Commit
·
c846a28
1
Parent(s):
4392c4c
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 +538 -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 sha256:84a957459fe94a7d70c7dc449dfcf9f1dd61321327ef80ad8e55905293e15a45
|
| 3 |
+
size 997295237
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:98f93deb91413d77bc13729b6aceb7957dc3dae1f891ceaf45e252f593ca55f2
|
| 3 |
+
size 498662069
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e110e91699b9baf9644ff0e21c7fe974d2b997af54c092377aa6bbf30e686873
|
| 3 |
+
size 14639
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b183c4200f97b2f91d18164080c6e8299090e1ddd1e0f6056c15ce1d8ff646b6
|
| 3 |
+
size 557
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e236ecf3f70271d763abd39f53175b6907d327d7f168b31d407f7298152a60bf
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,538 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.113589346408844,
|
| 3 |
+
"best_model_checkpoint": "./models/results_comb_453/checkpoint-425",
|
| 4 |
+
"epoch": 0.9424809424809425,
|
| 5 |
+
"global_step": 425,
|
| 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": 0.6901,
|
| 14 |
+
"step": 8
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.04,
|
| 18 |
+
"learning_rate": 8.000000000000001e-06,
|
| 19 |
+
"loss": 0.6684,
|
| 20 |
+
"step": 16
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.05,
|
| 24 |
+
"learning_rate": 1.2e-05,
|
| 25 |
+
"loss": 0.6534,
|
| 26 |
+
"step": 24
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.06,
|
| 30 |
+
"eval_accuracy": 0.5841623557500994,
|
| 31 |
+
"eval_f1": 0.7375031399145943,
|
| 32 |
+
"eval_loss": 0.6666890978813171,
|
| 33 |
+
"eval_precision": 0.5841623557500994,
|
| 34 |
+
"eval_recall": 1.0,
|
| 35 |
+
"eval_runtime": 38.749,
|
| 36 |
+
"eval_samples_per_second": 194.56,
|
| 37 |
+
"eval_steps_per_second": 24.336,
|
| 38 |
+
"step": 25
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.07,
|
| 42 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 43 |
+
"loss": 0.64,
|
| 44 |
+
"step": 32
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.09,
|
| 48 |
+
"learning_rate": 2e-05,
|
| 49 |
+
"loss": 0.6171,
|
| 50 |
+
"step": 40
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.11,
|
| 54 |
+
"learning_rate": 2.35e-05,
|
| 55 |
+
"loss": 0.4811,
|
| 56 |
+
"step": 48
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.11,
|
| 60 |
+
"eval_accuracy": 0.8840695052394216,
|
| 61 |
+
"eval_f1": 0.9018859452177818,
|
| 62 |
+
"eval_loss": 0.38738805055618286,
|
| 63 |
+
"eval_precision": 0.8918738898756661,
|
| 64 |
+
"eval_recall": 0.912125340599455,
|
| 65 |
+
"eval_runtime": 34.9973,
|
| 66 |
+
"eval_samples_per_second": 215.416,
|
| 67 |
+
"eval_steps_per_second": 26.945,
|
| 68 |
+
"step": 50
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.12,
|
| 72 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 73 |
+
"loss": 0.3376,
|
| 74 |
+
"step": 56
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.14,
|
| 78 |
+
"learning_rate": 3.15e-05,
|
| 79 |
+
"loss": 0.3194,
|
| 80 |
+
"step": 64
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.16,
|
| 84 |
+
"learning_rate": 3.55e-05,
|
| 85 |
+
"loss": 0.256,
|
| 86 |
+
"step": 72
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.17,
|
| 90 |
+
"eval_accuracy": 0.913914312243003,
|
| 91 |
+
"eval_f1": 0.9237636555855749,
|
| 92 |
+
"eval_loss": 0.21917007863521576,
|
| 93 |
+
"eval_precision": 0.9569238257483572,
|
| 94 |
+
"eval_recall": 0.8928247048138056,
|
| 95 |
+
"eval_runtime": 35.1296,
|
| 96 |
+
"eval_samples_per_second": 214.606,
|
| 97 |
+
"eval_steps_per_second": 26.843,
|
| 98 |
+
"step": 75
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.18,
|
| 102 |
+
"learning_rate": 3.9500000000000005e-05,
|
| 103 |
+
"loss": 0.243,
|
| 104 |
+
"step": 80
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.2,
|
| 108 |
+
"learning_rate": 4.35e-05,
|
| 109 |
+
"loss": 0.2168,
|
| 110 |
+
"step": 88
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.21,
|
| 114 |
+
"learning_rate": 4.75e-05,
|
| 115 |
+
"loss": 0.2272,
|
| 116 |
+
"step": 96
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.22,
|
| 120 |
+
"eval_accuracy": 0.9312906220984215,
|
| 121 |
+
"eval_f1": 0.9422390722569135,
|
| 122 |
+
"eval_loss": 0.18572893738746643,
|
| 123 |
+
"eval_precision": 0.9257230499561788,
|
| 124 |
+
"eval_recall": 0.959355131698456,
|
| 125 |
+
"eval_runtime": 34.8101,
|
| 126 |
+
"eval_samples_per_second": 216.575,
|
| 127 |
+
"eval_steps_per_second": 27.09,
|
| 128 |
+
"step": 100
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.23,
|
| 132 |
+
"learning_rate": 4.957142857142857e-05,
|
| 133 |
+
"loss": 0.1749,
|
| 134 |
+
"step": 104
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.25,
|
| 138 |
+
"learning_rate": 4.842857142857143e-05,
|
| 139 |
+
"loss": 0.2275,
|
| 140 |
+
"step": 112
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.27,
|
| 144 |
+
"learning_rate": 4.728571428571429e-05,
|
| 145 |
+
"loss": 0.218,
|
| 146 |
+
"step": 120
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.28,
|
| 150 |
+
"eval_accuracy": 0.9393818808860591,
|
| 151 |
+
"eval_f1": 0.949171393615838,
|
| 152 |
+
"eval_loss": 0.15581709146499634,
|
| 153 |
+
"eval_precision": 0.9302376280793547,
|
| 154 |
+
"eval_recall": 0.9688919164396004,
|
| 155 |
+
"eval_runtime": 35.2114,
|
| 156 |
+
"eval_samples_per_second": 214.107,
|
| 157 |
+
"eval_steps_per_second": 26.781,
|
| 158 |
+
"step": 125
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.28,
|
| 162 |
+
"learning_rate": 4.6142857142857145e-05,
|
| 163 |
+
"loss": 0.2285,
|
| 164 |
+
"step": 128
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.3,
|
| 168 |
+
"learning_rate": 4.5e-05,
|
| 169 |
+
"loss": 0.1953,
|
| 170 |
+
"step": 136
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.32,
|
| 174 |
+
"learning_rate": 4.385714285714286e-05,
|
| 175 |
+
"loss": 0.201,
|
| 176 |
+
"step": 144
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.33,
|
| 180 |
+
"eval_accuracy": 0.940708316752885,
|
| 181 |
+
"eval_f1": 0.9498260186328432,
|
| 182 |
+
"eval_loss": 0.154759481549263,
|
| 183 |
+
"eval_precision": 0.9391786903440622,
|
| 184 |
+
"eval_recall": 0.9607175295186194,
|
| 185 |
+
"eval_runtime": 34.6999,
|
| 186 |
+
"eval_samples_per_second": 217.263,
|
| 187 |
+
"eval_steps_per_second": 27.176,
|
| 188 |
+
"step": 150
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.34,
|
| 192 |
+
"learning_rate": 4.271428571428572e-05,
|
| 193 |
+
"loss": 0.1915,
|
| 194 |
+
"step": 152
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.35,
|
| 198 |
+
"learning_rate": 4.1571428571428575e-05,
|
| 199 |
+
"loss": 0.2163,
|
| 200 |
+
"step": 160
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.37,
|
| 204 |
+
"learning_rate": 4.042857142857143e-05,
|
| 205 |
+
"loss": 0.1934,
|
| 206 |
+
"step": 168
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.39,
|
| 210 |
+
"eval_accuracy": 0.9387186629526463,
|
| 211 |
+
"eval_f1": 0.9471032745591939,
|
| 212 |
+
"eval_loss": 0.1729520708322525,
|
| 213 |
+
"eval_precision": 0.9551963048498845,
|
| 214 |
+
"eval_recall": 0.9391462306993642,
|
| 215 |
+
"eval_runtime": 34.0751,
|
| 216 |
+
"eval_samples_per_second": 221.247,
|
| 217 |
+
"eval_steps_per_second": 27.674,
|
| 218 |
+
"step": 175
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.39,
|
| 222 |
+
"learning_rate": 3.928571428571429e-05,
|
| 223 |
+
"loss": 0.185,
|
| 224 |
+
"step": 176
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.41,
|
| 228 |
+
"learning_rate": 3.814285714285715e-05,
|
| 229 |
+
"loss": 0.2181,
|
| 230 |
+
"step": 184
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.43,
|
| 234 |
+
"learning_rate": 3.7e-05,
|
| 235 |
+
"loss": 0.1423,
|
| 236 |
+
"step": 192
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.44,
|
| 240 |
+
"learning_rate": 3.585714285714286e-05,
|
| 241 |
+
"loss": 0.1737,
|
| 242 |
+
"step": 200
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.44,
|
| 246 |
+
"eval_accuracy": 0.9441570500066322,
|
| 247 |
+
"eval_f1": 0.9513013302486986,
|
| 248 |
+
"eval_loss": 0.1504576951265335,
|
| 249 |
+
"eval_precision": 0.9695826456024522,
|
| 250 |
+
"eval_recall": 0.9336966394187103,
|
| 251 |
+
"eval_runtime": 34.4966,
|
| 252 |
+
"eval_samples_per_second": 218.544,
|
| 253 |
+
"eval_steps_per_second": 27.336,
|
| 254 |
+
"step": 200
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.46,
|
| 258 |
+
"learning_rate": 3.471428571428571e-05,
|
| 259 |
+
"loss": 0.1664,
|
| 260 |
+
"step": 208
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.48,
|
| 264 |
+
"learning_rate": 3.357142857142857e-05,
|
| 265 |
+
"loss": 0.1072,
|
| 266 |
+
"step": 216
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.5,
|
| 270 |
+
"learning_rate": 3.242857142857143e-05,
|
| 271 |
+
"loss": 0.1209,
|
| 272 |
+
"step": 224
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.5,
|
| 276 |
+
"eval_accuracy": 0.9538400318344608,
|
| 277 |
+
"eval_f1": 0.9605710401087695,
|
| 278 |
+
"eval_loss": 0.12554903328418732,
|
| 279 |
+
"eval_precision": 0.9586160108548168,
|
| 280 |
+
"eval_recall": 0.9625340599455041,
|
| 281 |
+
"eval_runtime": 34.0808,
|
| 282 |
+
"eval_samples_per_second": 221.21,
|
| 283 |
+
"eval_steps_per_second": 27.67,
|
| 284 |
+
"step": 225
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.51,
|
| 288 |
+
"learning_rate": 3.1285714285714285e-05,
|
| 289 |
+
"loss": 0.1808,
|
| 290 |
+
"step": 232
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.53,
|
| 294 |
+
"learning_rate": 3.0142857142857146e-05,
|
| 295 |
+
"loss": 0.1619,
|
| 296 |
+
"step": 240
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.55,
|
| 300 |
+
"learning_rate": 2.9e-05,
|
| 301 |
+
"loss": 0.1808,
|
| 302 |
+
"step": 248
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.55,
|
| 306 |
+
"eval_accuracy": 0.9460140602201883,
|
| 307 |
+
"eval_f1": 0.9544691799977626,
|
| 308 |
+
"eval_loss": 0.14133918285369873,
|
| 309 |
+
"eval_precision": 0.940683572216097,
|
| 310 |
+
"eval_recall": 0.9686648501362398,
|
| 311 |
+
"eval_runtime": 34.4019,
|
| 312 |
+
"eval_samples_per_second": 219.145,
|
| 313 |
+
"eval_steps_per_second": 27.411,
|
| 314 |
+
"step": 250
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.57,
|
| 318 |
+
"learning_rate": 2.785714285714286e-05,
|
| 319 |
+
"loss": 0.1802,
|
| 320 |
+
"step": 256
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.59,
|
| 324 |
+
"learning_rate": 2.6714285714285715e-05,
|
| 325 |
+
"loss": 0.1555,
|
| 326 |
+
"step": 264
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.6,
|
| 330 |
+
"learning_rate": 2.5571428571428572e-05,
|
| 331 |
+
"loss": 0.1274,
|
| 332 |
+
"step": 272
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.61,
|
| 336 |
+
"eval_accuracy": 0.9559623292213821,
|
| 337 |
+
"eval_f1": 0.9620137299771168,
|
| 338 |
+
"eval_loss": 0.1330484300851822,
|
| 339 |
+
"eval_precision": 0.9695571955719557,
|
| 340 |
+
"eval_recall": 0.9545867393278837,
|
| 341 |
+
"eval_runtime": 34.8961,
|
| 342 |
+
"eval_samples_per_second": 216.041,
|
| 343 |
+
"eval_steps_per_second": 27.023,
|
| 344 |
+
"step": 275
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.62,
|
| 348 |
+
"learning_rate": 2.442857142857143e-05,
|
| 349 |
+
"loss": 0.1112,
|
| 350 |
+
"step": 280
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.64,
|
| 354 |
+
"learning_rate": 2.3285714285714287e-05,
|
| 355 |
+
"loss": 0.1567,
|
| 356 |
+
"step": 288
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.66,
|
| 360 |
+
"learning_rate": 2.214285714285714e-05,
|
| 361 |
+
"loss": 0.1408,
|
| 362 |
+
"step": 296
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.67,
|
| 366 |
+
"eval_accuracy": 0.9571561215015254,
|
| 367 |
+
"eval_f1": 0.9634325823615986,
|
| 368 |
+
"eval_loss": 0.12022460252046585,
|
| 369 |
+
"eval_precision": 0.9607134793407089,
|
| 370 |
+
"eval_recall": 0.9661671207992734,
|
| 371 |
+
"eval_runtime": 35.3597,
|
| 372 |
+
"eval_samples_per_second": 213.209,
|
| 373 |
+
"eval_steps_per_second": 26.669,
|
| 374 |
+
"step": 300
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.67,
|
| 378 |
+
"learning_rate": 2.1e-05,
|
| 379 |
+
"loss": 0.1256,
|
| 380 |
+
"step": 304
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.69,
|
| 384 |
+
"learning_rate": 1.9857142857142856e-05,
|
| 385 |
+
"loss": 0.1378,
|
| 386 |
+
"step": 312
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.71,
|
| 390 |
+
"learning_rate": 1.8714285714285714e-05,
|
| 391 |
+
"loss": 0.0942,
|
| 392 |
+
"step": 320
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.72,
|
| 396 |
+
"eval_accuracy": 0.9551664677012867,
|
| 397 |
+
"eval_f1": 0.9612563044475011,
|
| 398 |
+
"eval_loss": 0.1314559429883957,
|
| 399 |
+
"eval_precision": 0.9706018518518519,
|
| 400 |
+
"eval_recall": 0.9520890099909174,
|
| 401 |
+
"eval_runtime": 35.9898,
|
| 402 |
+
"eval_samples_per_second": 209.476,
|
| 403 |
+
"eval_steps_per_second": 26.202,
|
| 404 |
+
"step": 325
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.73,
|
| 408 |
+
"learning_rate": 1.757142857142857e-05,
|
| 409 |
+
"loss": 0.1238,
|
| 410 |
+
"step": 328
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.75,
|
| 414 |
+
"learning_rate": 1.642857142857143e-05,
|
| 415 |
+
"loss": 0.1262,
|
| 416 |
+
"step": 336
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.76,
|
| 420 |
+
"learning_rate": 1.5285714285714286e-05,
|
| 421 |
+
"loss": 0.1627,
|
| 422 |
+
"step": 344
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.78,
|
| 426 |
+
"eval_accuracy": 0.9473404960870142,
|
| 427 |
+
"eval_f1": 0.9539603386292473,
|
| 428 |
+
"eval_loss": 0.14597517251968384,
|
| 429 |
+
"eval_precision": 0.9748755629296042,
|
| 430 |
+
"eval_recall": 0.9339237057220708,
|
| 431 |
+
"eval_runtime": 39.0107,
|
| 432 |
+
"eval_samples_per_second": 193.255,
|
| 433 |
+
"eval_steps_per_second": 24.173,
|
| 434 |
+
"step": 350
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.78,
|
| 438 |
+
"learning_rate": 1.4142857142857143e-05,
|
| 439 |
+
"loss": 0.1338,
|
| 440 |
+
"step": 352
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.8,
|
| 444 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 445 |
+
"loss": 0.1022,
|
| 446 |
+
"step": 360
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.82,
|
| 450 |
+
"learning_rate": 1.1857142857142858e-05,
|
| 451 |
+
"loss": 0.095,
|
| 452 |
+
"step": 368
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.83,
|
| 456 |
+
"eval_accuracy": 0.9562276163947473,
|
| 457 |
+
"eval_f1": 0.9622425629290619,
|
| 458 |
+
"eval_loss": 0.13147765398025513,
|
| 459 |
+
"eval_precision": 0.9697878228782287,
|
| 460 |
+
"eval_recall": 0.9548138056312443,
|
| 461 |
+
"eval_runtime": 37.6951,
|
| 462 |
+
"eval_samples_per_second": 199.999,
|
| 463 |
+
"eval_steps_per_second": 25.016,
|
| 464 |
+
"step": 375
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.83,
|
| 468 |
+
"learning_rate": 1.0714285714285714e-05,
|
| 469 |
+
"loss": 0.0907,
|
| 470 |
+
"step": 376
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.85,
|
| 474 |
+
"learning_rate": 9.571428571428572e-06,
|
| 475 |
+
"loss": 0.1099,
|
| 476 |
+
"step": 384
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.87,
|
| 480 |
+
"learning_rate": 8.428571428571429e-06,
|
| 481 |
+
"loss": 0.1498,
|
| 482 |
+
"step": 392
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.89,
|
| 486 |
+
"learning_rate": 7.285714285714286e-06,
|
| 487 |
+
"loss": 0.1221,
|
| 488 |
+
"step": 400
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.89,
|
| 492 |
+
"eval_accuracy": 0.9549011805279215,
|
| 493 |
+
"eval_f1": 0.961080586080586,
|
| 494 |
+
"eval_loss": 0.11751025170087814,
|
| 495 |
+
"eval_precision": 0.969067405355494,
|
| 496 |
+
"eval_recall": 0.9532243415077203,
|
| 497 |
+
"eval_runtime": 38.9296,
|
| 498 |
+
"eval_samples_per_second": 193.657,
|
| 499 |
+
"eval_steps_per_second": 24.223,
|
| 500 |
+
"step": 400
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.9,
|
| 504 |
+
"learning_rate": 6.142857142857143e-06,
|
| 505 |
+
"loss": 0.1109,
|
| 506 |
+
"step": 408
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.92,
|
| 510 |
+
"learning_rate": 5e-06,
|
| 511 |
+
"loss": 0.1069,
|
| 512 |
+
"step": 416
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.94,
|
| 516 |
+
"learning_rate": 3.857142857142857e-06,
|
| 517 |
+
"loss": 0.0952,
|
| 518 |
+
"step": 424
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.94,
|
| 522 |
+
"eval_accuracy": 0.9602069239952248,
|
| 523 |
+
"eval_f1": 0.9661552346570397,
|
| 524 |
+
"eval_loss": 0.113589346408844,
|
| 525 |
+
"eval_precision": 0.9600896860986547,
|
| 526 |
+
"eval_recall": 0.9722979109900091,
|
| 527 |
+
"eval_runtime": 39.1055,
|
| 528 |
+
"eval_samples_per_second": 192.786,
|
| 529 |
+
"eval_steps_per_second": 24.114,
|
| 530 |
+
"step": 425
|
| 531 |
+
}
|
| 532 |
+
],
|
| 533 |
+
"max_steps": 450,
|
| 534 |
+
"num_train_epochs": 1,
|
| 535 |
+
"total_flos": 4626643464096000.0,
|
| 536 |
+
"trial_name": null,
|
| 537 |
+
"trial_params": null
|
| 538 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0801ed8a1e7c36d496ab0b6de12f92609f20732b10feedd8031f99d7e85f7b24
|
| 3 |
+
size 3579
|