Initial commit
Browse files- config.json +52 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +466 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "roberta-large",
|
| 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": 1024,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "LABEL_0",
|
| 15 |
+
"1": "LABEL_1",
|
| 16 |
+
"2": "LABEL_2",
|
| 17 |
+
"3": "LABEL_3",
|
| 18 |
+
"4": "LABEL_4",
|
| 19 |
+
"5": "LABEL_5",
|
| 20 |
+
"6": "LABEL_6",
|
| 21 |
+
"7": "LABEL_7",
|
| 22 |
+
"8": "LABEL_8",
|
| 23 |
+
"9": "LABEL_9"
|
| 24 |
+
},
|
| 25 |
+
"initializer_range": 0.02,
|
| 26 |
+
"intermediate_size": 4096,
|
| 27 |
+
"label2id": {
|
| 28 |
+
"LABEL_0": 0,
|
| 29 |
+
"LABEL_1": 1,
|
| 30 |
+
"LABEL_2": 2,
|
| 31 |
+
"LABEL_3": 3,
|
| 32 |
+
"LABEL_4": 4,
|
| 33 |
+
"LABEL_5": 5,
|
| 34 |
+
"LABEL_6": 6,
|
| 35 |
+
"LABEL_7": 7,
|
| 36 |
+
"LABEL_8": 8,
|
| 37 |
+
"LABEL_9": 9
|
| 38 |
+
},
|
| 39 |
+
"layer_norm_eps": 1e-05,
|
| 40 |
+
"max_position_embeddings": 514,
|
| 41 |
+
"model_type": "roberta",
|
| 42 |
+
"num_attention_heads": 16,
|
| 43 |
+
"num_hidden_layers": 24,
|
| 44 |
+
"pad_token_id": 1,
|
| 45 |
+
"position_embedding_type": "absolute",
|
| 46 |
+
"problem_type": "single_label_classification",
|
| 47 |
+
"torch_dtype": "float32",
|
| 48 |
+
"transformers_version": "4.30.2",
|
| 49 |
+
"type_vocab_size": 1,
|
| 50 |
+
"use_cache": true,
|
| 51 |
+
"vocab_size": 50265
|
| 52 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:82d463bc89c2443310eb78dc4faec97f4903ee0a65a5f421a8d7ab1f40b60c05
|
| 3 |
+
size 2843188637
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d0ed97c57c19a805f0f01f9ddb2d98aae4c97c2c769cc94ace46848da606680b
|
| 3 |
+
size 1421617133
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3289d2eff02afc09eb0ae89af6dcbee574683fb90fd640177040b01c596204e
|
| 3 |
+
size 14503
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab9ba4151760aebd5075551d2af7eb2e9a6dab05ede86b5086da213164825ba4
|
| 3 |
+
size 623
|
trainer_state.json
ADDED
|
@@ -0,0 +1,466 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.951527777777778,
|
| 3 |
+
"best_model_checkpoint": "/n/holyscratch01/economics/esilcock/syria_models/8_5e-06/checkpoint-740",
|
| 4 |
+
"epoch": 17.61904761904762,
|
| 5 |
+
"global_step": 740,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.48,
|
| 12 |
+
"eval_accuracy": 0.18055555555555555,
|
| 13 |
+
"eval_f1": 0.05162907268170425,
|
| 14 |
+
"eval_loss": 2.1331899166107178,
|
| 15 |
+
"eval_precision": 0.05357142857142857,
|
| 16 |
+
"eval_recall": 0.08588235294117648,
|
| 17 |
+
"eval_runtime": 2.9184,
|
| 18 |
+
"eval_samples_per_second": 24.671,
|
| 19 |
+
"eval_steps_per_second": 3.084,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.95,
|
| 24 |
+
"eval_accuracy": 0.25,
|
| 25 |
+
"eval_f1": 0.10818399044205496,
|
| 26 |
+
"eval_loss": 2.088547468185425,
|
| 27 |
+
"eval_precision": 0.09869674185463659,
|
| 28 |
+
"eval_recall": 0.12676470588235295,
|
| 29 |
+
"eval_runtime": 2.6064,
|
| 30 |
+
"eval_samples_per_second": 27.625,
|
| 31 |
+
"eval_steps_per_second": 3.453,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 1.43,
|
| 36 |
+
"eval_accuracy": 0.2777777777777778,
|
| 37 |
+
"eval_f1": 0.09715938834550213,
|
| 38 |
+
"eval_loss": 1.9733208417892456,
|
| 39 |
+
"eval_precision": 0.08442028985507247,
|
| 40 |
+
"eval_recall": 0.13751633986928105,
|
| 41 |
+
"eval_runtime": 2.5925,
|
| 42 |
+
"eval_samples_per_second": 27.773,
|
| 43 |
+
"eval_steps_per_second": 3.472,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 1.9,
|
| 48 |
+
"eval_accuracy": 0.5555555555555556,
|
| 49 |
+
"eval_f1": 0.22415689149560114,
|
| 50 |
+
"eval_loss": 1.8020447492599487,
|
| 51 |
+
"eval_precision": 0.21013071895424837,
|
| 52 |
+
"eval_recall": 0.26833333333333337,
|
| 53 |
+
"eval_runtime": 2.5848,
|
| 54 |
+
"eval_samples_per_second": 27.855,
|
| 55 |
+
"eval_steps_per_second": 3.482,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.38,
|
| 60 |
+
"eval_accuracy": 0.5833333333333334,
|
| 61 |
+
"eval_f1": 0.2961904761904762,
|
| 62 |
+
"eval_loss": 1.4616996049880981,
|
| 63 |
+
"eval_precision": 0.3290064102564102,
|
| 64 |
+
"eval_recall": 0.335686274509804,
|
| 65 |
+
"eval_runtime": 2.9293,
|
| 66 |
+
"eval_samples_per_second": 24.579,
|
| 67 |
+
"eval_steps_per_second": 3.072,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 2.86,
|
| 72 |
+
"eval_accuracy": 0.6527777777777778,
|
| 73 |
+
"eval_f1": 0.43249639249639255,
|
| 74 |
+
"eval_loss": 1.2280499935150146,
|
| 75 |
+
"eval_precision": 0.4880341880341881,
|
| 76 |
+
"eval_recall": 0.4904901960784313,
|
| 77 |
+
"eval_runtime": 2.6424,
|
| 78 |
+
"eval_samples_per_second": 27.248,
|
| 79 |
+
"eval_steps_per_second": 3.406,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 3.33,
|
| 84 |
+
"eval_accuracy": 0.75,
|
| 85 |
+
"eval_f1": 0.6376634768740032,
|
| 86 |
+
"eval_loss": 0.9447543025016785,
|
| 87 |
+
"eval_precision": 0.7244047619047619,
|
| 88 |
+
"eval_recall": 0.6264705882352941,
|
| 89 |
+
"eval_runtime": 2.6104,
|
| 90 |
+
"eval_samples_per_second": 27.582,
|
| 91 |
+
"eval_steps_per_second": 3.448,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 3.81,
|
| 96 |
+
"eval_accuracy": 0.8055555555555556,
|
| 97 |
+
"eval_f1": 0.6916894699503395,
|
| 98 |
+
"eval_loss": 0.7263244986534119,
|
| 99 |
+
"eval_precision": 0.7515151515151516,
|
| 100 |
+
"eval_recall": 0.6826797385620915,
|
| 101 |
+
"eval_runtime": 2.6192,
|
| 102 |
+
"eval_samples_per_second": 27.489,
|
| 103 |
+
"eval_steps_per_second": 3.436,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 4.29,
|
| 108 |
+
"eval_accuracy": 0.8611111111111112,
|
| 109 |
+
"eval_f1": 0.780800727050727,
|
| 110 |
+
"eval_loss": 0.6345593333244324,
|
| 111 |
+
"eval_precision": 0.7710130718954249,
|
| 112 |
+
"eval_recall": 0.826797385620915,
|
| 113 |
+
"eval_runtime": 2.6126,
|
| 114 |
+
"eval_samples_per_second": 27.559,
|
| 115 |
+
"eval_steps_per_second": 3.445,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 4.76,
|
| 120 |
+
"eval_accuracy": 0.875,
|
| 121 |
+
"eval_f1": 0.8041340603840604,
|
| 122 |
+
"eval_loss": 0.5365733504295349,
|
| 123 |
+
"eval_precision": 0.8055159958720332,
|
| 124 |
+
"eval_recall": 0.8326797385620915,
|
| 125 |
+
"eval_runtime": 2.7985,
|
| 126 |
+
"eval_samples_per_second": 25.728,
|
| 127 |
+
"eval_steps_per_second": 3.216,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 5.24,
|
| 132 |
+
"eval_accuracy": 0.875,
|
| 133 |
+
"eval_f1": 0.8001352813852813,
|
| 134 |
+
"eval_loss": 0.41519981622695923,
|
| 135 |
+
"eval_precision": 0.800515995872033,
|
| 136 |
+
"eval_recall": 0.8187908496732026,
|
| 137 |
+
"eval_runtime": 2.6402,
|
| 138 |
+
"eval_samples_per_second": 27.271,
|
| 139 |
+
"eval_steps_per_second": 3.409,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 5.71,
|
| 144 |
+
"eval_accuracy": 0.8472222222222222,
|
| 145 |
+
"eval_f1": 0.774096712111418,
|
| 146 |
+
"eval_loss": 0.41259926557540894,
|
| 147 |
+
"eval_precision": 0.7647058823529411,
|
| 148 |
+
"eval_recall": 0.8209150326797385,
|
| 149 |
+
"eval_runtime": 2.6387,
|
| 150 |
+
"eval_samples_per_second": 27.286,
|
| 151 |
+
"eval_steps_per_second": 3.411,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 6.19,
|
| 156 |
+
"eval_accuracy": 0.875,
|
| 157 |
+
"eval_f1": 0.7898691749427044,
|
| 158 |
+
"eval_loss": 0.34084904193878174,
|
| 159 |
+
"eval_precision": 0.7705882352941177,
|
| 160 |
+
"eval_recall": 0.8379084967320262,
|
| 161 |
+
"eval_runtime": 2.7068,
|
| 162 |
+
"eval_samples_per_second": 26.6,
|
| 163 |
+
"eval_steps_per_second": 3.325,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 6.67,
|
| 168 |
+
"eval_accuracy": 0.8611111111111112,
|
| 169 |
+
"eval_f1": 0.7678326775480475,
|
| 170 |
+
"eval_loss": 0.39358699321746826,
|
| 171 |
+
"eval_precision": 0.7325162337662338,
|
| 172 |
+
"eval_recall": 0.8477124183006536,
|
| 173 |
+
"eval_runtime": 2.6003,
|
| 174 |
+
"eval_samples_per_second": 27.689,
|
| 175 |
+
"eval_steps_per_second": 3.461,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 7.14,
|
| 180 |
+
"eval_accuracy": 0.8888888888888888,
|
| 181 |
+
"eval_f1": 0.8779316203509753,
|
| 182 |
+
"eval_loss": 0.2763499915599823,
|
| 183 |
+
"eval_precision": 0.8507142857142858,
|
| 184 |
+
"eval_recall": 0.9372549019607843,
|
| 185 |
+
"eval_runtime": 2.6589,
|
| 186 |
+
"eval_samples_per_second": 27.078,
|
| 187 |
+
"eval_steps_per_second": 3.385,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 7.62,
|
| 192 |
+
"eval_accuracy": 0.9305555555555556,
|
| 193 |
+
"eval_f1": 0.9120108649520414,
|
| 194 |
+
"eval_loss": 0.217352032661438,
|
| 195 |
+
"eval_precision": 0.8800000000000001,
|
| 196 |
+
"eval_recall": 0.9653594771241831,
|
| 197 |
+
"eval_runtime": 2.6648,
|
| 198 |
+
"eval_samples_per_second": 27.019,
|
| 199 |
+
"eval_steps_per_second": 3.377,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 8.1,
|
| 204 |
+
"eval_accuracy": 0.8888888888888888,
|
| 205 |
+
"eval_f1": 0.8755410011444494,
|
| 206 |
+
"eval_loss": 0.22454272210597992,
|
| 207 |
+
"eval_precision": 0.8466666666666667,
|
| 208 |
+
"eval_recall": 0.9424836601307189,
|
| 209 |
+
"eval_runtime": 2.6406,
|
| 210 |
+
"eval_samples_per_second": 27.267,
|
| 211 |
+
"eval_steps_per_second": 3.408,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 8.57,
|
| 216 |
+
"eval_accuracy": 0.9444444444444444,
|
| 217 |
+
"eval_f1": 0.9410636043273615,
|
| 218 |
+
"eval_loss": 0.19036133587360382,
|
| 219 |
+
"eval_precision": 0.925,
|
| 220 |
+
"eval_recall": 0.9712418300653596,
|
| 221 |
+
"eval_runtime": 2.6164,
|
| 222 |
+
"eval_samples_per_second": 27.519,
|
| 223 |
+
"eval_steps_per_second": 3.44,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 9.05,
|
| 228 |
+
"eval_accuracy": 0.9444444444444444,
|
| 229 |
+
"eval_f1": 0.9299524932162504,
|
| 230 |
+
"eval_loss": 0.20558035373687744,
|
| 231 |
+
"eval_precision": 0.905,
|
| 232 |
+
"eval_recall": 0.9712418300653596,
|
| 233 |
+
"eval_runtime": 2.6164,
|
| 234 |
+
"eval_samples_per_second": 27.519,
|
| 235 |
+
"eval_steps_per_second": 3.44,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 9.52,
|
| 240 |
+
"eval_accuracy": 0.9444444444444444,
|
| 241 |
+
"eval_f1": 0.9299524932162504,
|
| 242 |
+
"eval_loss": 0.20008066296577454,
|
| 243 |
+
"eval_precision": 0.905,
|
| 244 |
+
"eval_recall": 0.9712418300653596,
|
| 245 |
+
"eval_runtime": 2.5909,
|
| 246 |
+
"eval_samples_per_second": 27.79,
|
| 247 |
+
"eval_steps_per_second": 3.474,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 10.0,
|
| 252 |
+
"eval_accuracy": 0.9444444444444444,
|
| 253 |
+
"eval_f1": 0.9410636043273615,
|
| 254 |
+
"eval_loss": 0.17470227181911469,
|
| 255 |
+
"eval_precision": 0.925,
|
| 256 |
+
"eval_recall": 0.9712418300653596,
|
| 257 |
+
"eval_runtime": 2.8957,
|
| 258 |
+
"eval_samples_per_second": 24.864,
|
| 259 |
+
"eval_steps_per_second": 3.108,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 10.48,
|
| 264 |
+
"eval_accuracy": 0.9583333333333334,
|
| 265 |
+
"eval_f1": 0.9476656268568033,
|
| 266 |
+
"eval_loss": 0.11614257842302322,
|
| 267 |
+
"eval_precision": 0.93,
|
| 268 |
+
"eval_recall": 0.977124183006536,
|
| 269 |
+
"eval_runtime": 2.6154,
|
| 270 |
+
"eval_samples_per_second": 27.53,
|
| 271 |
+
"eval_steps_per_second": 3.441,
|
| 272 |
+
"step": 440
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 10.95,
|
| 276 |
+
"eval_accuracy": 0.9305555555555556,
|
| 277 |
+
"eval_f1": 0.9205122655122654,
|
| 278 |
+
"eval_loss": 0.2461088001728058,
|
| 279 |
+
"eval_precision": 0.8983333333333334,
|
| 280 |
+
"eval_recall": 0.9601307189542485,
|
| 281 |
+
"eval_runtime": 2.6479,
|
| 282 |
+
"eval_samples_per_second": 27.191,
|
| 283 |
+
"eval_steps_per_second": 3.399,
|
| 284 |
+
"step": 460
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 11.43,
|
| 288 |
+
"eval_accuracy": 0.9305555555555556,
|
| 289 |
+
"eval_f1": 0.9205122655122654,
|
| 290 |
+
"eval_loss": 0.20940905809402466,
|
| 291 |
+
"eval_precision": 0.8983333333333334,
|
| 292 |
+
"eval_recall": 0.9601307189542485,
|
| 293 |
+
"eval_runtime": 2.5922,
|
| 294 |
+
"eval_samples_per_second": 27.776,
|
| 295 |
+
"eval_steps_per_second": 3.472,
|
| 296 |
+
"step": 480
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 11.9,
|
| 300 |
+
"learning_rate": 3.015873015873016e-06,
|
| 301 |
+
"loss": 0.6157,
|
| 302 |
+
"step": 500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 11.9,
|
| 306 |
+
"eval_accuracy": 0.9444444444444444,
|
| 307 |
+
"eval_f1": 0.9299524932162504,
|
| 308 |
+
"eval_loss": 0.19723846018314362,
|
| 309 |
+
"eval_precision": 0.905,
|
| 310 |
+
"eval_recall": 0.9712418300653596,
|
| 311 |
+
"eval_runtime": 2.6516,
|
| 312 |
+
"eval_samples_per_second": 27.154,
|
| 313 |
+
"eval_steps_per_second": 3.394,
|
| 314 |
+
"step": 500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 12.38,
|
| 318 |
+
"eval_accuracy": 0.9305555555555556,
|
| 319 |
+
"eval_f1": 0.9120108649520414,
|
| 320 |
+
"eval_loss": 0.20041564106941223,
|
| 321 |
+
"eval_precision": 0.8800000000000001,
|
| 322 |
+
"eval_recall": 0.9653594771241831,
|
| 323 |
+
"eval_runtime": 2.6152,
|
| 324 |
+
"eval_samples_per_second": 27.531,
|
| 325 |
+
"eval_steps_per_second": 3.441,
|
| 326 |
+
"step": 520
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 12.86,
|
| 330 |
+
"eval_accuracy": 0.9305555555555556,
|
| 331 |
+
"eval_f1": 0.9120108649520414,
|
| 332 |
+
"eval_loss": 0.1812739223241806,
|
| 333 |
+
"eval_precision": 0.8800000000000001,
|
| 334 |
+
"eval_recall": 0.9653594771241831,
|
| 335 |
+
"eval_runtime": 2.6055,
|
| 336 |
+
"eval_samples_per_second": 27.634,
|
| 337 |
+
"eval_steps_per_second": 3.454,
|
| 338 |
+
"step": 540
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 13.33,
|
| 342 |
+
"eval_accuracy": 0.9583333333333334,
|
| 343 |
+
"eval_f1": 0.9476656268568033,
|
| 344 |
+
"eval_loss": 0.16734400391578674,
|
| 345 |
+
"eval_precision": 0.93,
|
| 346 |
+
"eval_recall": 0.977124183006536,
|
| 347 |
+
"eval_runtime": 2.6389,
|
| 348 |
+
"eval_samples_per_second": 27.284,
|
| 349 |
+
"eval_steps_per_second": 3.411,
|
| 350 |
+
"step": 560
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 13.81,
|
| 354 |
+
"eval_accuracy": 0.9444444444444444,
|
| 355 |
+
"eval_f1": 0.9382070707070707,
|
| 356 |
+
"eval_loss": 0.17178066074848175,
|
| 357 |
+
"eval_precision": 0.9237500000000001,
|
| 358 |
+
"eval_recall": 0.966013071895425,
|
| 359 |
+
"eval_runtime": 2.7026,
|
| 360 |
+
"eval_samples_per_second": 26.641,
|
| 361 |
+
"eval_steps_per_second": 3.33,
|
| 362 |
+
"step": 580
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 14.29,
|
| 366 |
+
"eval_accuracy": 0.9444444444444444,
|
| 367 |
+
"eval_f1": 0.9382070707070707,
|
| 368 |
+
"eval_loss": 0.15890704095363617,
|
| 369 |
+
"eval_precision": 0.9237500000000001,
|
| 370 |
+
"eval_recall": 0.966013071895425,
|
| 371 |
+
"eval_runtime": 2.605,
|
| 372 |
+
"eval_samples_per_second": 27.639,
|
| 373 |
+
"eval_steps_per_second": 3.455,
|
| 374 |
+
"step": 600
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 14.76,
|
| 378 |
+
"eval_accuracy": 0.9583333333333334,
|
| 379 |
+
"eval_f1": 0.9476656268568033,
|
| 380 |
+
"eval_loss": 0.131246879696846,
|
| 381 |
+
"eval_precision": 0.93,
|
| 382 |
+
"eval_recall": 0.977124183006536,
|
| 383 |
+
"eval_runtime": 2.6046,
|
| 384 |
+
"eval_samples_per_second": 27.643,
|
| 385 |
+
"eval_steps_per_second": 3.455,
|
| 386 |
+
"step": 620
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 15.24,
|
| 390 |
+
"eval_accuracy": 0.9583333333333334,
|
| 391 |
+
"eval_f1": 0.9476656268568033,
|
| 392 |
+
"eval_loss": 0.1490662544965744,
|
| 393 |
+
"eval_precision": 0.93,
|
| 394 |
+
"eval_recall": 0.977124183006536,
|
| 395 |
+
"eval_runtime": 2.5887,
|
| 396 |
+
"eval_samples_per_second": 27.813,
|
| 397 |
+
"eval_steps_per_second": 3.477,
|
| 398 |
+
"step": 640
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 15.71,
|
| 402 |
+
"eval_accuracy": 0.9444444444444444,
|
| 403 |
+
"eval_f1": 0.9382070707070707,
|
| 404 |
+
"eval_loss": 0.22067798674106598,
|
| 405 |
+
"eval_precision": 0.9237500000000001,
|
| 406 |
+
"eval_recall": 0.966013071895425,
|
| 407 |
+
"eval_runtime": 2.6374,
|
| 408 |
+
"eval_samples_per_second": 27.299,
|
| 409 |
+
"eval_steps_per_second": 3.412,
|
| 410 |
+
"step": 660
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 16.19,
|
| 414 |
+
"eval_accuracy": 0.9583333333333334,
|
| 415 |
+
"eval_f1": 0.9476656268568033,
|
| 416 |
+
"eval_loss": 0.22119665145874023,
|
| 417 |
+
"eval_precision": 0.93,
|
| 418 |
+
"eval_recall": 0.977124183006536,
|
| 419 |
+
"eval_runtime": 2.6042,
|
| 420 |
+
"eval_samples_per_second": 27.647,
|
| 421 |
+
"eval_steps_per_second": 3.456,
|
| 422 |
+
"step": 680
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 16.67,
|
| 426 |
+
"eval_accuracy": 0.9583333333333334,
|
| 427 |
+
"eval_f1": 0.9476656268568033,
|
| 428 |
+
"eval_loss": 0.2090146392583847,
|
| 429 |
+
"eval_precision": 0.93,
|
| 430 |
+
"eval_recall": 0.977124183006536,
|
| 431 |
+
"eval_runtime": 2.6202,
|
| 432 |
+
"eval_samples_per_second": 27.478,
|
| 433 |
+
"eval_steps_per_second": 3.435,
|
| 434 |
+
"step": 700
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 17.14,
|
| 438 |
+
"eval_accuracy": 0.9583333333333334,
|
| 439 |
+
"eval_f1": 0.9476656268568033,
|
| 440 |
+
"eval_loss": 0.18439681828022003,
|
| 441 |
+
"eval_precision": 0.93,
|
| 442 |
+
"eval_recall": 0.977124183006536,
|
| 443 |
+
"eval_runtime": 2.5985,
|
| 444 |
+
"eval_samples_per_second": 27.708,
|
| 445 |
+
"eval_steps_per_second": 3.464,
|
| 446 |
+
"step": 720
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 17.62,
|
| 450 |
+
"eval_accuracy": 0.9583333333333334,
|
| 451 |
+
"eval_f1": 0.951527777777778,
|
| 452 |
+
"eval_loss": 0.17655248939990997,
|
| 453 |
+
"eval_precision": 0.9355555555555556,
|
| 454 |
+
"eval_recall": 0.977124183006536,
|
| 455 |
+
"eval_runtime": 2.893,
|
| 456 |
+
"eval_samples_per_second": 24.888,
|
| 457 |
+
"eval_steps_per_second": 3.111,
|
| 458 |
+
"step": 740
|
| 459 |
+
}
|
| 460 |
+
],
|
| 461 |
+
"max_steps": 1260,
|
| 462 |
+
"num_train_epochs": 30,
|
| 463 |
+
"total_flos": 5517182798069760.0,
|
| 464 |
+
"trial_name": null,
|
| 465 |
+
"trial_params": null
|
| 466 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a56512e6b197a9065fc0225985cb7fafdc9a9387c5857b57e407e961020ab433
|
| 3 |
+
size 4015
|