End of training
Browse files- README.md +1 -1
- all_results.json +11 -11
- eval_results.json +6 -6
- runs/Jul27_18-38-59_14dd8a901eca/events.out.tfevents.1722123329.14dd8a901eca +3 -0
- train_results.json +6 -6
- trainer_state.json +397 -153
README.md
CHANGED
|
@@ -32,7 +32,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 32 |
|
| 33 |
This model is a fine-tuned version of [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on the imagefolder dataset.
|
| 34 |
It achieves the following results on the evaluation set:
|
| 35 |
-
- Loss: 0.
|
| 36 |
- Accuracy: 0.8694
|
| 37 |
|
| 38 |
## Model description
|
|
|
|
| 32 |
|
| 33 |
This model is a fine-tuned version of [microsoft/swin-tiny-patch4-window7-224](https://huggingface.co/microsoft/swin-tiny-patch4-window7-224) on the imagefolder dataset.
|
| 34 |
It achieves the following results on the evaluation set:
|
| 35 |
+
- Loss: 0.3468
|
| 36 |
- Accuracy: 0.8694
|
| 37 |
|
| 38 |
## Model description
|
all_results.json
CHANGED
|
@@ -1,13 +1,13 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 0.
|
| 8 |
-
"total_flos":
|
| 9 |
-
"train_loss": 0.
|
| 10 |
-
"train_runtime":
|
| 11 |
-
"train_samples_per_second":
|
| 12 |
-
"train_steps_per_second": 0.
|
| 13 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.8693982074263764,
|
| 4 |
+
"eval_loss": 0.34680071473121643,
|
| 5 |
+
"eval_runtime": 75.6434,
|
| 6 |
+
"eval_samples_per_second": 10.325,
|
| 7 |
+
"eval_steps_per_second": 0.093,
|
| 8 |
+
"total_flos": 4.783917310653358e+18,
|
| 9 |
+
"train_loss": 0.3464228366550646,
|
| 10 |
+
"train_runtime": 17694.0061,
|
| 11 |
+
"train_samples_per_second": 10.877,
|
| 12 |
+
"train_steps_per_second": 0.021
|
| 13 |
}
|
eval_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
-
"eval_samples_per_second":
|
| 7 |
-
"eval_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_accuracy": 0.8693982074263764,
|
| 4 |
+
"eval_loss": 0.34680071473121643,
|
| 5 |
+
"eval_runtime": 75.6434,
|
| 6 |
+
"eval_samples_per_second": 10.325,
|
| 7 |
+
"eval_steps_per_second": 0.093
|
| 8 |
}
|
runs/Jul27_18-38-59_14dd8a901eca/events.out.tfevents.1722123329.14dd8a901eca
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9bf143752d02deccef806309f01bf7c31f6e450f5d36781ca075bba488a569f9
|
| 3 |
+
size 88
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"total_flos": 4.783917310653358e+18,
|
| 4 |
+
"train_loss": 0.3464228366550646,
|
| 5 |
+
"train_runtime": 17694.0061,
|
| 6 |
+
"train_samples_per_second": 10.877,
|
| 7 |
+
"train_steps_per_second": 0.021
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,229 +1,473 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "swin-tiny-patch4-window7-224-MM_Classification/checkpoint-
|
| 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 |
-
"grad_norm":
|
| 14 |
-
"learning_rate":
|
| 15 |
-
"loss": 1.
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
-
"epoch": 0
|
| 20 |
-
"eval_accuracy": 0.
|
| 21 |
-
"eval_loss": 0.
|
| 22 |
-
"eval_runtime":
|
| 23 |
-
"eval_samples_per_second": 7.
|
| 24 |
-
"eval_steps_per_second": 0.
|
| 25 |
-
"step":
|
| 26 |
},
|
| 27 |
{
|
| 28 |
-
"epoch": 1.
|
| 29 |
-
"grad_norm": 3.
|
| 30 |
-
"learning_rate":
|
| 31 |
-
"loss": 0.
|
| 32 |
"step": 20
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch": 1.
|
| 36 |
-
"grad_norm":
|
| 37 |
-
"learning_rate":
|
| 38 |
-
"loss": 0.
|
| 39 |
"step": 30
|
| 40 |
},
|
| 41 |
{
|
| 42 |
-
"epoch":
|
| 43 |
-
"eval_accuracy": 0.
|
| 44 |
-
"eval_loss": 0.
|
| 45 |
-
"eval_runtime":
|
| 46 |
-
"eval_samples_per_second":
|
| 47 |
-
"eval_steps_per_second": 0.
|
| 48 |
-
"step":
|
| 49 |
},
|
| 50 |
{
|
| 51 |
-
"epoch": 2.
|
| 52 |
-
"grad_norm":
|
| 53 |
-
"learning_rate": 4.
|
| 54 |
-
"loss": 0.
|
| 55 |
"step": 40
|
| 56 |
},
|
| 57 |
{
|
| 58 |
-
"epoch": 2.
|
| 59 |
-
"
|
| 60 |
-
"
|
| 61 |
-
"
|
| 62 |
-
"
|
| 63 |
-
"eval_steps_per_second": 0.062,
|
| 64 |
-
"step": 48
|
| 65 |
},
|
| 66 |
{
|
| 67 |
-
"epoch": 3.
|
| 68 |
-
"
|
| 69 |
-
"
|
| 70 |
-
"
|
| 71 |
-
"
|
|
|
|
|
|
|
| 72 |
},
|
| 73 |
{
|
| 74 |
-
"epoch": 3.
|
| 75 |
-
"grad_norm": 2.
|
| 76 |
-
"learning_rate":
|
| 77 |
-
"loss": 0.
|
| 78 |
"step": 60
|
| 79 |
},
|
| 80 |
{
|
| 81 |
-
"epoch":
|
| 82 |
-
"
|
| 83 |
-
"
|
| 84 |
-
"
|
| 85 |
-
"eval_samples_per_second": 7.154,
|
| 86 |
-
"eval_steps_per_second": 0.062,
|
| 87 |
-
"step": 65
|
| 88 |
-
},
|
| 89 |
-
{
|
| 90 |
-
"epoch": 4.3076923076923075,
|
| 91 |
-
"grad_norm": 2.523513078689575,
|
| 92 |
-
"learning_rate": 3.125e-05,
|
| 93 |
-
"loss": 0.3079,
|
| 94 |
"step": 70
|
| 95 |
},
|
| 96 |
{
|
| 97 |
-
"epoch": 4.
|
| 98 |
-
"
|
| 99 |
-
"
|
| 100 |
-
"
|
| 101 |
-
"
|
|
|
|
|
|
|
| 102 |
},
|
| 103 |
{
|
| 104 |
-
"epoch": 4.
|
| 105 |
-
"
|
| 106 |
-
"
|
| 107 |
-
"
|
| 108 |
-
"
|
| 109 |
-
"eval_steps_per_second": 0.063,
|
| 110 |
-
"step": 81
|
| 111 |
},
|
| 112 |
{
|
| 113 |
-
"epoch":
|
| 114 |
-
"grad_norm":
|
| 115 |
-
"learning_rate":
|
| 116 |
-
"loss": 0.
|
| 117 |
"step": 90
|
| 118 |
},
|
| 119 |
{
|
| 120 |
-
"epoch": 5.
|
| 121 |
-
"eval_accuracy": 0.
|
| 122 |
-
"eval_loss": 0.
|
| 123 |
-
"eval_runtime":
|
| 124 |
-
"eval_samples_per_second":
|
| 125 |
-
"eval_steps_per_second": 0.
|
| 126 |
-
"step":
|
| 127 |
},
|
| 128 |
{
|
| 129 |
-
"epoch":
|
| 130 |
-
"grad_norm":
|
| 131 |
-
"learning_rate":
|
| 132 |
-
"loss": 0.
|
| 133 |
"step": 100
|
| 134 |
},
|
| 135 |
{
|
| 136 |
-
"epoch":
|
| 137 |
-
"grad_norm":
|
| 138 |
-
"learning_rate":
|
| 139 |
-
"loss": 0.
|
| 140 |
"step": 110
|
| 141 |
},
|
| 142 |
{
|
| 143 |
-
"epoch": 6.
|
| 144 |
-
"eval_accuracy": 0.
|
| 145 |
-
"eval_loss": 0.
|
| 146 |
-
"eval_runtime":
|
| 147 |
-
"eval_samples_per_second":
|
| 148 |
-
"eval_steps_per_second": 0.
|
| 149 |
-
"step":
|
| 150 |
},
|
| 151 |
{
|
| 152 |
-
"epoch":
|
| 153 |
-
"grad_norm": 2.
|
| 154 |
-
"learning_rate":
|
| 155 |
-
"loss": 0.
|
| 156 |
"step": 120
|
| 157 |
},
|
| 158 |
{
|
| 159 |
-
"epoch":
|
| 160 |
-
"grad_norm":
|
| 161 |
-
"learning_rate":
|
| 162 |
-
"loss": 0.
|
| 163 |
"step": 130
|
| 164 |
},
|
| 165 |
{
|
| 166 |
-
"epoch":
|
| 167 |
-
"eval_accuracy": 0.
|
| 168 |
-
"eval_loss": 0.
|
| 169 |
-
"eval_runtime":
|
| 170 |
-
"eval_samples_per_second":
|
| 171 |
-
"eval_steps_per_second": 0.
|
| 172 |
-
"step":
|
| 173 |
},
|
| 174 |
{
|
| 175 |
-
"epoch":
|
| 176 |
-
"grad_norm": 2.
|
| 177 |
-
"learning_rate":
|
| 178 |
-
"loss": 0.
|
| 179 |
"step": 140
|
| 180 |
},
|
| 181 |
{
|
| 182 |
-
"epoch":
|
| 183 |
-
"
|
| 184 |
-
"
|
| 185 |
-
"
|
| 186 |
-
"
|
| 187 |
-
"eval_steps_per_second": 0.06,
|
| 188 |
-
"step": 146
|
| 189 |
},
|
| 190 |
{
|
| 191 |
-
"epoch":
|
| 192 |
-
"
|
| 193 |
-
"
|
| 194 |
-
"
|
| 195 |
-
"
|
|
|
|
|
|
|
| 196 |
},
|
| 197 |
{
|
| 198 |
-
"epoch":
|
| 199 |
-
"grad_norm": 3.
|
| 200 |
-
"learning_rate":
|
| 201 |
-
"loss": 0.
|
| 202 |
"step": 160
|
| 203 |
},
|
| 204 |
{
|
| 205 |
-
"epoch":
|
| 206 |
-
"
|
| 207 |
-
"
|
| 208 |
-
"
|
| 209 |
-
"
|
| 210 |
-
|
| 211 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 212 |
},
|
| 213 |
{
|
| 214 |
-
"epoch":
|
| 215 |
-
"step":
|
| 216 |
-
"total_flos":
|
| 217 |
-
"train_loss": 0.
|
| 218 |
-
"train_runtime":
|
| 219 |
-
"train_samples_per_second":
|
| 220 |
-
"train_steps_per_second": 0.
|
| 221 |
}
|
| 222 |
],
|
| 223 |
"logging_steps": 10,
|
| 224 |
-
"max_steps":
|
| 225 |
"num_input_tokens_seen": 0,
|
| 226 |
-
"num_train_epochs":
|
| 227 |
"save_steps": 500,
|
| 228 |
"stateful_callbacks": {
|
| 229 |
"TrainerControl": {
|
|
@@ -237,7 +481,7 @@
|
|
| 237 |
"attributes": {}
|
| 238 |
}
|
| 239 |
},
|
| 240 |
-
"total_flos":
|
| 241 |
"train_batch_size": 128,
|
| 242 |
"trial_name": null,
|
| 243 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.8693982074263764,
|
| 3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-MM_Classification/checkpoint-361",
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 380,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.5263157894736842,
|
| 13 |
+
"grad_norm": 3.3412868976593018,
|
| 14 |
+
"learning_rate": 1.3157894736842106e-05,
|
| 15 |
+
"loss": 1.0476,
|
| 16 |
"step": 10
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 1.0,
|
| 20 |
+
"eval_accuracy": 0.6530089628681178,
|
| 21 |
+
"eval_loss": 0.7707358002662659,
|
| 22 |
+
"eval_runtime": 104.799,
|
| 23 |
+
"eval_samples_per_second": 7.452,
|
| 24 |
+
"eval_steps_per_second": 0.067,
|
| 25 |
+
"step": 19
|
| 26 |
},
|
| 27 |
{
|
| 28 |
+
"epoch": 1.0526315789473684,
|
| 29 |
+
"grad_norm": 3.460216999053955,
|
| 30 |
+
"learning_rate": 2.6315789473684212e-05,
|
| 31 |
+
"loss": 0.7869,
|
| 32 |
"step": 20
|
| 33 |
},
|
| 34 |
{
|
| 35 |
+
"epoch": 1.5789473684210527,
|
| 36 |
+
"grad_norm": 3.0562336444854736,
|
| 37 |
+
"learning_rate": 3.9473684210526316e-05,
|
| 38 |
+
"loss": 0.6226,
|
| 39 |
"step": 30
|
| 40 |
},
|
| 41 |
{
|
| 42 |
+
"epoch": 2.0,
|
| 43 |
+
"eval_accuracy": 0.8104993597951344,
|
| 44 |
+
"eval_loss": 0.47430700063705444,
|
| 45 |
+
"eval_runtime": 75.387,
|
| 46 |
+
"eval_samples_per_second": 10.36,
|
| 47 |
+
"eval_steps_per_second": 0.093,
|
| 48 |
+
"step": 38
|
| 49 |
},
|
| 50 |
{
|
| 51 |
+
"epoch": 2.1052631578947367,
|
| 52 |
+
"grad_norm": 3.6400160789489746,
|
| 53 |
+
"learning_rate": 4.970760233918128e-05,
|
| 54 |
+
"loss": 0.5102,
|
| 55 |
"step": 40
|
| 56 |
},
|
| 57 |
{
|
| 58 |
+
"epoch": 2.6315789473684212,
|
| 59 |
+
"grad_norm": 2.88069486618042,
|
| 60 |
+
"learning_rate": 4.824561403508772e-05,
|
| 61 |
+
"loss": 0.4477,
|
| 62 |
+
"step": 50
|
|
|
|
|
|
|
| 63 |
},
|
| 64 |
{
|
| 65 |
+
"epoch": 3.0,
|
| 66 |
+
"eval_accuracy": 0.8322663252240717,
|
| 67 |
+
"eval_loss": 0.41332316398620605,
|
| 68 |
+
"eval_runtime": 75.6871,
|
| 69 |
+
"eval_samples_per_second": 10.319,
|
| 70 |
+
"eval_steps_per_second": 0.092,
|
| 71 |
+
"step": 57
|
| 72 |
},
|
| 73 |
{
|
| 74 |
+
"epoch": 3.1578947368421053,
|
| 75 |
+
"grad_norm": 2.5295557975769043,
|
| 76 |
+
"learning_rate": 4.678362573099415e-05,
|
| 77 |
+
"loss": 0.4205,
|
| 78 |
"step": 60
|
| 79 |
},
|
| 80 |
{
|
| 81 |
+
"epoch": 3.6842105263157894,
|
| 82 |
+
"grad_norm": 2.7619762420654297,
|
| 83 |
+
"learning_rate": 4.5321637426900585e-05,
|
| 84 |
+
"loss": 0.3963,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 85 |
"step": 70
|
| 86 |
},
|
| 87 |
{
|
| 88 |
+
"epoch": 4.0,
|
| 89 |
+
"eval_accuracy": 0.8476312419974392,
|
| 90 |
+
"eval_loss": 0.38133054971694946,
|
| 91 |
+
"eval_runtime": 75.5986,
|
| 92 |
+
"eval_samples_per_second": 10.331,
|
| 93 |
+
"eval_steps_per_second": 0.093,
|
| 94 |
+
"step": 76
|
| 95 |
},
|
| 96 |
{
|
| 97 |
+
"epoch": 4.2105263157894735,
|
| 98 |
+
"grad_norm": 2.971311330795288,
|
| 99 |
+
"learning_rate": 4.3859649122807014e-05,
|
| 100 |
+
"loss": 0.3788,
|
| 101 |
+
"step": 80
|
|
|
|
|
|
|
| 102 |
},
|
| 103 |
{
|
| 104 |
+
"epoch": 4.7368421052631575,
|
| 105 |
+
"grad_norm": 3.20125412940979,
|
| 106 |
+
"learning_rate": 4.239766081871345e-05,
|
| 107 |
+
"loss": 0.3694,
|
| 108 |
"step": 90
|
| 109 |
},
|
| 110 |
{
|
| 111 |
+
"epoch": 5.0,
|
| 112 |
+
"eval_accuracy": 0.8540332906530089,
|
| 113 |
+
"eval_loss": 0.37533658742904663,
|
| 114 |
+
"eval_runtime": 74.9213,
|
| 115 |
+
"eval_samples_per_second": 10.424,
|
| 116 |
+
"eval_steps_per_second": 0.093,
|
| 117 |
+
"step": 95
|
| 118 |
},
|
| 119 |
{
|
| 120 |
+
"epoch": 5.2631578947368425,
|
| 121 |
+
"grad_norm": 3.091187000274658,
|
| 122 |
+
"learning_rate": 4.093567251461988e-05,
|
| 123 |
+
"loss": 0.3499,
|
| 124 |
"step": 100
|
| 125 |
},
|
| 126 |
{
|
| 127 |
+
"epoch": 5.7894736842105265,
|
| 128 |
+
"grad_norm": 3.088123321533203,
|
| 129 |
+
"learning_rate": 3.9473684210526316e-05,
|
| 130 |
+
"loss": 0.3451,
|
| 131 |
"step": 110
|
| 132 |
},
|
| 133 |
{
|
| 134 |
+
"epoch": 6.0,
|
| 135 |
+
"eval_accuracy": 0.8489116517285531,
|
| 136 |
+
"eval_loss": 0.3586506247520447,
|
| 137 |
+
"eval_runtime": 75.7408,
|
| 138 |
+
"eval_samples_per_second": 10.311,
|
| 139 |
+
"eval_steps_per_second": 0.092,
|
| 140 |
+
"step": 114
|
| 141 |
},
|
| 142 |
{
|
| 143 |
+
"epoch": 6.315789473684211,
|
| 144 |
+
"grad_norm": 2.5398190021514893,
|
| 145 |
+
"learning_rate": 3.8011695906432746e-05,
|
| 146 |
+
"loss": 0.3421,
|
| 147 |
"step": 120
|
| 148 |
},
|
| 149 |
{
|
| 150 |
+
"epoch": 6.842105263157895,
|
| 151 |
+
"grad_norm": 3.2862489223480225,
|
| 152 |
+
"learning_rate": 3.654970760233918e-05,
|
| 153 |
+
"loss": 0.3382,
|
| 154 |
"step": 130
|
| 155 |
},
|
| 156 |
{
|
| 157 |
+
"epoch": 7.0,
|
| 158 |
+
"eval_accuracy": 0.8450704225352113,
|
| 159 |
+
"eval_loss": 0.3531467616558075,
|
| 160 |
+
"eval_runtime": 75.2863,
|
| 161 |
+
"eval_samples_per_second": 10.374,
|
| 162 |
+
"eval_steps_per_second": 0.093,
|
| 163 |
+
"step": 133
|
| 164 |
},
|
| 165 |
{
|
| 166 |
+
"epoch": 7.368421052631579,
|
| 167 |
+
"grad_norm": 2.175835609436035,
|
| 168 |
+
"learning_rate": 3.508771929824561e-05,
|
| 169 |
+
"loss": 0.3397,
|
| 170 |
"step": 140
|
| 171 |
},
|
| 172 |
{
|
| 173 |
+
"epoch": 7.894736842105263,
|
| 174 |
+
"grad_norm": 2.2023236751556396,
|
| 175 |
+
"learning_rate": 3.362573099415205e-05,
|
| 176 |
+
"loss": 0.3253,
|
| 177 |
+
"step": 150
|
|
|
|
|
|
|
| 178 |
},
|
| 179 |
{
|
| 180 |
+
"epoch": 8.0,
|
| 181 |
+
"eval_accuracy": 0.8578745198463509,
|
| 182 |
+
"eval_loss": 0.34979528188705444,
|
| 183 |
+
"eval_runtime": 75.6534,
|
| 184 |
+
"eval_samples_per_second": 10.323,
|
| 185 |
+
"eval_steps_per_second": 0.093,
|
| 186 |
+
"step": 152
|
| 187 |
},
|
| 188 |
{
|
| 189 |
+
"epoch": 8.421052631578947,
|
| 190 |
+
"grad_norm": 3.327239513397217,
|
| 191 |
+
"learning_rate": 3.216374269005848e-05,
|
| 192 |
+
"loss": 0.3156,
|
| 193 |
"step": 160
|
| 194 |
},
|
| 195 |
{
|
| 196 |
+
"epoch": 8.947368421052632,
|
| 197 |
+
"grad_norm": 3.0708658695220947,
|
| 198 |
+
"learning_rate": 3.0701754385964913e-05,
|
| 199 |
+
"loss": 0.3121,
|
| 200 |
+
"step": 170
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 9.0,
|
| 204 |
+
"eval_accuracy": 0.8578745198463509,
|
| 205 |
+
"eval_loss": 0.34373539686203003,
|
| 206 |
+
"eval_runtime": 74.9284,
|
| 207 |
+
"eval_samples_per_second": 10.423,
|
| 208 |
+
"eval_steps_per_second": 0.093,
|
| 209 |
+
"step": 171
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 9.473684210526315,
|
| 213 |
+
"grad_norm": 2.4968678951263428,
|
| 214 |
+
"learning_rate": 2.9239766081871346e-05,
|
| 215 |
+
"loss": 0.3042,
|
| 216 |
+
"step": 180
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 10.0,
|
| 220 |
+
"grad_norm": 2.865316152572632,
|
| 221 |
+
"learning_rate": 2.777777777777778e-05,
|
| 222 |
+
"loss": 0.2855,
|
| 223 |
+
"step": 190
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 10.0,
|
| 227 |
+
"eval_accuracy": 0.8655569782330346,
|
| 228 |
+
"eval_loss": 0.3446912467479706,
|
| 229 |
+
"eval_runtime": 75.4844,
|
| 230 |
+
"eval_samples_per_second": 10.347,
|
| 231 |
+
"eval_steps_per_second": 0.093,
|
| 232 |
+
"step": 190
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 10.526315789473685,
|
| 236 |
+
"grad_norm": 3.000030279159546,
|
| 237 |
+
"learning_rate": 2.6315789473684212e-05,
|
| 238 |
+
"loss": 0.2961,
|
| 239 |
+
"step": 200
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 11.0,
|
| 243 |
+
"eval_accuracy": 0.8617157490396927,
|
| 244 |
+
"eval_loss": 0.3350251019001007,
|
| 245 |
+
"eval_runtime": 74.9625,
|
| 246 |
+
"eval_samples_per_second": 10.419,
|
| 247 |
+
"eval_steps_per_second": 0.093,
|
| 248 |
+
"step": 209
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 11.052631578947368,
|
| 252 |
+
"grad_norm": 3.1556169986724854,
|
| 253 |
+
"learning_rate": 2.485380116959064e-05,
|
| 254 |
+
"loss": 0.291,
|
| 255 |
+
"step": 210
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 11.578947368421053,
|
| 259 |
+
"grad_norm": 2.82590389251709,
|
| 260 |
+
"learning_rate": 2.3391812865497074e-05,
|
| 261 |
+
"loss": 0.273,
|
| 262 |
+
"step": 220
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 12.0,
|
| 266 |
+
"eval_accuracy": 0.8565941101152369,
|
| 267 |
+
"eval_loss": 0.34841132164001465,
|
| 268 |
+
"eval_runtime": 75.6499,
|
| 269 |
+
"eval_samples_per_second": 10.324,
|
| 270 |
+
"eval_steps_per_second": 0.093,
|
| 271 |
+
"step": 228
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 12.105263157894736,
|
| 275 |
+
"grad_norm": 2.110739231109619,
|
| 276 |
+
"learning_rate": 2.1929824561403507e-05,
|
| 277 |
+
"loss": 0.2767,
|
| 278 |
+
"step": 230
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 12.631578947368421,
|
| 282 |
+
"grad_norm": 2.7739641666412354,
|
| 283 |
+
"learning_rate": 2.046783625730994e-05,
|
| 284 |
+
"loss": 0.2745,
|
| 285 |
+
"step": 240
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 13.0,
|
| 289 |
+
"eval_accuracy": 0.8604353393085787,
|
| 290 |
+
"eval_loss": 0.34332236647605896,
|
| 291 |
+
"eval_runtime": 75.3714,
|
| 292 |
+
"eval_samples_per_second": 10.362,
|
| 293 |
+
"eval_steps_per_second": 0.093,
|
| 294 |
+
"step": 247
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 13.157894736842104,
|
| 298 |
+
"grad_norm": 2.834440231323242,
|
| 299 |
+
"learning_rate": 1.9005847953216373e-05,
|
| 300 |
+
"loss": 0.2678,
|
| 301 |
+
"step": 250
|
| 302 |
+
},
|
| 303 |
+
{
|
| 304 |
+
"epoch": 13.68421052631579,
|
| 305 |
+
"grad_norm": 4.303690433502197,
|
| 306 |
+
"learning_rate": 1.7543859649122806e-05,
|
| 307 |
+
"loss": 0.2613,
|
| 308 |
+
"step": 260
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 14.0,
|
| 312 |
+
"eval_accuracy": 0.8642765685019206,
|
| 313 |
+
"eval_loss": 0.34982678294181824,
|
| 314 |
+
"eval_runtime": 76.0549,
|
| 315 |
+
"eval_samples_per_second": 10.269,
|
| 316 |
+
"eval_steps_per_second": 0.092,
|
| 317 |
+
"step": 266
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 14.210526315789474,
|
| 321 |
+
"grad_norm": 3.9326910972595215,
|
| 322 |
+
"learning_rate": 1.608187134502924e-05,
|
| 323 |
+
"loss": 0.2713,
|
| 324 |
+
"step": 270
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 14.736842105263158,
|
| 328 |
+
"grad_norm": 3.0511579513549805,
|
| 329 |
+
"learning_rate": 1.4619883040935673e-05,
|
| 330 |
+
"loss": 0.2527,
|
| 331 |
+
"step": 280
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 15.0,
|
| 335 |
+
"eval_accuracy": 0.8578745198463509,
|
| 336 |
+
"eval_loss": 0.33652085065841675,
|
| 337 |
+
"eval_runtime": 78.4572,
|
| 338 |
+
"eval_samples_per_second": 9.954,
|
| 339 |
+
"eval_steps_per_second": 0.089,
|
| 340 |
+
"step": 285
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 15.263157894736842,
|
| 344 |
+
"grad_norm": 2.9660778045654297,
|
| 345 |
+
"learning_rate": 1.3157894736842106e-05,
|
| 346 |
+
"loss": 0.246,
|
| 347 |
+
"step": 290
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 15.789473684210526,
|
| 351 |
+
"grad_norm": 2.621548652648926,
|
| 352 |
+
"learning_rate": 1.1695906432748537e-05,
|
| 353 |
+
"loss": 0.2619,
|
| 354 |
+
"step": 300
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 16.0,
|
| 358 |
+
"eval_accuracy": 0.8617157490396927,
|
| 359 |
+
"eval_loss": 0.3450033366680145,
|
| 360 |
+
"eval_runtime": 78.025,
|
| 361 |
+
"eval_samples_per_second": 10.01,
|
| 362 |
+
"eval_steps_per_second": 0.09,
|
| 363 |
+
"step": 304
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 16.31578947368421,
|
| 367 |
+
"grad_norm": 2.7999181747436523,
|
| 368 |
+
"learning_rate": 1.023391812865497e-05,
|
| 369 |
+
"loss": 0.2469,
|
| 370 |
+
"step": 310
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 16.842105263157894,
|
| 374 |
+
"grad_norm": 2.3670365810394287,
|
| 375 |
+
"learning_rate": 8.771929824561403e-06,
|
| 376 |
+
"loss": 0.2436,
|
| 377 |
+
"step": 320
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 17.0,
|
| 381 |
+
"eval_accuracy": 0.8681177976952625,
|
| 382 |
+
"eval_loss": 0.34535887837409973,
|
| 383 |
+
"eval_runtime": 78.3498,
|
| 384 |
+
"eval_samples_per_second": 9.968,
|
| 385 |
+
"eval_steps_per_second": 0.089,
|
| 386 |
+
"step": 323
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 17.36842105263158,
|
| 390 |
+
"grad_norm": 2.8293299674987793,
|
| 391 |
+
"learning_rate": 7.3099415204678366e-06,
|
| 392 |
+
"loss": 0.2457,
|
| 393 |
+
"step": 330
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 17.894736842105264,
|
| 397 |
+
"grad_norm": 2.825676441192627,
|
| 398 |
+
"learning_rate": 5.8479532163742686e-06,
|
| 399 |
+
"loss": 0.2518,
|
| 400 |
+
"step": 340
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 18.0,
|
| 404 |
+
"eval_accuracy": 0.8681177976952625,
|
| 405 |
+
"eval_loss": 0.3437488377094269,
|
| 406 |
+
"eval_runtime": 77.0059,
|
| 407 |
+
"eval_samples_per_second": 10.142,
|
| 408 |
+
"eval_steps_per_second": 0.091,
|
| 409 |
+
"step": 342
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 18.42105263157895,
|
| 413 |
+
"grad_norm": 2.9020776748657227,
|
| 414 |
+
"learning_rate": 4.3859649122807014e-06,
|
| 415 |
+
"loss": 0.2362,
|
| 416 |
+
"step": 350
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 18.94736842105263,
|
| 420 |
+
"grad_norm": 2.6419830322265625,
|
| 421 |
+
"learning_rate": 2.9239766081871343e-06,
|
| 422 |
+
"loss": 0.243,
|
| 423 |
+
"step": 360
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 19.0,
|
| 427 |
+
"eval_accuracy": 0.8693982074263764,
|
| 428 |
+
"eval_loss": 0.34680071473121643,
|
| 429 |
+
"eval_runtime": 76.7708,
|
| 430 |
+
"eval_samples_per_second": 10.173,
|
| 431 |
+
"eval_steps_per_second": 0.091,
|
| 432 |
+
"step": 361
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 19.473684210526315,
|
| 436 |
+
"grad_norm": 2.185123920440674,
|
| 437 |
+
"learning_rate": 1.4619883040935671e-06,
|
| 438 |
+
"loss": 0.2452,
|
| 439 |
+
"step": 370
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 20.0,
|
| 443 |
+
"grad_norm": 4.370210647583008,
|
| 444 |
+
"learning_rate": 0.0,
|
| 445 |
+
"loss": 0.2415,
|
| 446 |
+
"step": 380
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 20.0,
|
| 450 |
+
"eval_accuracy": 0.8693982074263764,
|
| 451 |
+
"eval_loss": 0.3454751670360565,
|
| 452 |
+
"eval_runtime": 76.6964,
|
| 453 |
+
"eval_samples_per_second": 10.183,
|
| 454 |
+
"eval_steps_per_second": 0.091,
|
| 455 |
+
"step": 380
|
| 456 |
},
|
| 457 |
{
|
| 458 |
+
"epoch": 20.0,
|
| 459 |
+
"step": 380,
|
| 460 |
+
"total_flos": 4.783917310653358e+18,
|
| 461 |
+
"train_loss": 0.3464228366550646,
|
| 462 |
+
"train_runtime": 17694.0061,
|
| 463 |
+
"train_samples_per_second": 10.877,
|
| 464 |
+
"train_steps_per_second": 0.021
|
| 465 |
}
|
| 466 |
],
|
| 467 |
"logging_steps": 10,
|
| 468 |
+
"max_steps": 380,
|
| 469 |
"num_input_tokens_seen": 0,
|
| 470 |
+
"num_train_epochs": 20,
|
| 471 |
"save_steps": 500,
|
| 472 |
"stateful_callbacks": {
|
| 473 |
"TrainerControl": {
|
|
|
|
| 481 |
"attributes": {}
|
| 482 |
}
|
| 483 |
},
|
| 484 |
+
"total_flos": 4.783917310653358e+18,
|
| 485 |
"train_batch_size": 128,
|
| 486 |
"trial_name": null,
|
| 487 |
"trial_params": null
|