Commit
·
2948d16
1
Parent(s):
b23cabc
Upload 8 files
Browse files- config.json +28 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- trainer_state.json +508 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/home/ambaranov/roberta-base/",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.1,
|
| 12 |
+
"hidden_size": 768,
|
| 13 |
+
"initializer_range": 0.02,
|
| 14 |
+
"intermediate_size": 3072,
|
| 15 |
+
"layer_norm_eps": 1e-05,
|
| 16 |
+
"max_position_embeddings": 514,
|
| 17 |
+
"model_type": "roberta",
|
| 18 |
+
"num_attention_heads": 12,
|
| 19 |
+
"num_hidden_layers": 12,
|
| 20 |
+
"pad_token_id": 1,
|
| 21 |
+
"position_embedding_type": "absolute",
|
| 22 |
+
"problem_type": "single_label_classification",
|
| 23 |
+
"torch_dtype": "float32",
|
| 24 |
+
"transformers_version": "4.27.4",
|
| 25 |
+
"type_vocab_size": 1,
|
| 26 |
+
"use_cache": true,
|
| 27 |
+
"vocab_size": 50265
|
| 28 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:125270372ae2d027557b459f60047f0a5aa1ad731c4161b52acb7d7a37f94010
|
| 3 |
+
size 997295237
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:daf68dd33d84ddc19048d4fcf35849ce990c4fedd043a33bd6b84812a5ad0ec9
|
| 3 |
+
size 498662069
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e0e0ed37db12f49dfa6eb3654219323511dd182d793c977e201d1d0fd1488c15
|
| 3 |
+
size 14639
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2f5c4ecd9e31b63121fc59deda946f92d5bea9e30c7113d58fd83b6e85a08197
|
| 3 |
+
size 557
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e07fb4750b77dd0d14d41ce5abea1031456e3cf252609ba64c97e1981069d362
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,508 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.10954569280147552,
|
| 3 |
+
"best_model_checkpoint": "./models/results_comb_693/checkpoint-400",
|
| 4 |
+
"epoch": 0.887040887040887,
|
| 5 |
+
"global_step": 400,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 4.000000000000001e-06,
|
| 13 |
+
"loss": 0.6881,
|
| 14 |
+
"step": 8
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.04,
|
| 18 |
+
"learning_rate": 7.5e-06,
|
| 19 |
+
"loss": 0.6788,
|
| 20 |
+
"step": 16
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.05,
|
| 24 |
+
"learning_rate": 1.1500000000000002e-05,
|
| 25 |
+
"loss": 0.6504,
|
| 26 |
+
"step": 24
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.06,
|
| 30 |
+
"eval_accuracy": 0.5841623557500994,
|
| 31 |
+
"eval_f1": 0.7375031399145943,
|
| 32 |
+
"eval_loss": 0.6672813892364502,
|
| 33 |
+
"eval_precision": 0.5841623557500994,
|
| 34 |
+
"eval_recall": 1.0,
|
| 35 |
+
"eval_runtime": 35.6463,
|
| 36 |
+
"eval_samples_per_second": 211.494,
|
| 37 |
+
"eval_steps_per_second": 26.454,
|
| 38 |
+
"step": 25
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.07,
|
| 42 |
+
"learning_rate": 1.55e-05,
|
| 43 |
+
"loss": 0.6723,
|
| 44 |
+
"step": 32
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.09,
|
| 48 |
+
"learning_rate": 1.9500000000000003e-05,
|
| 49 |
+
"loss": 0.598,
|
| 50 |
+
"step": 40
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.11,
|
| 54 |
+
"learning_rate": 2.35e-05,
|
| 55 |
+
"loss": 0.4717,
|
| 56 |
+
"step": 48
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.11,
|
| 60 |
+
"eval_accuracy": 0.8839368616527391,
|
| 61 |
+
"eval_f1": 0.9069841607313702,
|
| 62 |
+
"eval_loss": 0.3077848255634308,
|
| 63 |
+
"eval_precision": 0.8526883869678193,
|
| 64 |
+
"eval_recall": 0.9686648501362398,
|
| 65 |
+
"eval_runtime": 34.1268,
|
| 66 |
+
"eval_samples_per_second": 220.911,
|
| 67 |
+
"eval_steps_per_second": 27.632,
|
| 68 |
+
"step": 50
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.12,
|
| 72 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 73 |
+
"loss": 0.3128,
|
| 74 |
+
"step": 56
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.14,
|
| 78 |
+
"learning_rate": 3.15e-05,
|
| 79 |
+
"loss": 0.332,
|
| 80 |
+
"step": 64
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.16,
|
| 84 |
+
"learning_rate": 3.55e-05,
|
| 85 |
+
"loss": 0.2701,
|
| 86 |
+
"step": 72
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.17,
|
| 90 |
+
"eval_accuracy": 0.89322191272052,
|
| 91 |
+
"eval_f1": 0.9027425395674761,
|
| 92 |
+
"eval_loss": 0.25026363134384155,
|
| 93 |
+
"eval_precision": 0.9646269042086238,
|
| 94 |
+
"eval_recall": 0.8483197093551317,
|
| 95 |
+
"eval_runtime": 33.4359,
|
| 96 |
+
"eval_samples_per_second": 225.476,
|
| 97 |
+
"eval_steps_per_second": 28.203,
|
| 98 |
+
"step": 75
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.18,
|
| 102 |
+
"learning_rate": 3.9500000000000005e-05,
|
| 103 |
+
"loss": 0.284,
|
| 104 |
+
"step": 80
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.2,
|
| 108 |
+
"learning_rate": 4.35e-05,
|
| 109 |
+
"loss": 0.283,
|
| 110 |
+
"step": 88
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.21,
|
| 114 |
+
"learning_rate": 4.75e-05,
|
| 115 |
+
"loss": 0.1908,
|
| 116 |
+
"step": 96
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.22,
|
| 120 |
+
"eval_accuracy": 0.9243931555909272,
|
| 121 |
+
"eval_f1": 0.9354327140915271,
|
| 122 |
+
"eval_loss": 0.2004704475402832,
|
| 123 |
+
"eval_precision": 0.9333182640144665,
|
| 124 |
+
"eval_recall": 0.9375567665758402,
|
| 125 |
+
"eval_runtime": 33.8959,
|
| 126 |
+
"eval_samples_per_second": 222.416,
|
| 127 |
+
"eval_steps_per_second": 27.82,
|
| 128 |
+
"step": 100
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.23,
|
| 132 |
+
"learning_rate": 4.957142857142857e-05,
|
| 133 |
+
"loss": 0.237,
|
| 134 |
+
"step": 104
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.25,
|
| 138 |
+
"learning_rate": 4.842857142857143e-05,
|
| 139 |
+
"loss": 0.2298,
|
| 140 |
+
"step": 112
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.27,
|
| 144 |
+
"learning_rate": 4.728571428571429e-05,
|
| 145 |
+
"loss": 0.2824,
|
| 146 |
+
"step": 120
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.28,
|
| 150 |
+
"eval_accuracy": 0.8917628332670116,
|
| 151 |
+
"eval_f1": 0.9143397018685703,
|
| 152 |
+
"eval_loss": 0.301138699054718,
|
| 153 |
+
"eval_precision": 0.850253807106599,
|
| 154 |
+
"eval_recall": 0.9888737511353315,
|
| 155 |
+
"eval_runtime": 33.5038,
|
| 156 |
+
"eval_samples_per_second": 225.019,
|
| 157 |
+
"eval_steps_per_second": 28.146,
|
| 158 |
+
"step": 125
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.28,
|
| 162 |
+
"learning_rate": 4.6142857142857145e-05,
|
| 163 |
+
"loss": 0.2522,
|
| 164 |
+
"step": 128
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.3,
|
| 168 |
+
"learning_rate": 4.5e-05,
|
| 169 |
+
"loss": 0.1874,
|
| 170 |
+
"step": 136
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.32,
|
| 174 |
+
"learning_rate": 4.385714285714286e-05,
|
| 175 |
+
"loss": 0.1865,
|
| 176 |
+
"step": 144
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.33,
|
| 180 |
+
"eval_accuracy": 0.9358005040456294,
|
| 181 |
+
"eval_f1": 0.9457033879291002,
|
| 182 |
+
"eval_loss": 0.16087646782398224,
|
| 183 |
+
"eval_precision": 0.934589800443459,
|
| 184 |
+
"eval_recall": 0.9570844686648501,
|
| 185 |
+
"eval_runtime": 33.8925,
|
| 186 |
+
"eval_samples_per_second": 222.439,
|
| 187 |
+
"eval_steps_per_second": 27.823,
|
| 188 |
+
"step": 150
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.34,
|
| 192 |
+
"learning_rate": 4.271428571428572e-05,
|
| 193 |
+
"loss": 0.2079,
|
| 194 |
+
"step": 152
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.35,
|
| 198 |
+
"learning_rate": 4.1571428571428575e-05,
|
| 199 |
+
"loss": 0.2041,
|
| 200 |
+
"step": 160
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.37,
|
| 204 |
+
"learning_rate": 4.042857142857143e-05,
|
| 205 |
+
"loss": 0.1723,
|
| 206 |
+
"step": 168
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.39,
|
| 210 |
+
"eval_accuracy": 0.9396471680594243,
|
| 211 |
+
"eval_f1": 0.9485468732330657,
|
| 212 |
+
"eval_loss": 0.16514933109283447,
|
| 213 |
+
"eval_precision": 0.9448073890515882,
|
| 214 |
+
"eval_recall": 0.952316076294278,
|
| 215 |
+
"eval_runtime": 32.9214,
|
| 216 |
+
"eval_samples_per_second": 229.0,
|
| 217 |
+
"eval_steps_per_second": 28.644,
|
| 218 |
+
"step": 175
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.39,
|
| 222 |
+
"learning_rate": 3.928571428571429e-05,
|
| 223 |
+
"loss": 0.1804,
|
| 224 |
+
"step": 176
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.41,
|
| 228 |
+
"learning_rate": 3.814285714285715e-05,
|
| 229 |
+
"loss": 0.1687,
|
| 230 |
+
"step": 184
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.43,
|
| 234 |
+
"learning_rate": 3.7e-05,
|
| 235 |
+
"loss": 0.2033,
|
| 236 |
+
"step": 192
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.44,
|
| 240 |
+
"learning_rate": 3.585714285714286e-05,
|
| 241 |
+
"loss": 0.1858,
|
| 242 |
+
"step": 200
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.44,
|
| 246 |
+
"eval_accuracy": 0.9452181987000928,
|
| 247 |
+
"eval_f1": 0.9521492295214923,
|
| 248 |
+
"eval_loss": 0.15048551559448242,
|
| 249 |
+
"eval_precision": 0.9720842204873432,
|
| 250 |
+
"eval_recall": 0.9330154405086285,
|
| 251 |
+
"eval_runtime": 32.8586,
|
| 252 |
+
"eval_samples_per_second": 229.438,
|
| 253 |
+
"eval_steps_per_second": 28.699,
|
| 254 |
+
"step": 200
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.46,
|
| 258 |
+
"learning_rate": 3.471428571428571e-05,
|
| 259 |
+
"loss": 0.1741,
|
| 260 |
+
"step": 208
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.48,
|
| 264 |
+
"learning_rate": 3.357142857142857e-05,
|
| 265 |
+
"loss": 0.1503,
|
| 266 |
+
"step": 216
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.5,
|
| 270 |
+
"learning_rate": 3.242857142857143e-05,
|
| 271 |
+
"loss": 0.1293,
|
| 272 |
+
"step": 224
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.5,
|
| 276 |
+
"eval_accuracy": 0.9112614405093514,
|
| 277 |
+
"eval_f1": 0.9184643510054844,
|
| 278 |
+
"eval_loss": 0.25113266706466675,
|
| 279 |
+
"eval_precision": 0.9913180741910024,
|
| 280 |
+
"eval_recall": 0.8555858310626703,
|
| 281 |
+
"eval_runtime": 33.3517,
|
| 282 |
+
"eval_samples_per_second": 226.045,
|
| 283 |
+
"eval_steps_per_second": 28.274,
|
| 284 |
+
"step": 225
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.51,
|
| 288 |
+
"learning_rate": 3.1285714285714285e-05,
|
| 289 |
+
"loss": 0.2328,
|
| 290 |
+
"step": 232
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.53,
|
| 294 |
+
"learning_rate": 3.0142857142857146e-05,
|
| 295 |
+
"loss": 0.1614,
|
| 296 |
+
"step": 240
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.55,
|
| 300 |
+
"learning_rate": 2.9e-05,
|
| 301 |
+
"loss": 0.1339,
|
| 302 |
+
"step": 248
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.55,
|
| 306 |
+
"eval_accuracy": 0.9495954370606181,
|
| 307 |
+
"eval_f1": 0.9561909153792945,
|
| 308 |
+
"eval_loss": 0.12891609966754913,
|
| 309 |
+
"eval_precision": 0.9711943793911006,
|
| 310 |
+
"eval_recall": 0.9416439600363307,
|
| 311 |
+
"eval_runtime": 32.9115,
|
| 312 |
+
"eval_samples_per_second": 229.069,
|
| 313 |
+
"eval_steps_per_second": 28.653,
|
| 314 |
+
"step": 250
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.57,
|
| 318 |
+
"learning_rate": 2.785714285714286e-05,
|
| 319 |
+
"loss": 0.1318,
|
| 320 |
+
"step": 256
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.59,
|
| 324 |
+
"learning_rate": 2.6714285714285715e-05,
|
| 325 |
+
"loss": 0.1038,
|
| 326 |
+
"step": 264
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.6,
|
| 330 |
+
"learning_rate": 2.5571428571428572e-05,
|
| 331 |
+
"loss": 0.1714,
|
| 332 |
+
"step": 272
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.61,
|
| 336 |
+
"eval_accuracy": 0.9599416368218596,
|
| 337 |
+
"eval_f1": 0.9657052009993187,
|
| 338 |
+
"eval_loss": 0.12271451950073242,
|
| 339 |
+
"eval_precision": 0.9659245797364834,
|
| 340 |
+
"eval_recall": 0.9654859218891917,
|
| 341 |
+
"eval_runtime": 33.9975,
|
| 342 |
+
"eval_samples_per_second": 221.752,
|
| 343 |
+
"eval_steps_per_second": 27.737,
|
| 344 |
+
"step": 275
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.62,
|
| 348 |
+
"learning_rate": 2.442857142857143e-05,
|
| 349 |
+
"loss": 0.1444,
|
| 350 |
+
"step": 280
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.64,
|
| 354 |
+
"learning_rate": 2.3285714285714287e-05,
|
| 355 |
+
"loss": 0.1771,
|
| 356 |
+
"step": 288
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.66,
|
| 360 |
+
"learning_rate": 2.214285714285714e-05,
|
| 361 |
+
"loss": 0.1438,
|
| 362 |
+
"step": 296
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.67,
|
| 366 |
+
"eval_accuracy": 0.954105319007826,
|
| 367 |
+
"eval_f1": 0.9600369600369599,
|
| 368 |
+
"eval_loss": 0.12011300027370453,
|
| 369 |
+
"eval_precision": 0.9769628584861307,
|
| 370 |
+
"eval_recall": 0.9436875567665758,
|
| 371 |
+
"eval_runtime": 34.0522,
|
| 372 |
+
"eval_samples_per_second": 221.395,
|
| 373 |
+
"eval_steps_per_second": 27.693,
|
| 374 |
+
"step": 300
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.67,
|
| 378 |
+
"learning_rate": 2.1e-05,
|
| 379 |
+
"loss": 0.1213,
|
| 380 |
+
"step": 304
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.69,
|
| 384 |
+
"learning_rate": 1.9857142857142856e-05,
|
| 385 |
+
"loss": 0.1517,
|
| 386 |
+
"step": 312
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.71,
|
| 390 |
+
"learning_rate": 1.8714285714285714e-05,
|
| 391 |
+
"loss": 0.132,
|
| 392 |
+
"step": 320
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.72,
|
| 396 |
+
"eval_accuracy": 0.9594110624751293,
|
| 397 |
+
"eval_f1": 0.9649885583524028,
|
| 398 |
+
"eval_loss": 0.11286678165197372,
|
| 399 |
+
"eval_precision": 0.9725553505535055,
|
| 400 |
+
"eval_recall": 0.9575386012715713,
|
| 401 |
+
"eval_runtime": 33.5869,
|
| 402 |
+
"eval_samples_per_second": 224.463,
|
| 403 |
+
"eval_steps_per_second": 28.076,
|
| 404 |
+
"step": 325
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.73,
|
| 408 |
+
"learning_rate": 1.757142857142857e-05,
|
| 409 |
+
"loss": 0.1399,
|
| 410 |
+
"step": 328
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.75,
|
| 414 |
+
"learning_rate": 1.642857142857143e-05,
|
| 415 |
+
"loss": 0.1393,
|
| 416 |
+
"step": 336
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.76,
|
| 420 |
+
"learning_rate": 1.5285714285714286e-05,
|
| 421 |
+
"loss": 0.123,
|
| 422 |
+
"step": 344
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.78,
|
| 426 |
+
"eval_accuracy": 0.9590131317150816,
|
| 427 |
+
"eval_f1": 0.964802369290352,
|
| 428 |
+
"eval_loss": 0.11320723593235016,
|
| 429 |
+
"eval_precision": 0.968,
|
| 430 |
+
"eval_recall": 0.9616257947320618,
|
| 431 |
+
"eval_runtime": 37.7864,
|
| 432 |
+
"eval_samples_per_second": 199.516,
|
| 433 |
+
"eval_steps_per_second": 24.956,
|
| 434 |
+
"step": 350
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.78,
|
| 438 |
+
"learning_rate": 1.4142857142857143e-05,
|
| 439 |
+
"loss": 0.1067,
|
| 440 |
+
"step": 352
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.8,
|
| 444 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 445 |
+
"loss": 0.1252,
|
| 446 |
+
"step": 360
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.82,
|
| 450 |
+
"learning_rate": 1.1857142857142858e-05,
|
| 451 |
+
"loss": 0.1513,
|
| 452 |
+
"step": 368
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.83,
|
| 456 |
+
"eval_accuracy": 0.9595437060618119,
|
| 457 |
+
"eval_f1": 0.9653684569092766,
|
| 458 |
+
"eval_loss": 0.11097515374422073,
|
| 459 |
+
"eval_precision": 0.9654780831251419,
|
| 460 |
+
"eval_recall": 0.965258855585831,
|
| 461 |
+
"eval_runtime": 37.8195,
|
| 462 |
+
"eval_samples_per_second": 199.342,
|
| 463 |
+
"eval_steps_per_second": 24.934,
|
| 464 |
+
"step": 375
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.83,
|
| 468 |
+
"learning_rate": 1.0714285714285714e-05,
|
| 469 |
+
"loss": 0.1322,
|
| 470 |
+
"step": 376
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.85,
|
| 474 |
+
"learning_rate": 9.571428571428572e-06,
|
| 475 |
+
"loss": 0.1199,
|
| 476 |
+
"step": 384
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.87,
|
| 480 |
+
"learning_rate": 8.428571428571429e-06,
|
| 481 |
+
"loss": 0.1171,
|
| 482 |
+
"step": 392
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.89,
|
| 486 |
+
"learning_rate": 7.285714285714286e-06,
|
| 487 |
+
"loss": 0.1583,
|
| 488 |
+
"step": 400
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.89,
|
| 492 |
+
"eval_accuracy": 0.9600742804085423,
|
| 493 |
+
"eval_f1": 0.965643191416505,
|
| 494 |
+
"eval_loss": 0.10954569280147552,
|
| 495 |
+
"eval_precision": 0.970851503327978,
|
| 496 |
+
"eval_recall": 0.9604904632152589,
|
| 497 |
+
"eval_runtime": 37.3506,
|
| 498 |
+
"eval_samples_per_second": 201.844,
|
| 499 |
+
"eval_steps_per_second": 25.247,
|
| 500 |
+
"step": 400
|
| 501 |
+
}
|
| 502 |
+
],
|
| 503 |
+
"max_steps": 450,
|
| 504 |
+
"num_train_epochs": 1,
|
| 505 |
+
"total_flos": 4354487966208000.0,
|
| 506 |
+
"trial_name": null,
|
| 507 |
+
"trial_params": null
|
| 508 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:029d10122f479d9abcfafee6fc4e9cfc495d0ceeef61ca9231b326ec13bdc7f4
|
| 3 |
+
size 3579
|