accuracy": 0.8086108107426412,
|
| 483 |
+
"step": 285
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.859896219421794,
|
| 487 |
+
"grad_norm": 0.43931896763540745,
|
| 488 |
+
"learning_rate": 1.1640423526166987e-06,
|
| 489 |
+
"loss": 0.5834,
|
| 490 |
+
"mean_token_accuracy": 0.8111981074481076,
|
| 491 |
+
"step": 290
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.8747220163083765,
|
| 495 |
+
"grad_norm": 0.4456895234302146,
|
| 496 |
+
"learning_rate": 9.332739882292752e-07,
|
| 497 |
+
"loss": 0.5842,
|
| 498 |
+
"mean_token_accuracy": 0.8116254578754578,
|
| 499 |
+
"step": 295
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.8895478131949592,
|
| 503 |
+
"grad_norm": 0.43426061796264454,
|
| 504 |
+
"learning_rate": 7.268673311786378e-07,
|
| 505 |
+
"loss": 0.572,
|
| 506 |
+
"mean_token_accuracy": 0.8145253357753358,
|
| 507 |
+
"step": 300
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.8895478131949592,
|
| 511 |
+
"eval_loss": 0.6066814064979553,
|
| 512 |
+
"eval_mean_token_accuracy": 0.8039550018326881,
|
| 513 |
+
"eval_runtime": 5.5504,
|
| 514 |
+
"eval_samples_per_second": 23.061,
|
| 515 |
+
"eval_steps_per_second": 1.441,
|
| 516 |
+
"step": 300
|
| 517 |
+
},
|
| 518 |
+
{
|
| 519 |
+
"epoch": 0.9043736100815419,
|
| 520 |
+
"grad_norm": 0.45927392759405916,
|
| 521 |
+
"learning_rate": 5.453769828241872e-07,
|
| 522 |
+
"loss": 0.5848,
|
| 523 |
+
"mean_token_accuracy": 0.8108882783882784,
|
| 524 |
+
"step": 305
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.9191994069681245,
|
| 528 |
+
"grad_norm": 0.44015278909822164,
|
| 529 |
+
"learning_rate": 3.8929059601275463e-07,
|
| 530 |
+
"loss": 0.5883,
|
| 531 |
+
"mean_token_accuracy": 0.8098282967032967,
|
| 532 |
+
"step": 310
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 0.9340252038547072,
|
| 536 |
+
"grad_norm": 0.46373639838618946,
|
| 537 |
+
"learning_rate": 2.5902756478688674e-07,
|
| 538 |
+
"loss": 0.5893,
|
| 539 |
+
"mean_token_accuracy": 0.8085653235653236,
|
| 540 |
+
"step": 315
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 0.9488510007412898,
|
| 544 |
+
"grad_norm": 0.4453169356064979,
|
| 545 |
+
"learning_rate": 1.5493789750014032e-07,
|
| 546 |
+
"loss": 0.5934,
|
| 547 |
+
"mean_token_accuracy": 0.8079464285714286,
|
| 548 |
+
"step": 320
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.9636767976278725,
|
| 552 |
+
"grad_norm": 0.44080298077757607,
|
| 553 |
+
"learning_rate": 7.730127636723539e-08,
|
| 554 |
+
"loss": 0.5835,
|
| 555 |
+
"mean_token_accuracy": 0.8108723706229716,
|
| 556 |
+
"step": 325
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.9785025945144552,
|
| 560 |
+
"grad_norm": 0.42873433987601206,
|
| 561 |
+
"learning_rate": 2.6326305976001054e-08,
|
| 562 |
+
"loss": 0.5817,
|
| 563 |
+
"mean_token_accuracy": 0.8121665140415143,
|
| 564 |
+
"step": 330
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.9933283914010378,
|
| 568 |
+
"grad_norm": 0.44474948977305667,
|
| 569 |
+
"learning_rate": 2.149952780321485e-09,
|
| 570 |
+
"loss": 0.5933,
|
| 571 |
+
"mean_token_accuracy": 0.8080050057957257,
|
| 572 |
+
"step": 335
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.9992587101556709,
|
| 576 |
+
"mean_token_accuracy": 0.8132173382173382,
|
| 577 |
+
"step": 337,
|
| 578 |
+
"total_flos": 176608216350720.0,
|
| 579 |
+
"train_loss": 0.6238384791580789,
|
| 580 |
+
"train_runtime": 3258.6411,
|
| 581 |
+
"train_samples_per_second": 6.624,
|
| 582 |
+
"train_steps_per_second": 0.103
|
| 583 |
+
}
|
| 584 |
+
],
|
| 585 |
+
"logging_steps": 5,
|
| 586 |
+
"max_steps": 337,
|
| 587 |
+
"num_input_tokens_seen": 0,
|
| 588 |
+
"num_train_epochs": 1,
|
| 589 |
+
"save_steps": 500,
|
| 590 |
+
"stateful_callbacks": {
|
| 591 |
+
"TrainerControl": {
|
| 592 |
+
"args": {
|
| 593 |
+
"should_epoch_stop": false,
|
| 594 |
+
"should_evaluate": false,
|
| 595 |
+
"should_log": false,
|
| 596 |
+
"should_save": false,
|
| 597 |
+
"should_training_stop": false
|
| 598 |
+
},
|
| 599 |
+
"attributes": {}
|
| 600 |
+
}
|
| 601 |
+
},
|
| 602 |
+
"total_flos": 176608216350720.0,
|
| 603 |
+
"train_batch_size": 2,
|
| 604 |
+
"trial_name": null,
|
| 605 |
+
"trial_params": null
|
| 606 |
+
}
|