Upload folder using huggingface_hub
Browse files- adapter_model.bin +1 -1
- all_results.json +14 -14
- eval_results.json +9 -9
- train_results.json +5 -5
- trainer_state.json +152 -63
adapter_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 160086542
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:247fc2b9ff6d183a31785d3f2a3b4287662c9c74c22d00915bfc7758ee61e6a1
|
| 3 |
size 160086542
|
all_results.json
CHANGED
|
@@ -1,17 +1,17 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_f1_score": 0.
|
| 5 |
-
"eval_gmean": 0.
|
| 6 |
-
"eval_loss": 0.
|
| 7 |
-
"eval_precision": 0.
|
| 8 |
-
"eval_recall": 0.
|
| 9 |
-
"eval_runtime":
|
| 10 |
-
"eval_samples_per_second": 0.
|
| 11 |
-
"eval_steps_per_second": 0.
|
| 12 |
-
"total_flos":
|
| 13 |
-
"train_loss": 0.
|
| 14 |
-
"train_runtime":
|
| 15 |
-
"train_samples_per_second":
|
| 16 |
-
"train_steps_per_second": 0.
|
| 17 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"eval_accuracy": 0.7666666666666667,
|
| 4 |
+
"eval_f1_score": 0.7609087348217781,
|
| 5 |
+
"eval_gmean": 0.7486572528549951,
|
| 6 |
+
"eval_loss": 0.5165690183639526,
|
| 7 |
+
"eval_precision": 0.7716666666666666,
|
| 8 |
+
"eval_recall": 0.7666666666666667,
|
| 9 |
+
"eval_runtime": 170.2423,
|
| 10 |
+
"eval_samples_per_second": 0.352,
|
| 11 |
+
"eval_steps_per_second": 0.047,
|
| 12 |
+
"total_flos": 2.597706419798016e+16,
|
| 13 |
+
"train_loss": 0.7354364809782609,
|
| 14 |
+
"train_runtime": 6240.8638,
|
| 15 |
+
"train_samples_per_second": 2.838,
|
| 16 |
+
"train_steps_per_second": 0.022
|
| 17 |
}
|
eval_results.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_f1_score": 0.
|
| 5 |
-
"eval_gmean": 0.
|
| 6 |
-
"eval_loss": 0.
|
| 7 |
-
"eval_precision": 0.
|
| 8 |
-
"eval_recall": 0.
|
| 9 |
-
"eval_runtime":
|
| 10 |
-
"eval_samples_per_second": 0.
|
| 11 |
-
"eval_steps_per_second": 0.
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 3 |
+
"eval_accuracy": 0.7666666666666667,
|
| 4 |
+
"eval_f1_score": 0.7609087348217781,
|
| 5 |
+
"eval_gmean": 0.7486572528549951,
|
| 6 |
+
"eval_loss": 0.5165690183639526,
|
| 7 |
+
"eval_precision": 0.7716666666666666,
|
| 8 |
+
"eval_recall": 0.7666666666666667,
|
| 9 |
+
"eval_runtime": 170.2423,
|
| 10 |
+
"eval_samples_per_second": 0.352,
|
| 11 |
+
"eval_steps_per_second": 0.047
|
| 12 |
}
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
"epoch": 1.0,
|
| 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": 1.0,
|
| 3 |
+
"total_flos": 2.597706419798016e+16,
|
| 4 |
+
"train_loss": 0.7354364809782609,
|
| 5 |
+
"train_runtime": 6240.8638,
|
| 6 |
+
"train_samples_per_second": 2.838,
|
| 7 |
+
"train_steps_per_second": 0.022
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,112 +1,201 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 28,
|
| 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": 1.
|
| 15 |
-
"loss": 0.
|
| 16 |
"step": 8
|
| 17 |
},
|
| 18 |
{
|
| 19 |
-
"epoch": 0.
|
| 20 |
-
"grad_norm":
|
| 21 |
-
"learning_rate": 1.
|
| 22 |
-
"loss": 0.
|
| 23 |
"step": 16
|
| 24 |
},
|
| 25 |
{
|
| 26 |
-
"epoch": 0.
|
| 27 |
-
"grad_norm":
|
| 28 |
-
"learning_rate": 1.
|
| 29 |
-
"loss": 0.
|
| 30 |
"step": 24
|
| 31 |
},
|
| 32 |
{
|
| 33 |
-
"epoch": 0.
|
| 34 |
-
"eval_accuracy": 0.
|
| 35 |
-
"eval_f1_score": 0.
|
| 36 |
-
"eval_gmean": 0.
|
| 37 |
-
"eval_loss": 0.
|
| 38 |
-
"eval_precision": 0.
|
| 39 |
-
"eval_recall": 0.
|
| 40 |
-
"eval_runtime":
|
| 41 |
-
"eval_samples_per_second": 0.
|
| 42 |
-
"eval_steps_per_second": 0.
|
| 43 |
"step": 28
|
| 44 |
},
|
| 45 |
{
|
| 46 |
-
"epoch": 0.
|
| 47 |
-
"grad_norm":
|
| 48 |
-
"learning_rate": 1.
|
| 49 |
-
"loss": 0.
|
| 50 |
"step": 32
|
| 51 |
},
|
| 52 |
{
|
| 53 |
-
"epoch": 0.
|
| 54 |
-
"grad_norm": 7.
|
| 55 |
-
"learning_rate":
|
| 56 |
-
"loss": 0.
|
| 57 |
"step": 40
|
| 58 |
},
|
| 59 |
{
|
| 60 |
-
"epoch": 0.
|
| 61 |
-
"grad_norm":
|
| 62 |
-
"learning_rate":
|
| 63 |
-
"loss": 0.
|
| 64 |
"step": 48
|
| 65 |
},
|
| 66 |
{
|
| 67 |
-
"epoch": 0.
|
| 68 |
-
"grad_norm":
|
| 69 |
-
"learning_rate":
|
| 70 |
-
"loss": 0.
|
| 71 |
"step": 56
|
| 72 |
},
|
| 73 |
{
|
| 74 |
-
"epoch": 0.
|
| 75 |
-
"eval_accuracy": 0.
|
| 76 |
-
"eval_f1_score": 0.
|
| 77 |
-
"eval_gmean": 0.
|
| 78 |
-
"eval_loss": 0.
|
| 79 |
-
"eval_precision": 0.
|
| 80 |
-
"eval_recall": 0.
|
| 81 |
-
"eval_runtime":
|
| 82 |
-
"eval_samples_per_second": 0.
|
| 83 |
-
"eval_steps_per_second": 0.
|
| 84 |
"step": 56
|
| 85 |
},
|
| 86 |
{
|
| 87 |
-
"epoch": 0.
|
| 88 |
-
"grad_norm":
|
| 89 |
-
"learning_rate":
|
| 90 |
-
"loss": 0.
|
| 91 |
"step": 64
|
| 92 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93 |
{
|
| 94 |
"epoch": 1.0,
|
| 95 |
-
"step":
|
| 96 |
-
"total_flos":
|
| 97 |
-
"train_loss": 0.
|
| 98 |
-
"train_runtime":
|
| 99 |
-
"train_samples_per_second":
|
| 100 |
-
"train_steps_per_second": 0.
|
| 101 |
}
|
| 102 |
],
|
| 103 |
"logging_steps": 8,
|
| 104 |
-
"max_steps":
|
| 105 |
"num_input_tokens_seen": 0,
|
| 106 |
"num_train_epochs": 1,
|
| 107 |
"save_steps": 500,
|
| 108 |
-
"total_flos":
|
| 109 |
-
"train_batch_size":
|
| 110 |
"trial_name": null,
|
| 111 |
"trial_params": null
|
| 112 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9963898916967509,
|
| 5 |
"eval_steps": 28,
|
| 6 |
+
"global_step": 138,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.06,
|
| 13 |
+
"grad_norm": 4.96875,
|
| 14 |
+
"learning_rate": 1.911764705882353e-05,
|
| 15 |
+
"loss": 0.7457,
|
| 16 |
"step": 8
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 0.12,
|
| 20 |
+
"grad_norm": 12.75,
|
| 21 |
+
"learning_rate": 1.7941176470588237e-05,
|
| 22 |
+
"loss": 0.7997,
|
| 23 |
"step": 16
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.17,
|
| 27 |
+
"grad_norm": 7.5625,
|
| 28 |
+
"learning_rate": 1.6764705882352943e-05,
|
| 29 |
+
"loss": 0.7616,
|
| 30 |
"step": 24
|
| 31 |
},
|
| 32 |
{
|
| 33 |
+
"epoch": 0.2,
|
| 34 |
+
"eval_accuracy": 0.7666666666666667,
|
| 35 |
+
"eval_f1_score": 0.7609087348217781,
|
| 36 |
+
"eval_gmean": 0.7486572528549951,
|
| 37 |
+
"eval_loss": 0.5174153447151184,
|
| 38 |
+
"eval_precision": 0.7716666666666666,
|
| 39 |
+
"eval_recall": 0.7666666666666667,
|
| 40 |
+
"eval_runtime": 170.072,
|
| 41 |
+
"eval_samples_per_second": 0.353,
|
| 42 |
+
"eval_steps_per_second": 0.047,
|
| 43 |
"step": 28
|
| 44 |
},
|
| 45 |
{
|
| 46 |
+
"epoch": 0.23,
|
| 47 |
+
"grad_norm": 18.0,
|
| 48 |
+
"learning_rate": 1.558823529411765e-05,
|
| 49 |
+
"loss": 0.7458,
|
| 50 |
"step": 32
|
| 51 |
},
|
| 52 |
{
|
| 53 |
+
"epoch": 0.29,
|
| 54 |
+
"grad_norm": 7.78125,
|
| 55 |
+
"learning_rate": 1.4411764705882353e-05,
|
| 56 |
+
"loss": 0.691,
|
| 57 |
"step": 40
|
| 58 |
},
|
| 59 |
{
|
| 60 |
+
"epoch": 0.35,
|
| 61 |
+
"grad_norm": 5.78125,
|
| 62 |
+
"learning_rate": 1.323529411764706e-05,
|
| 63 |
+
"loss": 0.7286,
|
| 64 |
"step": 48
|
| 65 |
},
|
| 66 |
{
|
| 67 |
+
"epoch": 0.4,
|
| 68 |
+
"grad_norm": 6.40625,
|
| 69 |
+
"learning_rate": 1.2058823529411765e-05,
|
| 70 |
+
"loss": 0.8156,
|
| 71 |
"step": 56
|
| 72 |
},
|
| 73 |
{
|
| 74 |
+
"epoch": 0.4,
|
| 75 |
+
"eval_accuracy": 0.7666666666666667,
|
| 76 |
+
"eval_f1_score": 0.7609087348217781,
|
| 77 |
+
"eval_gmean": 0.7486572528549951,
|
| 78 |
+
"eval_loss": 0.5176106691360474,
|
| 79 |
+
"eval_precision": 0.7716666666666666,
|
| 80 |
+
"eval_recall": 0.7666666666666667,
|
| 81 |
+
"eval_runtime": 170.1289,
|
| 82 |
+
"eval_samples_per_second": 0.353,
|
| 83 |
+
"eval_steps_per_second": 0.047,
|
| 84 |
"step": 56
|
| 85 |
},
|
| 86 |
{
|
| 87 |
+
"epoch": 0.46,
|
| 88 |
+
"grad_norm": 10.3125,
|
| 89 |
+
"learning_rate": 1.0882352941176471e-05,
|
| 90 |
+
"loss": 0.7198,
|
| 91 |
"step": 64
|
| 92 |
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 0.52,
|
| 95 |
+
"grad_norm": 9.0625,
|
| 96 |
+
"learning_rate": 9.705882352941177e-06,
|
| 97 |
+
"loss": 0.7421,
|
| 98 |
+
"step": 72
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.58,
|
| 102 |
+
"grad_norm": 6.625,
|
| 103 |
+
"learning_rate": 8.529411764705883e-06,
|
| 104 |
+
"loss": 0.7488,
|
| 105 |
+
"step": 80
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.61,
|
| 109 |
+
"eval_accuracy": 0.7666666666666667,
|
| 110 |
+
"eval_f1_score": 0.7609087348217781,
|
| 111 |
+
"eval_gmean": 0.7486572528549951,
|
| 112 |
+
"eval_loss": 0.5164387822151184,
|
| 113 |
+
"eval_precision": 0.7716666666666666,
|
| 114 |
+
"eval_recall": 0.7666666666666667,
|
| 115 |
+
"eval_runtime": 170.831,
|
| 116 |
+
"eval_samples_per_second": 0.351,
|
| 117 |
+
"eval_steps_per_second": 0.047,
|
| 118 |
+
"step": 84
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 0.64,
|
| 122 |
+
"grad_norm": 6.125,
|
| 123 |
+
"learning_rate": 7.352941176470589e-06,
|
| 124 |
+
"loss": 0.7045,
|
| 125 |
+
"step": 88
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.69,
|
| 129 |
+
"grad_norm": 4.625,
|
| 130 |
+
"learning_rate": 6.176470588235295e-06,
|
| 131 |
+
"loss": 0.6564,
|
| 132 |
+
"step": 96
|
| 133 |
+
},
|
| 134 |
+
{
|
| 135 |
+
"epoch": 0.75,
|
| 136 |
+
"grad_norm": 8.3125,
|
| 137 |
+
"learning_rate": 5e-06,
|
| 138 |
+
"loss": 0.7003,
|
| 139 |
+
"step": 104
|
| 140 |
+
},
|
| 141 |
+
{
|
| 142 |
+
"epoch": 0.81,
|
| 143 |
+
"grad_norm": 12.9375,
|
| 144 |
+
"learning_rate": 3.8235294117647055e-06,
|
| 145 |
+
"loss": 0.7528,
|
| 146 |
+
"step": 112
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.81,
|
| 150 |
+
"eval_accuracy": 0.7666666666666667,
|
| 151 |
+
"eval_f1_score": 0.7609087348217781,
|
| 152 |
+
"eval_gmean": 0.7486572528549951,
|
| 153 |
+
"eval_loss": 0.5166015625,
|
| 154 |
+
"eval_precision": 0.7716666666666666,
|
| 155 |
+
"eval_recall": 0.7666666666666667,
|
| 156 |
+
"eval_runtime": 170.3596,
|
| 157 |
+
"eval_samples_per_second": 0.352,
|
| 158 |
+
"eval_steps_per_second": 0.047,
|
| 159 |
+
"step": 112
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.87,
|
| 163 |
+
"grad_norm": 7.46875,
|
| 164 |
+
"learning_rate": 2.647058823529412e-06,
|
| 165 |
+
"loss": 0.7045,
|
| 166 |
+
"step": 120
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 0.92,
|
| 170 |
+
"grad_norm": 5.8125,
|
| 171 |
+
"learning_rate": 1.4705882352941177e-06,
|
| 172 |
+
"loss": 0.7029,
|
| 173 |
+
"step": 128
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.98,
|
| 177 |
+
"grad_norm": 10.1875,
|
| 178 |
+
"learning_rate": 2.9411764705882356e-07,
|
| 179 |
+
"loss": 0.7869,
|
| 180 |
+
"step": 136
|
| 181 |
+
},
|
| 182 |
{
|
| 183 |
"epoch": 1.0,
|
| 184 |
+
"step": 138,
|
| 185 |
+
"total_flos": 2.597706419798016e+16,
|
| 186 |
+
"train_loss": 0.7354364809782609,
|
| 187 |
+
"train_runtime": 6240.8638,
|
| 188 |
+
"train_samples_per_second": 2.838,
|
| 189 |
+
"train_steps_per_second": 0.022
|
| 190 |
}
|
| 191 |
],
|
| 192 |
"logging_steps": 8,
|
| 193 |
+
"max_steps": 138,
|
| 194 |
"num_input_tokens_seen": 0,
|
| 195 |
"num_train_epochs": 1,
|
| 196 |
"save_steps": 500,
|
| 197 |
+
"total_flos": 2.597706419798016e+16,
|
| 198 |
+
"train_batch_size": 32,
|
| 199 |
"trial_name": null,
|
| 200 |
"trial_params": null
|
| 201 |
}
|