Commit
•
e9f741c
1
Parent(s):
7cf5aec
End of training
Browse files- all_results.json +15 -0
- eval_results.json +10 -0
- train_results.json +8 -0
- trainer_state.json +451 -0
all_results.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 100.16,
|
| 3 |
+
"eval_cer": 0.10498671252149445,
|
| 4 |
+
"eval_loss": 0.4146759808063507,
|
| 5 |
+
"eval_runtime": 116.6347,
|
| 6 |
+
"eval_samples": 2796,
|
| 7 |
+
"eval_samples_per_second": 23.972,
|
| 8 |
+
"eval_steps_per_second": 0.377,
|
| 9 |
+
"eval_wer": 0.3171619442524764,
|
| 10 |
+
"train_loss": 1.6543282796931436,
|
| 11 |
+
"train_runtime": 29157.3737,
|
| 12 |
+
"train_samples": 6538,
|
| 13 |
+
"train_samples_per_second": 22.424,
|
| 14 |
+
"train_steps_per_second": 0.175
|
| 15 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 100.16,
|
| 3 |
+
"eval_cer": 0.10498671252149445,
|
| 4 |
+
"eval_loss": 0.4146759808063507,
|
| 5 |
+
"eval_runtime": 116.6347,
|
| 6 |
+
"eval_samples": 2796,
|
| 7 |
+
"eval_samples_per_second": 23.972,
|
| 8 |
+
"eval_steps_per_second": 0.377,
|
| 9 |
+
"eval_wer": 0.3171619442524764
|
| 10 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 100.16,
|
| 3 |
+
"train_loss": 1.6543282796931436,
|
| 4 |
+
"train_runtime": 29157.3737,
|
| 5 |
+
"train_samples": 6538,
|
| 6 |
+
"train_samples_per_second": 22.424,
|
| 7 |
+
"train_steps_per_second": 0.175
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,451 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 100.15533980582525,
|
| 5 |
+
"global_step": 5108,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 1.95,
|
| 12 |
+
"learning_rate": 1.487205882352941e-05,
|
| 13 |
+
"loss": 16.4084,
|
| 14 |
+
"step": 100
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 3.91,
|
| 18 |
+
"learning_rate": 2.9430882352941174e-05,
|
| 19 |
+
"loss": 6.2142,
|
| 20 |
+
"step": 200
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 5.87,
|
| 24 |
+
"learning_rate": 4.398970588235294e-05,
|
| 25 |
+
"loss": 4.1261,
|
| 26 |
+
"step": 300
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 7.83,
|
| 30 |
+
"learning_rate": 5.8548529411764694e-05,
|
| 31 |
+
"loss": 3.2894,
|
| 32 |
+
"step": 400
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 7.83,
|
| 36 |
+
"eval_cer": 1.0,
|
| 37 |
+
"eval_loss": 3.1501412391662598,
|
| 38 |
+
"eval_runtime": 111.2121,
|
| 39 |
+
"eval_samples_per_second": 25.141,
|
| 40 |
+
"eval_steps_per_second": 0.396,
|
| 41 |
+
"eval_wer": 1.0,
|
| 42 |
+
"step": 400
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 9.8,
|
| 46 |
+
"learning_rate": 7.310735294117646e-05,
|
| 47 |
+
"loss": 3.083,
|
| 48 |
+
"step": 500
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 11.76,
|
| 52 |
+
"learning_rate": 7.5e-05,
|
| 53 |
+
"loss": 2.9638,
|
| 54 |
+
"step": 600
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 13.72,
|
| 58 |
+
"learning_rate": 7.5e-05,
|
| 59 |
+
"loss": 2.534,
|
| 60 |
+
"step": 700
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 15.68,
|
| 64 |
+
"learning_rate": 7.5e-05,
|
| 65 |
+
"loss": 1.8586,
|
| 66 |
+
"step": 800
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 15.68,
|
| 70 |
+
"eval_cer": 0.24023761138033453,
|
| 71 |
+
"eval_loss": 0.8871385455131531,
|
| 72 |
+
"eval_runtime": 114.0709,
|
| 73 |
+
"eval_samples_per_second": 24.511,
|
| 74 |
+
"eval_steps_per_second": 0.386,
|
| 75 |
+
"eval_wer": 0.6721492743607463,
|
| 76 |
+
"step": 800
|
| 77 |
+
},
|
| 78 |
+
{
|
| 79 |
+
"epoch": 17.64,
|
| 80 |
+
"learning_rate": 7.5e-05,
|
| 81 |
+
"loss": 1.6008,
|
| 82 |
+
"step": 900
|
| 83 |
+
},
|
| 84 |
+
{
|
| 85 |
+
"epoch": 19.6,
|
| 86 |
+
"learning_rate": 7.5e-05,
|
| 87 |
+
"loss": 1.4898,
|
| 88 |
+
"step": 1000
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 21.56,
|
| 92 |
+
"learning_rate": 7.5e-05,
|
| 93 |
+
"loss": 1.4105,
|
| 94 |
+
"step": 1100
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 23.52,
|
| 98 |
+
"learning_rate": 7.5e-05,
|
| 99 |
+
"loss": 1.3431,
|
| 100 |
+
"step": 1200
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 23.52,
|
| 104 |
+
"eval_cer": 0.19387212755979366,
|
| 105 |
+
"eval_loss": 0.5813264846801758,
|
| 106 |
+
"eval_runtime": 110.7438,
|
| 107 |
+
"eval_samples_per_second": 25.247,
|
| 108 |
+
"eval_steps_per_second": 0.397,
|
| 109 |
+
"eval_wer": 0.5502418797512094,
|
| 110 |
+
"step": 1200
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 25.49,
|
| 114 |
+
"learning_rate": 7.5e-05,
|
| 115 |
+
"loss": 1.2993,
|
| 116 |
+
"step": 1300
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 27.45,
|
| 120 |
+
"learning_rate": 7.5e-05,
|
| 121 |
+
"loss": 1.2563,
|
| 122 |
+
"step": 1400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 29.41,
|
| 126 |
+
"learning_rate": 7.5e-05,
|
| 127 |
+
"loss": 1.2298,
|
| 128 |
+
"step": 1500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 31.37,
|
| 132 |
+
"learning_rate": 7.5e-05,
|
| 133 |
+
"loss": 1.2052,
|
| 134 |
+
"step": 1600
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 31.37,
|
| 138 |
+
"eval_cer": 0.1664738679589391,
|
| 139 |
+
"eval_loss": 0.49559420347213745,
|
| 140 |
+
"eval_runtime": 111.374,
|
| 141 |
+
"eval_samples_per_second": 25.105,
|
| 142 |
+
"eval_steps_per_second": 0.395,
|
| 143 |
+
"eval_wer": 0.47878369039391844,
|
| 144 |
+
"step": 1600
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 33.33,
|
| 148 |
+
"learning_rate": 7.5e-05,
|
| 149 |
+
"loss": 1.186,
|
| 150 |
+
"step": 1700
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 35.29,
|
| 154 |
+
"learning_rate": 7.5e-05,
|
| 155 |
+
"loss": 1.1473,
|
| 156 |
+
"step": 1800
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 37.25,
|
| 160 |
+
"learning_rate": 7.5e-05,
|
| 161 |
+
"loss": 1.1235,
|
| 162 |
+
"step": 1900
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 39.21,
|
| 166 |
+
"learning_rate": 7.5e-05,
|
| 167 |
+
"loss": 1.1097,
|
| 168 |
+
"step": 2000
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 39.21,
|
| 172 |
+
"eval_cer": 0.1397321661195352,
|
| 173 |
+
"eval_loss": 0.44468843936920166,
|
| 174 |
+
"eval_runtime": 109.918,
|
| 175 |
+
"eval_samples_per_second": 25.437,
|
| 176 |
+
"eval_steps_per_second": 0.4,
|
| 177 |
+
"eval_wer": 0.41428242340474547,
|
| 178 |
+
"step": 2000
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 41.17,
|
| 182 |
+
"learning_rate": 7.5e-05,
|
| 183 |
+
"loss": 1.0951,
|
| 184 |
+
"step": 2100
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 43.14,
|
| 188 |
+
"learning_rate": 7.5e-05,
|
| 189 |
+
"loss": 1.0785,
|
| 190 |
+
"step": 2200
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 45.1,
|
| 194 |
+
"learning_rate": 7.5e-05,
|
| 195 |
+
"loss": 1.0575,
|
| 196 |
+
"step": 2300
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 47.06,
|
| 200 |
+
"learning_rate": 7.5e-05,
|
| 201 |
+
"loss": 1.0528,
|
| 202 |
+
"step": 2400
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 47.06,
|
| 206 |
+
"eval_cer": 0.13331249022979522,
|
| 207 |
+
"eval_loss": 0.4439217448234558,
|
| 208 |
+
"eval_runtime": 111.2432,
|
| 209 |
+
"eval_samples_per_second": 25.134,
|
| 210 |
+
"eval_steps_per_second": 0.396,
|
| 211 |
+
"eval_wer": 0.3960838516470859,
|
| 212 |
+
"step": 2400
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 49.02,
|
| 216 |
+
"learning_rate": 7.5e-05,
|
| 217 |
+
"loss": 1.0359,
|
| 218 |
+
"step": 2500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 50.97,
|
| 222 |
+
"learning_rate": 7.377299412915851e-05,
|
| 223 |
+
"loss": 1.0118,
|
| 224 |
+
"step": 2600
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 52.93,
|
| 228 |
+
"learning_rate": 7.098434442270058e-05,
|
| 229 |
+
"loss": 1.0083,
|
| 230 |
+
"step": 2700
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 54.89,
|
| 234 |
+
"learning_rate": 6.819569471624266e-05,
|
| 235 |
+
"loss": 0.9939,
|
| 236 |
+
"step": 2800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 54.89,
|
| 240 |
+
"eval_cer": 0.13785628680110468,
|
| 241 |
+
"eval_loss": 0.43479105830192566,
|
| 242 |
+
"eval_runtime": 112.3814,
|
| 243 |
+
"eval_samples_per_second": 24.88,
|
| 244 |
+
"eval_steps_per_second": 0.392,
|
| 245 |
+
"eval_wer": 0.40138217000691084,
|
| 246 |
+
"step": 2800
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 56.85,
|
| 250 |
+
"learning_rate": 6.540704500978474e-05,
|
| 251 |
+
"loss": 0.9884,
|
| 252 |
+
"step": 2900
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 58.82,
|
| 256 |
+
"learning_rate": 6.26183953033268e-05,
|
| 257 |
+
"loss": 0.9771,
|
| 258 |
+
"step": 3000
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 60.78,
|
| 262 |
+
"learning_rate": 5.982974559686888e-05,
|
| 263 |
+
"loss": 0.9577,
|
| 264 |
+
"step": 3100
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 62.74,
|
| 268 |
+
"learning_rate": 5.704109589041095e-05,
|
| 269 |
+
"loss": 0.9441,
|
| 270 |
+
"step": 3200
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 62.74,
|
| 274 |
+
"eval_cer": 0.12230733156166954,
|
| 275 |
+
"eval_loss": 0.42357733845710754,
|
| 276 |
+
"eval_runtime": 114.6173,
|
| 277 |
+
"eval_samples_per_second": 24.394,
|
| 278 |
+
"eval_steps_per_second": 0.384,
|
| 279 |
+
"eval_wer": 0.365261460492974,
|
| 280 |
+
"step": 3200
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 64.7,
|
| 284 |
+
"learning_rate": 5.425244618395303e-05,
|
| 285 |
+
"loss": 0.9296,
|
| 286 |
+
"step": 3300
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 66.66,
|
| 290 |
+
"learning_rate": 5.14637964774951e-05,
|
| 291 |
+
"loss": 0.9203,
|
| 292 |
+
"step": 3400
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 68.62,
|
| 296 |
+
"learning_rate": 4.867514677103718e-05,
|
| 297 |
+
"loss": 0.9215,
|
| 298 |
+
"step": 3500
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 70.58,
|
| 302 |
+
"learning_rate": 4.5886497064579256e-05,
|
| 303 |
+
"loss": 0.913,
|
| 304 |
+
"step": 3600
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 70.58,
|
| 308 |
+
"eval_cer": 0.11566880308477932,
|
| 309 |
+
"eval_loss": 0.43085047602653503,
|
| 310 |
+
"eval_runtime": 112.648,
|
| 311 |
+
"eval_samples_per_second": 24.821,
|
| 312 |
+
"eval_steps_per_second": 0.391,
|
| 313 |
+
"eval_wer": 0.3475236120709514,
|
| 314 |
+
"step": 3600
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 72.54,
|
| 318 |
+
"learning_rate": 4.309784735812133e-05,
|
| 319 |
+
"loss": 0.9033,
|
| 320 |
+
"step": 3700
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 74.5,
|
| 324 |
+
"learning_rate": 4.03091976516634e-05,
|
| 325 |
+
"loss": 0.8945,
|
| 326 |
+
"step": 3800
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 76.47,
|
| 330 |
+
"learning_rate": 3.752054794520548e-05,
|
| 331 |
+
"loss": 0.8714,
|
| 332 |
+
"step": 3900
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 78.43,
|
| 336 |
+
"learning_rate": 3.473189823874755e-05,
|
| 337 |
+
"loss": 0.8678,
|
| 338 |
+
"step": 4000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 78.43,
|
| 342 |
+
"eval_cer": 0.1109895263404721,
|
| 343 |
+
"eval_loss": 0.42695942521095276,
|
| 344 |
+
"eval_runtime": 109.8691,
|
| 345 |
+
"eval_samples_per_second": 25.448,
|
| 346 |
+
"eval_steps_per_second": 0.4,
|
| 347 |
+
"eval_wer": 0.3337479843354066,
|
| 348 |
+
"step": 4000
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 80.39,
|
| 352 |
+
"learning_rate": 3.1943248532289626e-05,
|
| 353 |
+
"loss": 0.8612,
|
| 354 |
+
"step": 4100
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 82.35,
|
| 358 |
+
"learning_rate": 2.9154598825831697e-05,
|
| 359 |
+
"loss": 0.8603,
|
| 360 |
+
"step": 4200
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 84.31,
|
| 364 |
+
"learning_rate": 2.6365949119373778e-05,
|
| 365 |
+
"loss": 0.8451,
|
| 366 |
+
"step": 4300
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 86.27,
|
| 370 |
+
"learning_rate": 2.357729941291585e-05,
|
| 371 |
+
"loss": 0.8414,
|
| 372 |
+
"step": 4400
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 86.27,
|
| 376 |
+
"eval_cer": 0.10695638580584649,
|
| 377 |
+
"eval_loss": 0.41582536697387695,
|
| 378 |
+
"eval_runtime": 110.6892,
|
| 379 |
+
"eval_samples_per_second": 25.26,
|
| 380 |
+
"eval_steps_per_second": 0.398,
|
| 381 |
+
"eval_wer": 0.32204561161022804,
|
| 382 |
+
"step": 4400
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 88.23,
|
| 386 |
+
"learning_rate": 2.0788649706457926e-05,
|
| 387 |
+
"loss": 0.8394,
|
| 388 |
+
"step": 4500
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 90.19,
|
| 392 |
+
"learning_rate": 1.7999999999999997e-05,
|
| 393 |
+
"loss": 0.8304,
|
| 394 |
+
"step": 4600
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 92.16,
|
| 398 |
+
"learning_rate": 1.5211350293542076e-05,
|
| 399 |
+
"loss": 0.8188,
|
| 400 |
+
"step": 4700
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 94.12,
|
| 404 |
+
"learning_rate": 1.2422700587084147e-05,
|
| 405 |
+
"loss": 0.817,
|
| 406 |
+
"step": 4800
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 94.12,
|
| 410 |
+
"eval_cer": 0.10721692460007295,
|
| 411 |
+
"eval_loss": 0.4184626638889313,
|
| 412 |
+
"eval_runtime": 112.2012,
|
| 413 |
+
"eval_samples_per_second": 24.92,
|
| 414 |
+
"eval_steps_per_second": 0.392,
|
| 415 |
+
"eval_wer": 0.323105275282193,
|
| 416 |
+
"step": 4800
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 96.08,
|
| 420 |
+
"learning_rate": 9.634050880626226e-06,
|
| 421 |
+
"loss": 0.8126,
|
| 422 |
+
"step": 4900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 98.04,
|
| 426 |
+
"learning_rate": 6.845401174168297e-06,
|
| 427 |
+
"loss": 0.8105,
|
| 428 |
+
"step": 5000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 99.99,
|
| 432 |
+
"learning_rate": 4.056751467710376e-06,
|
| 433 |
+
"loss": 0.7959,
|
| 434 |
+
"step": 5100
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 100.16,
|
| 438 |
+
"step": 5108,
|
| 439 |
+
"total_flos": 8.553278225483388e+19,
|
| 440 |
+
"train_loss": 1.6543282796931436,
|
| 441 |
+
"train_runtime": 29157.3737,
|
| 442 |
+
"train_samples_per_second": 22.424,
|
| 443 |
+
"train_steps_per_second": 0.175
|
| 444 |
+
}
|
| 445 |
+
],
|
| 446 |
+
"max_steps": 5108,
|
| 447 |
+
"num_train_epochs": 101,
|
| 448 |
+
"total_flos": 8.553278225483388e+19,
|
| 449 |
+
"trial_name": null,
|
| 450 |
+
"trial_params": null
|
| 451 |
+
}
|