Commit
·
ca13b08
1
Parent(s):
4a06feb
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 +472 -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 sha256:0d4aa068bed43db418971110c8b1d771eff074a71f4ae5e947e06fba86b06a32
|
| 3 |
+
size 997295237
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:03a227d22cb747063aed5197909b7754baa6b3e2db664c15aa713f4624d03b8d
|
| 3 |
+
size 498662069
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9c7d3576a2a0c6c53ae82d76a1d9eaacbdb9f7c1fa8c011f958033c6947df1b2
|
| 3 |
+
size 14575
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5ba21dd811eba354dd2811b10590ff002ddacc6908dc696938c3a0bf95fddb11
|
| 3 |
+
size 557
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:27ab3c6f1c9890eda95497d1f529fd666f9d6241b3af8e9edb3785b10c19e06d
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,472 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.10937219858169556,
|
| 3 |
+
"best_model_checkpoint": "./models/results_comb_977/checkpoint-375",
|
| 4 |
+
"epoch": 0.8316008316008316,
|
| 5 |
+
"global_step": 375,
|
| 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": 0.6852,
|
| 14 |
+
"step": 8
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.04,
|
| 18 |
+
"learning_rate": 8.000000000000001e-06,
|
| 19 |
+
"loss": 0.6763,
|
| 20 |
+
"step": 16
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.05,
|
| 24 |
+
"learning_rate": 1.2e-05,
|
| 25 |
+
"loss": 0.6574,
|
| 26 |
+
"step": 24
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.06,
|
| 30 |
+
"eval_accuracy": 0.5841623557500994,
|
| 31 |
+
"eval_f1": 0.7375031399145943,
|
| 32 |
+
"eval_loss": 0.6670724749565125,
|
| 33 |
+
"eval_precision": 0.5841623557500994,
|
| 34 |
+
"eval_recall": 1.0,
|
| 35 |
+
"eval_runtime": 49.517,
|
| 36 |
+
"eval_samples_per_second": 152.251,
|
| 37 |
+
"eval_steps_per_second": 19.044,
|
| 38 |
+
"step": 25
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.07,
|
| 42 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 43 |
+
"loss": 0.6057,
|
| 44 |
+
"step": 32
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.09,
|
| 48 |
+
"learning_rate": 2e-05,
|
| 49 |
+
"loss": 0.6139,
|
| 50 |
+
"step": 40
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.11,
|
| 54 |
+
"learning_rate": 2.4e-05,
|
| 55 |
+
"loss": 0.5282,
|
| 56 |
+
"step": 48
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.11,
|
| 60 |
+
"eval_accuracy": 0.7591192465844276,
|
| 61 |
+
"eval_f1": 0.8288084464555052,
|
| 62 |
+
"eval_loss": 0.43177902698516846,
|
| 63 |
+
"eval_precision": 0.708575112830432,
|
| 64 |
+
"eval_recall": 0.9981834695731153,
|
| 65 |
+
"eval_runtime": 49.8289,
|
| 66 |
+
"eval_samples_per_second": 151.298,
|
| 67 |
+
"eval_steps_per_second": 18.925,
|
| 68 |
+
"step": 50
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.12,
|
| 72 |
+
"learning_rate": 2.8000000000000003e-05,
|
| 73 |
+
"loss": 0.421,
|
| 74 |
+
"step": 56
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.14,
|
| 78 |
+
"learning_rate": 3.2000000000000005e-05,
|
| 79 |
+
"loss": 0.3004,
|
| 80 |
+
"step": 64
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.16,
|
| 84 |
+
"learning_rate": 3.6e-05,
|
| 85 |
+
"loss": 0.3156,
|
| 86 |
+
"step": 72
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.17,
|
| 90 |
+
"eval_accuracy": 0.9250563735243401,
|
| 91 |
+
"eval_f1": 0.9366094468753505,
|
| 92 |
+
"eval_loss": 0.19340644776821136,
|
| 93 |
+
"eval_precision": 0.9257041472610334,
|
| 94 |
+
"eval_recall": 0.9477747502270663,
|
| 95 |
+
"eval_runtime": 48.8142,
|
| 96 |
+
"eval_samples_per_second": 154.443,
|
| 97 |
+
"eval_steps_per_second": 19.318,
|
| 98 |
+
"step": 75
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.18,
|
| 102 |
+
"learning_rate": 4e-05,
|
| 103 |
+
"loss": 0.275,
|
| 104 |
+
"step": 80
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.2,
|
| 108 |
+
"learning_rate": 4.4000000000000006e-05,
|
| 109 |
+
"loss": 0.2258,
|
| 110 |
+
"step": 88
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.21,
|
| 114 |
+
"learning_rate": 4.75e-05,
|
| 115 |
+
"loss": 0.2682,
|
| 116 |
+
"step": 96
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.22,
|
| 120 |
+
"eval_accuracy": 0.931157978511739,
|
| 121 |
+
"eval_f1": 0.9422370617696161,
|
| 122 |
+
"eval_loss": 0.19519653916358948,
|
| 123 |
+
"eval_precision": 0.9240340537000655,
|
| 124 |
+
"eval_recall": 0.9611716621253406,
|
| 125 |
+
"eval_runtime": 48.2615,
|
| 126 |
+
"eval_samples_per_second": 156.212,
|
| 127 |
+
"eval_steps_per_second": 19.539,
|
| 128 |
+
"step": 100
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.23,
|
| 132 |
+
"learning_rate": 4.957142857142857e-05,
|
| 133 |
+
"loss": 0.2448,
|
| 134 |
+
"step": 104
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.25,
|
| 138 |
+
"learning_rate": 4.842857142857143e-05,
|
| 139 |
+
"loss": 0.2179,
|
| 140 |
+
"step": 112
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.27,
|
| 144 |
+
"learning_rate": 4.728571428571429e-05,
|
| 145 |
+
"loss": 0.2161,
|
| 146 |
+
"step": 120
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.28,
|
| 150 |
+
"eval_accuracy": 0.9363310783923597,
|
| 151 |
+
"eval_f1": 0.9444444444444444,
|
| 152 |
+
"eval_loss": 0.1711500585079193,
|
| 153 |
+
"eval_precision": 0.9631728045325779,
|
| 154 |
+
"eval_recall": 0.9264305177111717,
|
| 155 |
+
"eval_runtime": 49.0831,
|
| 156 |
+
"eval_samples_per_second": 153.597,
|
| 157 |
+
"eval_steps_per_second": 19.212,
|
| 158 |
+
"step": 125
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.28,
|
| 162 |
+
"learning_rate": 4.6142857142857145e-05,
|
| 163 |
+
"loss": 0.185,
|
| 164 |
+
"step": 128
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.3,
|
| 168 |
+
"learning_rate": 4.5e-05,
|
| 169 |
+
"loss": 0.213,
|
| 170 |
+
"step": 136
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.32,
|
| 174 |
+
"learning_rate": 4.385714285714286e-05,
|
| 175 |
+
"loss": 0.2209,
|
| 176 |
+
"step": 144
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.33,
|
| 180 |
+
"eval_accuracy": 0.9320864836185171,
|
| 181 |
+
"eval_f1": 0.9418842224744609,
|
| 182 |
+
"eval_loss": 0.18172995746135712,
|
| 183 |
+
"eval_precision": 0.9416704493871992,
|
| 184 |
+
"eval_recall": 0.9420980926430518,
|
| 185 |
+
"eval_runtime": 48.8331,
|
| 186 |
+
"eval_samples_per_second": 154.383,
|
| 187 |
+
"eval_steps_per_second": 19.311,
|
| 188 |
+
"step": 150
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.34,
|
| 192 |
+
"learning_rate": 4.271428571428572e-05,
|
| 193 |
+
"loss": 0.1561,
|
| 194 |
+
"step": 152
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.35,
|
| 198 |
+
"learning_rate": 4.1571428571428575e-05,
|
| 199 |
+
"loss": 0.2087,
|
| 200 |
+
"step": 160
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.37,
|
| 204 |
+
"learning_rate": 4.042857142857143e-05,
|
| 205 |
+
"loss": 0.2013,
|
| 206 |
+
"step": 168
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.39,
|
| 210 |
+
"eval_accuracy": 0.9460140602201883,
|
| 211 |
+
"eval_f1": 0.9541511771995044,
|
| 212 |
+
"eval_loss": 0.14321394264698029,
|
| 213 |
+
"eval_precision": 0.94679186228482,
|
| 214 |
+
"eval_recall": 0.9616257947320618,
|
| 215 |
+
"eval_runtime": 48.5882,
|
| 216 |
+
"eval_samples_per_second": 155.161,
|
| 217 |
+
"eval_steps_per_second": 19.408,
|
| 218 |
+
"step": 175
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.39,
|
| 222 |
+
"learning_rate": 3.928571428571429e-05,
|
| 223 |
+
"loss": 0.1913,
|
| 224 |
+
"step": 176
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.41,
|
| 228 |
+
"learning_rate": 3.814285714285715e-05,
|
| 229 |
+
"loss": 0.1259,
|
| 230 |
+
"step": 184
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.43,
|
| 234 |
+
"learning_rate": 3.7e-05,
|
| 235 |
+
"loss": 0.1834,
|
| 236 |
+
"step": 192
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.44,
|
| 240 |
+
"learning_rate": 3.585714285714286e-05,
|
| 241 |
+
"loss": 0.2194,
|
| 242 |
+
"step": 200
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.44,
|
| 246 |
+
"eval_accuracy": 0.9450855551134103,
|
| 247 |
+
"eval_f1": 0.9540816326530612,
|
| 248 |
+
"eval_loss": 0.1561821550130844,
|
| 249 |
+
"eval_precision": 0.9325672159583694,
|
| 250 |
+
"eval_recall": 0.9766121707538601,
|
| 251 |
+
"eval_runtime": 48.3284,
|
| 252 |
+
"eval_samples_per_second": 155.995,
|
| 253 |
+
"eval_steps_per_second": 19.512,
|
| 254 |
+
"step": 200
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.46,
|
| 258 |
+
"learning_rate": 3.471428571428571e-05,
|
| 259 |
+
"loss": 0.1455,
|
| 260 |
+
"step": 208
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.48,
|
| 264 |
+
"learning_rate": 3.357142857142857e-05,
|
| 265 |
+
"loss": 0.1617,
|
| 266 |
+
"step": 216
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.5,
|
| 270 |
+
"learning_rate": 3.242857142857143e-05,
|
| 271 |
+
"loss": 0.1336,
|
| 272 |
+
"step": 224
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.5,
|
| 276 |
+
"eval_accuracy": 0.9474731396736967,
|
| 277 |
+
"eval_f1": 0.9546495648190564,
|
| 278 |
+
"eval_loss": 0.13737089931964874,
|
| 279 |
+
"eval_precision": 0.9630314232902033,
|
| 280 |
+
"eval_recall": 0.9464123524069028,
|
| 281 |
+
"eval_runtime": 48.3619,
|
| 282 |
+
"eval_samples_per_second": 155.887,
|
| 283 |
+
"eval_steps_per_second": 19.499,
|
| 284 |
+
"step": 225
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.51,
|
| 288 |
+
"learning_rate": 3.1285714285714285e-05,
|
| 289 |
+
"loss": 0.1541,
|
| 290 |
+
"step": 232
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.53,
|
| 294 |
+
"learning_rate": 3.0142857142857146e-05,
|
| 295 |
+
"loss": 0.1724,
|
| 296 |
+
"step": 240
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.55,
|
| 300 |
+
"learning_rate": 2.9e-05,
|
| 301 |
+
"loss": 0.1387,
|
| 302 |
+
"step": 248
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.55,
|
| 306 |
+
"eval_accuracy": 0.9465446345669187,
|
| 307 |
+
"eval_f1": 0.9532102635550912,
|
| 308 |
+
"eval_loss": 0.14909207820892334,
|
| 309 |
+
"eval_precision": 0.9752910430030887,
|
| 310 |
+
"eval_recall": 0.9321071752951862,
|
| 311 |
+
"eval_runtime": 48.4793,
|
| 312 |
+
"eval_samples_per_second": 155.51,
|
| 313 |
+
"eval_steps_per_second": 19.452,
|
| 314 |
+
"step": 250
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.57,
|
| 318 |
+
"learning_rate": 2.785714285714286e-05,
|
| 319 |
+
"loss": 0.1559,
|
| 320 |
+
"step": 256
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.59,
|
| 324 |
+
"learning_rate": 2.6714285714285715e-05,
|
| 325 |
+
"loss": 0.1444,
|
| 326 |
+
"step": 264
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.6,
|
| 330 |
+
"learning_rate": 2.5571428571428572e-05,
|
| 331 |
+
"loss": 0.1196,
|
| 332 |
+
"step": 272
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.61,
|
| 336 |
+
"eval_accuracy": 0.955697042048017,
|
| 337 |
+
"eval_f1": 0.9622342831298055,
|
| 338 |
+
"eval_loss": 0.12214687466621399,
|
| 339 |
+
"eval_precision": 0.9583333333333334,
|
| 340 |
+
"eval_recall": 0.9661671207992734,
|
| 341 |
+
"eval_runtime": 48.8578,
|
| 342 |
+
"eval_samples_per_second": 154.305,
|
| 343 |
+
"eval_steps_per_second": 19.301,
|
| 344 |
+
"step": 275
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.62,
|
| 348 |
+
"learning_rate": 2.442857142857143e-05,
|
| 349 |
+
"loss": 0.1397,
|
| 350 |
+
"step": 280
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.64,
|
| 354 |
+
"learning_rate": 2.3285714285714287e-05,
|
| 355 |
+
"loss": 0.1284,
|
| 356 |
+
"step": 288
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.66,
|
| 360 |
+
"learning_rate": 2.214285714285714e-05,
|
| 361 |
+
"loss": 0.1779,
|
| 362 |
+
"step": 296
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.67,
|
| 366 |
+
"eval_accuracy": 0.9549011805279215,
|
| 367 |
+
"eval_f1": 0.9617289509230076,
|
| 368 |
+
"eval_loss": 0.12309886515140533,
|
| 369 |
+
"eval_precision": 0.9535714285714286,
|
| 370 |
+
"eval_recall": 0.9700272479564033,
|
| 371 |
+
"eval_runtime": 48.6426,
|
| 372 |
+
"eval_samples_per_second": 154.987,
|
| 373 |
+
"eval_steps_per_second": 19.386,
|
| 374 |
+
"step": 300
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.67,
|
| 378 |
+
"learning_rate": 2.1e-05,
|
| 379 |
+
"loss": 0.1118,
|
| 380 |
+
"step": 304
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.69,
|
| 384 |
+
"learning_rate": 1.9857142857142856e-05,
|
| 385 |
+
"loss": 0.1375,
|
| 386 |
+
"step": 312
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.71,
|
| 390 |
+
"learning_rate": 1.8714285714285714e-05,
|
| 391 |
+
"loss": 0.097,
|
| 392 |
+
"step": 320
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.72,
|
| 396 |
+
"eval_accuracy": 0.9547685369412389,
|
| 397 |
+
"eval_f1": 0.9607549775578317,
|
| 398 |
+
"eval_loss": 0.13004465401172638,
|
| 399 |
+
"eval_precision": 0.974095682613769,
|
| 400 |
+
"eval_recall": 0.9477747502270663,
|
| 401 |
+
"eval_runtime": 48.3897,
|
| 402 |
+
"eval_samples_per_second": 155.797,
|
| 403 |
+
"eval_steps_per_second": 19.488,
|
| 404 |
+
"step": 325
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.73,
|
| 408 |
+
"learning_rate": 1.757142857142857e-05,
|
| 409 |
+
"loss": 0.1579,
|
| 410 |
+
"step": 328
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.75,
|
| 414 |
+
"learning_rate": 1.642857142857143e-05,
|
| 415 |
+
"loss": 0.1502,
|
| 416 |
+
"step": 336
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.76,
|
| 420 |
+
"learning_rate": 1.5285714285714286e-05,
|
| 421 |
+
"loss": 0.101,
|
| 422 |
+
"step": 344
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.78,
|
| 426 |
+
"eval_accuracy": 0.9599416368218596,
|
| 427 |
+
"eval_f1": 0.9658834161771351,
|
| 428 |
+
"eval_loss": 0.11177459359169006,
|
| 429 |
+
"eval_precision": 0.9611061151079137,
|
| 430 |
+
"eval_recall": 0.970708446866485,
|
| 431 |
+
"eval_runtime": 48.7108,
|
| 432 |
+
"eval_samples_per_second": 154.771,
|
| 433 |
+
"eval_steps_per_second": 19.359,
|
| 434 |
+
"step": 350
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.78,
|
| 438 |
+
"learning_rate": 1.4142857142857143e-05,
|
| 439 |
+
"loss": 0.1073,
|
| 440 |
+
"step": 352
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.8,
|
| 444 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 445 |
+
"loss": 0.1431,
|
| 446 |
+
"step": 360
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.82,
|
| 450 |
+
"learning_rate": 1.1857142857142858e-05,
|
| 451 |
+
"loss": 0.1708,
|
| 452 |
+
"step": 368
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.83,
|
| 456 |
+
"eval_accuracy": 0.9583499137816687,
|
| 457 |
+
"eval_f1": 0.9646714671467146,
|
| 458 |
+
"eval_loss": 0.10937219858169556,
|
| 459 |
+
"eval_precision": 0.9560660124888493,
|
| 460 |
+
"eval_recall": 0.973433242506812,
|
| 461 |
+
"eval_runtime": 49.3579,
|
| 462 |
+
"eval_samples_per_second": 152.742,
|
| 463 |
+
"eval_steps_per_second": 19.105,
|
| 464 |
+
"step": 375
|
| 465 |
+
}
|
| 466 |
+
],
|
| 467 |
+
"max_steps": 450,
|
| 468 |
+
"num_train_epochs": 1,
|
| 469 |
+
"total_flos": 4082332468320000.0,
|
| 470 |
+
"trial_name": null,
|
| 471 |
+
"trial_params": null
|
| 472 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:64f15fab0ebcad6cc6b9a3d4e52b9e592cb46e7153068c3ae342bfb1a12de17c
|
| 3 |
+
size 3579
|