Commit
•
d7b30b2
1
Parent(s):
09a269e
End of training
Browse files- all_results.json +5 -5
- train_results.json +5 -5
- trainer_state.json +222 -102
all_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss":
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 128272,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 0.3749944605037362,
|
| 4 |
+
"train_runtime": 1805.8732,
|
| 5 |
"train_samples": 128272,
|
| 6 |
+
"train_samples_per_second": 355.152,
|
| 7 |
+
"train_steps_per_second": 2.774
|
| 8 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"train_loss":
|
| 4 |
-
"train_runtime":
|
| 5 |
"train_samples": 128272,
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second":
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 0.3749944605037362,
|
| 4 |
+
"train_runtime": 1805.8732,
|
| 5 |
"train_samples": 128272,
|
| 6 |
+
"train_samples_per_second": 355.152,
|
| 7 |
+
"train_steps_per_second": 2.774
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,208 +1,328 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 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.
|
| 13 |
-
"learning_rate": 4.
|
| 14 |
-
"loss": 1.
|
| 15 |
"step": 100
|
| 16 |
},
|
| 17 |
{
|
| 18 |
-
"epoch": 0.
|
| 19 |
-
"learning_rate": 4.
|
| 20 |
-
"loss": 1.
|
| 21 |
"step": 200
|
| 22 |
},
|
| 23 |
{
|
| 24 |
-
"epoch": 0.
|
| 25 |
-
"learning_rate": 4.
|
| 26 |
-
"loss": 1.
|
| 27 |
"step": 300
|
| 28 |
},
|
| 29 |
{
|
| 30 |
-
"epoch": 0.
|
| 31 |
-
"learning_rate": 4.
|
| 32 |
-
"loss": 1.
|
| 33 |
"step": 400
|
| 34 |
},
|
| 35 |
{
|
| 36 |
-
"epoch": 0.
|
| 37 |
-
"learning_rate": 4.
|
| 38 |
-
"loss": 1.
|
| 39 |
"step": 500
|
| 40 |
},
|
| 41 |
{
|
| 42 |
-
"epoch": 0.
|
| 43 |
-
"learning_rate": 4.
|
| 44 |
-
"loss": 1.
|
| 45 |
"step": 600
|
| 46 |
},
|
| 47 |
{
|
| 48 |
-
"epoch": 0.
|
| 49 |
-
"learning_rate":
|
| 50 |
-
"loss": 1.
|
| 51 |
"step": 700
|
| 52 |
},
|
| 53 |
{
|
| 54 |
-
"epoch": 0.
|
| 55 |
-
"learning_rate":
|
| 56 |
-
"loss": 1.
|
| 57 |
"step": 800
|
| 58 |
},
|
| 59 |
{
|
| 60 |
-
"epoch": 0.
|
| 61 |
-
"learning_rate":
|
| 62 |
-
"loss": 1.
|
| 63 |
"step": 900
|
| 64 |
},
|
| 65 |
{
|
| 66 |
-
"epoch":
|
| 67 |
-
"learning_rate":
|
| 68 |
-
"loss": 1.
|
| 69 |
"step": 1000
|
| 70 |
},
|
| 71 |
{
|
| 72 |
-
"epoch":
|
| 73 |
-
"learning_rate":
|
| 74 |
-
"loss": 1.
|
| 75 |
"step": 1100
|
| 76 |
},
|
| 77 |
{
|
| 78 |
-
"epoch":
|
| 79 |
-
"learning_rate":
|
| 80 |
-
"loss": 1.
|
| 81 |
"step": 1200
|
| 82 |
},
|
| 83 |
{
|
| 84 |
-
"epoch":
|
| 85 |
-
"learning_rate":
|
| 86 |
-
"loss": 1.
|
| 87 |
"step": 1300
|
| 88 |
},
|
| 89 |
{
|
| 90 |
-
"epoch":
|
| 91 |
-
"learning_rate":
|
| 92 |
-
"loss": 1.
|
| 93 |
"step": 1400
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch":
|
| 97 |
-
"learning_rate":
|
| 98 |
-
"loss": 1.
|
| 99 |
"step": 1500
|
| 100 |
},
|
| 101 |
{
|
| 102 |
-
"epoch":
|
| 103 |
-
"learning_rate":
|
| 104 |
-
"loss": 1.
|
| 105 |
"step": 1600
|
| 106 |
},
|
| 107 |
{
|
| 108 |
-
"epoch":
|
| 109 |
-
"learning_rate":
|
| 110 |
-
"loss": 1.
|
| 111 |
"step": 1700
|
| 112 |
},
|
| 113 |
{
|
| 114 |
-
"epoch":
|
| 115 |
-
"learning_rate":
|
| 116 |
-
"loss": 1.
|
| 117 |
"step": 1800
|
| 118 |
},
|
| 119 |
{
|
| 120 |
-
"epoch":
|
| 121 |
-
"learning_rate":
|
| 122 |
-
"loss": 1.
|
| 123 |
"step": 1900
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch":
|
| 127 |
-
"learning_rate":
|
| 128 |
-
"loss": 1.
|
| 129 |
"step": 2000
|
| 130 |
},
|
| 131 |
{
|
| 132 |
-
"epoch":
|
| 133 |
-
"learning_rate":
|
| 134 |
-
"loss": 1.
|
| 135 |
"step": 2100
|
| 136 |
},
|
| 137 |
{
|
| 138 |
-
"epoch":
|
| 139 |
-
"learning_rate":
|
| 140 |
-
"loss": 1.
|
| 141 |
"step": 2200
|
| 142 |
},
|
| 143 |
{
|
| 144 |
-
"epoch":
|
| 145 |
-
"learning_rate":
|
| 146 |
-
"loss": 1.
|
| 147 |
"step": 2300
|
| 148 |
},
|
| 149 |
{
|
| 150 |
-
"epoch":
|
| 151 |
-
"learning_rate":
|
| 152 |
-
"loss": 1.
|
| 153 |
"step": 2400
|
| 154 |
},
|
| 155 |
{
|
| 156 |
-
"epoch":
|
| 157 |
-
"learning_rate":
|
| 158 |
-
"loss": 1.
|
| 159 |
"step": 2500
|
| 160 |
},
|
| 161 |
{
|
| 162 |
-
"epoch":
|
| 163 |
-
"learning_rate":
|
| 164 |
-
"loss": 1.
|
| 165 |
"step": 2600
|
| 166 |
},
|
| 167 |
{
|
| 168 |
-
"epoch":
|
| 169 |
-
"learning_rate":
|
| 170 |
-
"loss": 1.
|
| 171 |
"step": 2700
|
| 172 |
},
|
| 173 |
{
|
| 174 |
-
"epoch":
|
| 175 |
-
"learning_rate":
|
| 176 |
-
"loss": 1.
|
| 177 |
"step": 2800
|
| 178 |
},
|
| 179 |
{
|
| 180 |
-
"epoch":
|
| 181 |
-
"learning_rate":
|
| 182 |
-
"loss": 1.
|
| 183 |
"step": 2900
|
| 184 |
},
|
| 185 |
{
|
| 186 |
-
"epoch":
|
| 187 |
-
"learning_rate":
|
| 188 |
-
"loss": 1.
|
| 189 |
"step": 3000
|
| 190 |
},
|
| 191 |
{
|
| 192 |
-
"epoch":
|
| 193 |
-
"
|
| 194 |
-
"
|
| 195 |
-
"
|
| 196 |
-
|
| 197 |
-
|
| 198 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 199 |
}
|
| 200 |
],
|
| 201 |
"logging_steps": 100,
|
| 202 |
-
"max_steps":
|
| 203 |
-
"num_train_epochs":
|
| 204 |
"save_steps": 500,
|
| 205 |
-
"total_flos":
|
| 206 |
"trial_name": null,
|
| 207 |
"trial_params": null
|
| 208 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 4.999750530123488,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 5010,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.02,
|
| 13 |
+
"learning_rate": 4.9731096052490054e-05,
|
| 14 |
+
"loss": 1.7709,
|
| 15 |
"step": 100
|
| 16 |
},
|
| 17 |
{
|
| 18 |
+
"epoch": 0.03,
|
| 19 |
+
"learning_rate": 4.9462192104980105e-05,
|
| 20 |
+
"loss": 1.6756,
|
| 21 |
"step": 200
|
| 22 |
},
|
| 23 |
{
|
| 24 |
+
"epoch": 0.05,
|
| 25 |
+
"learning_rate": 4.919328815747015e-05,
|
| 26 |
+
"loss": 1.6585,
|
| 27 |
"step": 300
|
| 28 |
},
|
| 29 |
{
|
| 30 |
+
"epoch": 0.06,
|
| 31 |
+
"learning_rate": 4.892438420996021e-05,
|
| 32 |
+
"loss": 1.6405,
|
| 33 |
"step": 400
|
| 34 |
},
|
| 35 |
{
|
| 36 |
+
"epoch": 0.08,
|
| 37 |
+
"learning_rate": 4.865548026245025e-05,
|
| 38 |
+
"loss": 1.5871,
|
| 39 |
"step": 500
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 0.1,
|
| 43 |
+
"learning_rate": 4.838657631494031e-05,
|
| 44 |
+
"loss": 1.5809,
|
| 45 |
"step": 600
|
| 46 |
},
|
| 47 |
{
|
| 48 |
+
"epoch": 0.11,
|
| 49 |
+
"learning_rate": 4.8117672367430355e-05,
|
| 50 |
+
"loss": 1.5877,
|
| 51 |
"step": 700
|
| 52 |
},
|
| 53 |
{
|
| 54 |
+
"epoch": 0.13,
|
| 55 |
+
"learning_rate": 4.7848768419920406e-05,
|
| 56 |
+
"loss": 1.5475,
|
| 57 |
"step": 800
|
| 58 |
},
|
| 59 |
{
|
| 60 |
+
"epoch": 0.15,
|
| 61 |
+
"learning_rate": 4.757986447241046e-05,
|
| 62 |
+
"loss": 1.5461,
|
| 63 |
"step": 900
|
| 64 |
},
|
| 65 |
{
|
| 66 |
+
"epoch": 0.16,
|
| 67 |
+
"learning_rate": 4.731096052490051e-05,
|
| 68 |
+
"loss": 1.5553,
|
| 69 |
"step": 1000
|
| 70 |
},
|
| 71 |
{
|
| 72 |
+
"epoch": 0.18,
|
| 73 |
+
"learning_rate": 4.704205657739056e-05,
|
| 74 |
+
"loss": 1.543,
|
| 75 |
"step": 1100
|
| 76 |
},
|
| 77 |
{
|
| 78 |
+
"epoch": 0.19,
|
| 79 |
+
"learning_rate": 4.6773152629880605e-05,
|
| 80 |
+
"loss": 1.5603,
|
| 81 |
"step": 1200
|
| 82 |
},
|
| 83 |
{
|
| 84 |
+
"epoch": 0.21,
|
| 85 |
+
"learning_rate": 4.650424868237066e-05,
|
| 86 |
+
"loss": 1.5441,
|
| 87 |
"step": 1300
|
| 88 |
},
|
| 89 |
{
|
| 90 |
+
"epoch": 0.23,
|
| 91 |
+
"learning_rate": 4.623534473486071e-05,
|
| 92 |
+
"loss": 1.5416,
|
| 93 |
"step": 1400
|
| 94 |
},
|
| 95 |
{
|
| 96 |
+
"epoch": 0.24,
|
| 97 |
+
"learning_rate": 4.596644078735076e-05,
|
| 98 |
+
"loss": 1.5159,
|
| 99 |
"step": 1500
|
| 100 |
},
|
| 101 |
{
|
| 102 |
+
"epoch": 0.26,
|
| 103 |
+
"learning_rate": 4.569753683984081e-05,
|
| 104 |
+
"loss": 1.5225,
|
| 105 |
"step": 1600
|
| 106 |
},
|
| 107 |
{
|
| 108 |
+
"epoch": 0.27,
|
| 109 |
+
"learning_rate": 4.542863289233086e-05,
|
| 110 |
+
"loss": 1.5281,
|
| 111 |
"step": 1700
|
| 112 |
},
|
| 113 |
{
|
| 114 |
+
"epoch": 0.29,
|
| 115 |
+
"learning_rate": 4.515972894482091e-05,
|
| 116 |
+
"loss": 1.5128,
|
| 117 |
"step": 1800
|
| 118 |
},
|
| 119 |
{
|
| 120 |
+
"epoch": 0.31,
|
| 121 |
+
"learning_rate": 4.4890824997310964e-05,
|
| 122 |
+
"loss": 1.5215,
|
| 123 |
"step": 1900
|
| 124 |
},
|
| 125 |
{
|
| 126 |
+
"epoch": 0.32,
|
| 127 |
+
"learning_rate": 4.4621921049801015e-05,
|
| 128 |
+
"loss": 1.4975,
|
| 129 |
"step": 2000
|
| 130 |
},
|
| 131 |
{
|
| 132 |
+
"epoch": 0.34,
|
| 133 |
+
"learning_rate": 4.435301710229106e-05,
|
| 134 |
+
"loss": 1.4772,
|
| 135 |
"step": 2100
|
| 136 |
},
|
| 137 |
{
|
| 138 |
+
"epoch": 0.35,
|
| 139 |
+
"learning_rate": 4.408411315478112e-05,
|
| 140 |
+
"loss": 1.4776,
|
| 141 |
"step": 2200
|
| 142 |
},
|
| 143 |
{
|
| 144 |
+
"epoch": 0.37,
|
| 145 |
+
"learning_rate": 4.381520920727116e-05,
|
| 146 |
+
"loss": 1.5157,
|
| 147 |
"step": 2300
|
| 148 |
},
|
| 149 |
{
|
| 150 |
+
"epoch": 0.39,
|
| 151 |
+
"learning_rate": 4.3546305259761214e-05,
|
| 152 |
+
"loss": 1.4968,
|
| 153 |
"step": 2400
|
| 154 |
},
|
| 155 |
{
|
| 156 |
+
"epoch": 0.4,
|
| 157 |
+
"learning_rate": 4.3277401312251265e-05,
|
| 158 |
+
"loss": 1.4835,
|
| 159 |
"step": 2500
|
| 160 |
},
|
| 161 |
{
|
| 162 |
+
"epoch": 0.42,
|
| 163 |
+
"learning_rate": 4.3008497364741317e-05,
|
| 164 |
+
"loss": 1.4805,
|
| 165 |
"step": 2600
|
| 166 |
},
|
| 167 |
{
|
| 168 |
+
"epoch": 0.44,
|
| 169 |
+
"learning_rate": 4.273959341723137e-05,
|
| 170 |
+
"loss": 1.4892,
|
| 171 |
"step": 2700
|
| 172 |
},
|
| 173 |
{
|
| 174 |
+
"epoch": 0.45,
|
| 175 |
+
"learning_rate": 4.247068946972142e-05,
|
| 176 |
+
"loss": 1.4705,
|
| 177 |
"step": 2800
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 0.47,
|
| 181 |
+
"learning_rate": 4.220178552221147e-05,
|
| 182 |
+
"loss": 1.4829,
|
| 183 |
"step": 2900
|
| 184 |
},
|
| 185 |
{
|
| 186 |
+
"epoch": 0.48,
|
| 187 |
+
"learning_rate": 4.193288157470152e-05,
|
| 188 |
+
"loss": 1.4688,
|
| 189 |
"step": 3000
|
| 190 |
},
|
| 191 |
{
|
| 192 |
+
"epoch": 0.5,
|
| 193 |
+
"learning_rate": 4.1663977627191566e-05,
|
| 194 |
+
"loss": 1.4711,
|
| 195 |
+
"step": 3100
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.52,
|
| 199 |
+
"learning_rate": 4.139507367968162e-05,
|
| 200 |
+
"loss": 1.4813,
|
| 201 |
+
"step": 3200
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.53,
|
| 205 |
+
"learning_rate": 4.112616973217167e-05,
|
| 206 |
+
"loss": 1.4541,
|
| 207 |
+
"step": 3300
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.55,
|
| 211 |
+
"learning_rate": 4.085726578466172e-05,
|
| 212 |
+
"loss": 1.4574,
|
| 213 |
+
"step": 3400
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.56,
|
| 217 |
+
"learning_rate": 4.058836183715177e-05,
|
| 218 |
+
"loss": 1.4634,
|
| 219 |
+
"step": 3500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 3.59,
|
| 223 |
+
"learning_rate": 1.407185628742515e-05,
|
| 224 |
+
"loss": 1.2902,
|
| 225 |
+
"step": 3600
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.69,
|
| 229 |
+
"learning_rate": 1.3073852295409184e-05,
|
| 230 |
+
"loss": 1.2773,
|
| 231 |
+
"step": 3700
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 3.79,
|
| 235 |
+
"learning_rate": 1.2075848303393214e-05,
|
| 236 |
+
"loss": 1.2647,
|
| 237 |
+
"step": 3800
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 3.89,
|
| 241 |
+
"learning_rate": 1.1077844311377246e-05,
|
| 242 |
+
"loss": 1.2524,
|
| 243 |
+
"step": 3900
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 3.99,
|
| 247 |
+
"learning_rate": 1.0079840319361278e-05,
|
| 248 |
+
"loss": 1.2543,
|
| 249 |
+
"step": 4000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 4.09,
|
| 253 |
+
"learning_rate": 9.08183632734531e-06,
|
| 254 |
+
"loss": 1.2499,
|
| 255 |
+
"step": 4100
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 4.19,
|
| 259 |
+
"learning_rate": 8.083832335329342e-06,
|
| 260 |
+
"loss": 1.2418,
|
| 261 |
+
"step": 4200
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 4.29,
|
| 265 |
+
"learning_rate": 7.085828343313373e-06,
|
| 266 |
+
"loss": 1.2379,
|
| 267 |
+
"step": 4300
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 4.39,
|
| 271 |
+
"learning_rate": 6.0878243512974054e-06,
|
| 272 |
+
"loss": 1.2325,
|
| 273 |
+
"step": 4400
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 4.49,
|
| 277 |
+
"learning_rate": 5.0898203592814375e-06,
|
| 278 |
+
"loss": 1.2246,
|
| 279 |
+
"step": 4500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 4.59,
|
| 283 |
+
"learning_rate": 4.091816367265469e-06,
|
| 284 |
+
"loss": 1.2331,
|
| 285 |
+
"step": 4600
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.69,
|
| 289 |
+
"learning_rate": 3.093812375249501e-06,
|
| 290 |
+
"loss": 1.2236,
|
| 291 |
+
"step": 4700
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 4.79,
|
| 295 |
+
"learning_rate": 2.095808383233533e-06,
|
| 296 |
+
"loss": 1.2281,
|
| 297 |
+
"step": 4800
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 4.89,
|
| 301 |
+
"learning_rate": 1.097804391217565e-06,
|
| 302 |
+
"loss": 1.2254,
|
| 303 |
+
"step": 4900
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 4.99,
|
| 307 |
+
"learning_rate": 9.98003992015968e-08,
|
| 308 |
+
"loss": 1.225,
|
| 309 |
+
"step": 5000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 5.0,
|
| 313 |
+
"step": 5010,
|
| 314 |
+
"total_flos": 8.106812063308723e+16,
|
| 315 |
+
"train_loss": 0.3749944605037362,
|
| 316 |
+
"train_runtime": 1805.8732,
|
| 317 |
+
"train_samples_per_second": 355.152,
|
| 318 |
+
"train_steps_per_second": 2.774
|
| 319 |
}
|
| 320 |
],
|
| 321 |
"logging_steps": 100,
|
| 322 |
+
"max_steps": 5010,
|
| 323 |
+
"num_train_epochs": 5,
|
| 324 |
"save_steps": 500,
|
| 325 |
+
"total_flos": 8.106812063308723e+16,
|
| 326 |
"trial_name": null,
|
| 327 |
"trial_params": null
|
| 328 |
}
|