Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- cross_cell_type_generization/L6/B cells/.DS_Store +0 -0
- cross_cell_type_generization/L6/B cells/all_results.json +8 -0
- cross_cell_type_generization/L6/B cells/config.json +25 -0
- cross_cell_type_generization/L6/B cells/eval_results.json +8 -0
- cross_cell_type_generization/L6/B cells/trainer_state.json +450 -0
- cross_cell_type_generization/L6/Fibroblasts/trainer_state.json +450 -0
- data_curation&frozen_layers/dataset1_all/F0/optimizer.pt +3 -0
- data_curation&frozen_layers/dataset3_2048/F2/rng_state.pth +3 -0
- data_curation&frozen_layers/dataset3_2048/F2/scheduler.pt +3 -0
- data_curation&frozen_layers/dataset3_2048/F4/predictions.pickle +3 -0
- data_curation&frozen_layers/dataset3_2048/F4/training_args.bin +3 -0
- scaling_performance/.DS_Store +0 -0
- scaling_performance/1000/.DS_Store +0 -0
- scaling_performance/1000/L2/.DS_Store +0 -0
- scaling_performance/1000/L2/all_results.json +8 -0
- scaling_performance/1000/L2/config.json +25 -0
- scaling_performance/1000/L2/eval_results.json +8 -0
- scaling_performance/1000/L2/trainer_state.json +712 -0
- scaling_performance/1000/L4/.DS_Store +0 -0
- scaling_performance/1000/L4/all_results.json +8 -0
- scaling_performance/1000/L4/config.json +25 -0
- scaling_performance/1000/L4/eval_results.json +8 -0
- scaling_performance/1000/L4/trainer_state.json +480 -0
- scaling_performance/1000/fine-tuned/.DS_Store +0 -0
- scaling_performance/1000/fine-tuned/all_results.json +8 -0
- scaling_performance/1000/fine-tuned/config.json +27 -0
- scaling_performance/1000/fine-tuned/eval_results.json +8 -0
- scaling_performance/1000/fine-tuned/trainer_state.json +1756 -0
- scaling_performance/2000/.DS_Store +0 -0
- scaling_performance/2000/L1/.DS_Store +0 -0
- scaling_performance/2000/L1/all_results.json +8 -0
- scaling_performance/2000/L1/config.json +25 -0
- scaling_performance/2000/L1/eval_results.json +8 -0
- scaling_performance/2000/L1/trainer_state.json +886 -0
- scaling_performance/2000/L2/.DS_Store +0 -0
- scaling_performance/2000/L2/all_results.json +8 -0
- scaling_performance/2000/L2/config.json +25 -0
- scaling_performance/2000/L2/eval_results.json +8 -0
- scaling_performance/2000/L2/trainer_state.json +596 -0
- scaling_performance/2000/L4/.DS_Store +0 -0
- scaling_performance/2000/L4/all_results.json +8 -0
- scaling_performance/2000/L4/config.json +25 -0
- scaling_performance/2000/L4/eval_results.json +8 -0
- scaling_performance/2000/L4/trainer_state.json +596 -0
- scaling_performance/2000/L6/.DS_Store +0 -0
- scaling_performance/2000/L6/all_results.json +8 -0
- scaling_performance/2000/L6/config.json +25 -0
- scaling_performance/2000/L6/eval_results.json +8 -0
- scaling_performance/2000/L6/trainer_state.json +596 -0
- scaling_performance/2000/fine-tuned/.DS_Store +0 -0
cross_cell_type_generization/L6/B cells/.DS_Store
CHANGED
|
Binary files a/cross_cell_type_generization/L6/B cells/.DS_Store and b/cross_cell_type_generization/L6/B cells/.DS_Store differ
|
|
|
cross_cell_type_generization/L6/B cells/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.9055555555555556,
|
| 3 |
+
"test_loss": 0.26022690534591675,
|
| 4 |
+
"test_macro_f1": 0.881574364332985,
|
| 5 |
+
"test_runtime": 3.0564,
|
| 6 |
+
"test_samples_per_second": 294.462,
|
| 7 |
+
"test_steps_per_second": 2.945
|
| 8 |
+
}
|
cross_cell_type_generization/L6/B cells/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.02,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.02,
|
| 9 |
+
"hidden_size": 256,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 512,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 2048,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 4,
|
| 16 |
+
"num_hidden_layers": 6,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"problem_type": "single_label_classification",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.28.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
cross_cell_type_generization/L6/B cells/eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.9055555555555556,
|
| 3 |
+
"test_loss": 0.26022690534591675,
|
| 4 |
+
"test_macro_f1": 0.881574364332985,
|
| 5 |
+
"test_runtime": 3.0564,
|
| 6 |
+
"test_samples_per_second": 294.462,
|
| 7 |
+
"test_steps_per_second": 2.945
|
| 8 |
+
}
|
cross_cell_type_generization/L6/B cells/trainer_state.json
ADDED
|
@@ -0,0 +1,450 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.40387290716171265,
|
| 3 |
+
"best_model_checkpoint": "/vsphhome/fengguoqing/Geneformer/models/data_diversity/L6/B cells/fold4/checkpoint-1122",
|
| 4 |
+
"epoch": 8.0,
|
| 5 |
+
"global_step": 1496,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.13,
|
| 12 |
+
"learning_rate": 2.5e-06,
|
| 13 |
+
"loss": 0.6926,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.27,
|
| 18 |
+
"learning_rate": 5e-06,
|
| 19 |
+
"loss": 0.6933,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.4,
|
| 24 |
+
"learning_rate": 7.5e-06,
|
| 25 |
+
"loss": 0.6935,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.53,
|
| 30 |
+
"learning_rate": 1e-05,
|
| 31 |
+
"loss": 0.6936,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.67,
|
| 36 |
+
"learning_rate": 1.25e-05,
|
| 37 |
+
"loss": 0.6928,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.8,
|
| 42 |
+
"learning_rate": 1.5e-05,
|
| 43 |
+
"loss": 0.692,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.94,
|
| 48 |
+
"learning_rate": 1.75e-05,
|
| 49 |
+
"loss": 0.6907,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"eval_accuracy": 0.5210482872472142,
|
| 55 |
+
"eval_loss": 0.6905837059020996,
|
| 56 |
+
"eval_macro_f1": 0.41147454390525884,
|
| 57 |
+
"eval_runtime": 17.4321,
|
| 58 |
+
"eval_samples_per_second": 277.992,
|
| 59 |
+
"eval_steps_per_second": 2.696,
|
| 60 |
+
"step": 187
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 1.07,
|
| 64 |
+
"learning_rate": 2e-05,
|
| 65 |
+
"loss": 0.6904,
|
| 66 |
+
"step": 200
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.2,
|
| 70 |
+
"learning_rate": 2.25e-05,
|
| 71 |
+
"loss": 0.6898,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.34,
|
| 76 |
+
"learning_rate": 2.5e-05,
|
| 77 |
+
"loss": 0.6865,
|
| 78 |
+
"step": 250
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.47,
|
| 82 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 83 |
+
"loss": 0.6869,
|
| 84 |
+
"step": 275
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.6,
|
| 88 |
+
"learning_rate": 3e-05,
|
| 89 |
+
"loss": 0.6814,
|
| 90 |
+
"step": 300
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.74,
|
| 94 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 95 |
+
"loss": 0.6762,
|
| 96 |
+
"step": 325
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.87,
|
| 100 |
+
"learning_rate": 3.5e-05,
|
| 101 |
+
"loss": 0.6747,
|
| 102 |
+
"step": 350
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 2.0,
|
| 106 |
+
"eval_accuracy": 0.5984316962443252,
|
| 107 |
+
"eval_loss": 0.6632421016693115,
|
| 108 |
+
"eval_macro_f1": 0.5897740135113385,
|
| 109 |
+
"eval_runtime": 16.5825,
|
| 110 |
+
"eval_samples_per_second": 292.235,
|
| 111 |
+
"eval_steps_per_second": 2.834,
|
| 112 |
+
"step": 374
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.01,
|
| 116 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 117 |
+
"loss": 0.6664,
|
| 118 |
+
"step": 375
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 2.14,
|
| 122 |
+
"learning_rate": 4e-05,
|
| 123 |
+
"loss": 0.6356,
|
| 124 |
+
"step": 400
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.27,
|
| 128 |
+
"learning_rate": 4.25e-05,
|
| 129 |
+
"loss": 0.6121,
|
| 130 |
+
"step": 425
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.41,
|
| 134 |
+
"learning_rate": 4.5e-05,
|
| 135 |
+
"loss": 0.5833,
|
| 136 |
+
"step": 450
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.54,
|
| 140 |
+
"learning_rate": 4.75e-05,
|
| 141 |
+
"loss": 0.5619,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.67,
|
| 146 |
+
"learning_rate": 5e-05,
|
| 147 |
+
"loss": 0.5355,
|
| 148 |
+
"step": 500
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.81,
|
| 152 |
+
"learning_rate": 4.9858757062146896e-05,
|
| 153 |
+
"loss": 0.4877,
|
| 154 |
+
"step": 525
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.94,
|
| 158 |
+
"learning_rate": 4.971751412429379e-05,
|
| 159 |
+
"loss": 0.4662,
|
| 160 |
+
"step": 550
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 3.0,
|
| 164 |
+
"eval_accuracy": 0.7631035905901775,
|
| 165 |
+
"eval_loss": 0.48640862107276917,
|
| 166 |
+
"eval_macro_f1": 0.7547183753731113,
|
| 167 |
+
"eval_runtime": 17.7844,
|
| 168 |
+
"eval_samples_per_second": 272.486,
|
| 169 |
+
"eval_steps_per_second": 2.643,
|
| 170 |
+
"step": 561
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 3.07,
|
| 174 |
+
"learning_rate": 4.957627118644068e-05,
|
| 175 |
+
"loss": 0.4818,
|
| 176 |
+
"step": 575
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.21,
|
| 180 |
+
"learning_rate": 4.9435028248587575e-05,
|
| 181 |
+
"loss": 0.458,
|
| 182 |
+
"step": 600
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.34,
|
| 186 |
+
"learning_rate": 4.929378531073446e-05,
|
| 187 |
+
"loss": 0.424,
|
| 188 |
+
"step": 625
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.48,
|
| 192 |
+
"learning_rate": 4.915254237288136e-05,
|
| 193 |
+
"loss": 0.4373,
|
| 194 |
+
"step": 650
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.61,
|
| 198 |
+
"learning_rate": 4.9011299435028255e-05,
|
| 199 |
+
"loss": 0.4121,
|
| 200 |
+
"step": 675
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.74,
|
| 204 |
+
"learning_rate": 4.887005649717514e-05,
|
| 205 |
+
"loss": 0.397,
|
| 206 |
+
"step": 700
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.88,
|
| 210 |
+
"learning_rate": 4.8728813559322034e-05,
|
| 211 |
+
"loss": 0.3926,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 4.0,
|
| 216 |
+
"eval_accuracy": 0.7961205117622782,
|
| 217 |
+
"eval_loss": 0.42608675360679626,
|
| 218 |
+
"eval_macro_f1": 0.7956681990046659,
|
| 219 |
+
"eval_runtime": 17.1151,
|
| 220 |
+
"eval_samples_per_second": 283.141,
|
| 221 |
+
"eval_steps_per_second": 2.746,
|
| 222 |
+
"step": 748
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 4.01,
|
| 226 |
+
"learning_rate": 4.8587570621468934e-05,
|
| 227 |
+
"loss": 0.3983,
|
| 228 |
+
"step": 750
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 4.14,
|
| 232 |
+
"learning_rate": 4.844632768361582e-05,
|
| 233 |
+
"loss": 0.3468,
|
| 234 |
+
"step": 775
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 4.28,
|
| 238 |
+
"learning_rate": 4.8305084745762714e-05,
|
| 239 |
+
"loss": 0.3599,
|
| 240 |
+
"step": 800
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.41,
|
| 244 |
+
"learning_rate": 4.816384180790961e-05,
|
| 245 |
+
"loss": 0.3619,
|
| 246 |
+
"step": 825
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.55,
|
| 250 |
+
"learning_rate": 4.80225988700565e-05,
|
| 251 |
+
"loss": 0.3428,
|
| 252 |
+
"step": 850
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.68,
|
| 256 |
+
"learning_rate": 4.788135593220339e-05,
|
| 257 |
+
"loss": 0.3622,
|
| 258 |
+
"step": 875
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.81,
|
| 262 |
+
"learning_rate": 4.7740112994350286e-05,
|
| 263 |
+
"loss": 0.3451,
|
| 264 |
+
"step": 900
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.95,
|
| 268 |
+
"learning_rate": 4.759887005649718e-05,
|
| 269 |
+
"loss": 0.3714,
|
| 270 |
+
"step": 925
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 5.0,
|
| 274 |
+
"eval_accuracy": 0.7845645893520429,
|
| 275 |
+
"eval_loss": 0.4528542459011078,
|
| 276 |
+
"eval_macro_f1": 0.7833690643653564,
|
| 277 |
+
"eval_runtime": 17.1144,
|
| 278 |
+
"eval_samples_per_second": 283.153,
|
| 279 |
+
"eval_steps_per_second": 2.746,
|
| 280 |
+
"step": 935
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 5.08,
|
| 284 |
+
"learning_rate": 4.745762711864407e-05,
|
| 285 |
+
"loss": 0.339,
|
| 286 |
+
"step": 950
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 5.21,
|
| 290 |
+
"learning_rate": 4.7316384180790966e-05,
|
| 291 |
+
"loss": 0.2876,
|
| 292 |
+
"step": 975
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.35,
|
| 296 |
+
"learning_rate": 4.717514124293785e-05,
|
| 297 |
+
"loss": 0.355,
|
| 298 |
+
"step": 1000
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.48,
|
| 302 |
+
"learning_rate": 4.703389830508475e-05,
|
| 303 |
+
"loss": 0.3099,
|
| 304 |
+
"step": 1025
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.61,
|
| 308 |
+
"learning_rate": 4.689265536723164e-05,
|
| 309 |
+
"loss": 0.322,
|
| 310 |
+
"step": 1050
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.75,
|
| 314 |
+
"learning_rate": 4.675141242937853e-05,
|
| 315 |
+
"loss": 0.3246,
|
| 316 |
+
"step": 1075
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.88,
|
| 320 |
+
"learning_rate": 4.6610169491525425e-05,
|
| 321 |
+
"loss": 0.3719,
|
| 322 |
+
"step": 1100
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 6.0,
|
| 326 |
+
"eval_accuracy": 0.8186132893107718,
|
| 327 |
+
"eval_loss": 0.40387290716171265,
|
| 328 |
+
"eval_macro_f1": 0.8179837351502454,
|
| 329 |
+
"eval_runtime": 17.1237,
|
| 330 |
+
"eval_samples_per_second": 282.999,
|
| 331 |
+
"eval_steps_per_second": 2.745,
|
| 332 |
+
"step": 1122
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 6.02,
|
| 336 |
+
"learning_rate": 4.646892655367232e-05,
|
| 337 |
+
"loss": 0.2755,
|
| 338 |
+
"step": 1125
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 6.15,
|
| 342 |
+
"learning_rate": 4.632768361581921e-05,
|
| 343 |
+
"loss": 0.293,
|
| 344 |
+
"step": 1150
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 6.28,
|
| 348 |
+
"learning_rate": 4.6186440677966104e-05,
|
| 349 |
+
"loss": 0.3708,
|
| 350 |
+
"step": 1175
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 6.42,
|
| 354 |
+
"learning_rate": 4.6045197740113e-05,
|
| 355 |
+
"loss": 0.2885,
|
| 356 |
+
"step": 1200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.55,
|
| 360 |
+
"learning_rate": 4.590395480225989e-05,
|
| 361 |
+
"loss": 0.2562,
|
| 362 |
+
"step": 1225
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.68,
|
| 366 |
+
"learning_rate": 4.5762711864406784e-05,
|
| 367 |
+
"loss": 0.2814,
|
| 368 |
+
"step": 1250
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.82,
|
| 372 |
+
"learning_rate": 4.562146892655367e-05,
|
| 373 |
+
"loss": 0.3036,
|
| 374 |
+
"step": 1275
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.95,
|
| 378 |
+
"learning_rate": 4.548022598870056e-05,
|
| 379 |
+
"loss": 0.278,
|
| 380 |
+
"step": 1300
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 7.0,
|
| 384 |
+
"eval_accuracy": 0.7992158481221626,
|
| 385 |
+
"eval_loss": 0.44179922342300415,
|
| 386 |
+
"eval_macro_f1": 0.7983638814289351,
|
| 387 |
+
"eval_runtime": 16.9988,
|
| 388 |
+
"eval_samples_per_second": 285.079,
|
| 389 |
+
"eval_steps_per_second": 2.765,
|
| 390 |
+
"step": 1309
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 7.09,
|
| 394 |
+
"learning_rate": 4.533898305084746e-05,
|
| 395 |
+
"loss": 0.2443,
|
| 396 |
+
"step": 1325
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 7.22,
|
| 400 |
+
"learning_rate": 4.519774011299435e-05,
|
| 401 |
+
"loss": 0.2577,
|
| 402 |
+
"step": 1350
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 7.35,
|
| 406 |
+
"learning_rate": 4.505649717514124e-05,
|
| 407 |
+
"loss": 0.2615,
|
| 408 |
+
"step": 1375
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 7.49,
|
| 412 |
+
"learning_rate": 4.491525423728814e-05,
|
| 413 |
+
"loss": 0.2527,
|
| 414 |
+
"step": 1400
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.62,
|
| 418 |
+
"learning_rate": 4.477401129943503e-05,
|
| 419 |
+
"loss": 0.3119,
|
| 420 |
+
"step": 1425
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.75,
|
| 424 |
+
"learning_rate": 4.463276836158192e-05,
|
| 425 |
+
"loss": 0.2463,
|
| 426 |
+
"step": 1450
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.89,
|
| 430 |
+
"learning_rate": 4.4491525423728816e-05,
|
| 431 |
+
"loss": 0.2449,
|
| 432 |
+
"step": 1475
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 8.0,
|
| 436 |
+
"eval_accuracy": 0.8196450680973999,
|
| 437 |
+
"eval_loss": 0.4183129370212555,
|
| 438 |
+
"eval_macro_f1": 0.8195520921951007,
|
| 439 |
+
"eval_runtime": 16.9753,
|
| 440 |
+
"eval_samples_per_second": 285.473,
|
| 441 |
+
"eval_steps_per_second": 2.769,
|
| 442 |
+
"step": 1496
|
| 443 |
+
}
|
| 444 |
+
],
|
| 445 |
+
"max_steps": 9350,
|
| 446 |
+
"num_train_epochs": 50,
|
| 447 |
+
"total_flos": 6154099081543680.0,
|
| 448 |
+
"trial_name": null,
|
| 449 |
+
"trial_params": null
|
| 450 |
+
}
|
cross_cell_type_generization/L6/Fibroblasts/trainer_state.json
ADDED
|
@@ -0,0 +1,450 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.40387290716171265,
|
| 3 |
+
"best_model_checkpoint": "/vsphhome/fengguoqing/Geneformer/models/data_diversity/L6/Fibroblasts/fold4/checkpoint-1122",
|
| 4 |
+
"epoch": 8.0,
|
| 5 |
+
"global_step": 1496,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.13,
|
| 12 |
+
"learning_rate": 2.5e-06,
|
| 13 |
+
"loss": 0.6926,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.27,
|
| 18 |
+
"learning_rate": 5e-06,
|
| 19 |
+
"loss": 0.6933,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.4,
|
| 24 |
+
"learning_rate": 7.5e-06,
|
| 25 |
+
"loss": 0.6935,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.53,
|
| 30 |
+
"learning_rate": 1e-05,
|
| 31 |
+
"loss": 0.6936,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.67,
|
| 36 |
+
"learning_rate": 1.25e-05,
|
| 37 |
+
"loss": 0.6928,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.8,
|
| 42 |
+
"learning_rate": 1.5e-05,
|
| 43 |
+
"loss": 0.692,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.94,
|
| 48 |
+
"learning_rate": 1.75e-05,
|
| 49 |
+
"loss": 0.6907,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"eval_accuracy": 0.5210482872472142,
|
| 55 |
+
"eval_loss": 0.6905837059020996,
|
| 56 |
+
"eval_macro_f1": 0.41147454390525884,
|
| 57 |
+
"eval_runtime": 15.8319,
|
| 58 |
+
"eval_samples_per_second": 306.091,
|
| 59 |
+
"eval_steps_per_second": 2.969,
|
| 60 |
+
"step": 187
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"epoch": 1.07,
|
| 64 |
+
"learning_rate": 2e-05,
|
| 65 |
+
"loss": 0.6904,
|
| 66 |
+
"step": 200
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.2,
|
| 70 |
+
"learning_rate": 2.25e-05,
|
| 71 |
+
"loss": 0.6898,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.34,
|
| 76 |
+
"learning_rate": 2.5e-05,
|
| 77 |
+
"loss": 0.6865,
|
| 78 |
+
"step": 250
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.47,
|
| 82 |
+
"learning_rate": 2.7500000000000004e-05,
|
| 83 |
+
"loss": 0.6869,
|
| 84 |
+
"step": 275
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.6,
|
| 88 |
+
"learning_rate": 3e-05,
|
| 89 |
+
"loss": 0.6814,
|
| 90 |
+
"step": 300
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.74,
|
| 94 |
+
"learning_rate": 3.2500000000000004e-05,
|
| 95 |
+
"loss": 0.6762,
|
| 96 |
+
"step": 325
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.87,
|
| 100 |
+
"learning_rate": 3.5e-05,
|
| 101 |
+
"loss": 0.6747,
|
| 102 |
+
"step": 350
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 2.0,
|
| 106 |
+
"eval_accuracy": 0.5984316962443252,
|
| 107 |
+
"eval_loss": 0.6632421016693115,
|
| 108 |
+
"eval_macro_f1": 0.5897740135113385,
|
| 109 |
+
"eval_runtime": 16.4963,
|
| 110 |
+
"eval_samples_per_second": 293.762,
|
| 111 |
+
"eval_steps_per_second": 2.849,
|
| 112 |
+
"step": 374
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.01,
|
| 116 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 117 |
+
"loss": 0.6664,
|
| 118 |
+
"step": 375
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 2.14,
|
| 122 |
+
"learning_rate": 4e-05,
|
| 123 |
+
"loss": 0.6356,
|
| 124 |
+
"step": 400
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.27,
|
| 128 |
+
"learning_rate": 4.25e-05,
|
| 129 |
+
"loss": 0.6121,
|
| 130 |
+
"step": 425
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.41,
|
| 134 |
+
"learning_rate": 4.5e-05,
|
| 135 |
+
"loss": 0.5833,
|
| 136 |
+
"step": 450
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.54,
|
| 140 |
+
"learning_rate": 4.75e-05,
|
| 141 |
+
"loss": 0.5619,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.67,
|
| 146 |
+
"learning_rate": 5e-05,
|
| 147 |
+
"loss": 0.5355,
|
| 148 |
+
"step": 500
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.81,
|
| 152 |
+
"learning_rate": 4.9858757062146896e-05,
|
| 153 |
+
"loss": 0.4877,
|
| 154 |
+
"step": 525
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.94,
|
| 158 |
+
"learning_rate": 4.971751412429379e-05,
|
| 159 |
+
"loss": 0.4662,
|
| 160 |
+
"step": 550
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 3.0,
|
| 164 |
+
"eval_accuracy": 0.7631035905901775,
|
| 165 |
+
"eval_loss": 0.48640862107276917,
|
| 166 |
+
"eval_macro_f1": 0.7547183753731113,
|
| 167 |
+
"eval_runtime": 16.3862,
|
| 168 |
+
"eval_samples_per_second": 295.737,
|
| 169 |
+
"eval_steps_per_second": 2.868,
|
| 170 |
+
"step": 561
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 3.07,
|
| 174 |
+
"learning_rate": 4.957627118644068e-05,
|
| 175 |
+
"loss": 0.4818,
|
| 176 |
+
"step": 575
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.21,
|
| 180 |
+
"learning_rate": 4.9435028248587575e-05,
|
| 181 |
+
"loss": 0.458,
|
| 182 |
+
"step": 600
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.34,
|
| 186 |
+
"learning_rate": 4.929378531073446e-05,
|
| 187 |
+
"loss": 0.424,
|
| 188 |
+
"step": 625
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.48,
|
| 192 |
+
"learning_rate": 4.915254237288136e-05,
|
| 193 |
+
"loss": 0.4373,
|
| 194 |
+
"step": 650
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.61,
|
| 198 |
+
"learning_rate": 4.9011299435028255e-05,
|
| 199 |
+
"loss": 0.4121,
|
| 200 |
+
"step": 675
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.74,
|
| 204 |
+
"learning_rate": 4.887005649717514e-05,
|
| 205 |
+
"loss": 0.397,
|
| 206 |
+
"step": 700
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.88,
|
| 210 |
+
"learning_rate": 4.8728813559322034e-05,
|
| 211 |
+
"loss": 0.3926,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 4.0,
|
| 216 |
+
"eval_accuracy": 0.7961205117622782,
|
| 217 |
+
"eval_loss": 0.42608675360679626,
|
| 218 |
+
"eval_macro_f1": 0.7956681990046659,
|
| 219 |
+
"eval_runtime": 16.1617,
|
| 220 |
+
"eval_samples_per_second": 299.845,
|
| 221 |
+
"eval_steps_per_second": 2.908,
|
| 222 |
+
"step": 748
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 4.01,
|
| 226 |
+
"learning_rate": 4.8587570621468934e-05,
|
| 227 |
+
"loss": 0.3983,
|
| 228 |
+
"step": 750
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 4.14,
|
| 232 |
+
"learning_rate": 4.844632768361582e-05,
|
| 233 |
+
"loss": 0.3468,
|
| 234 |
+
"step": 775
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 4.28,
|
| 238 |
+
"learning_rate": 4.8305084745762714e-05,
|
| 239 |
+
"loss": 0.3599,
|
| 240 |
+
"step": 800
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.41,
|
| 244 |
+
"learning_rate": 4.816384180790961e-05,
|
| 245 |
+
"loss": 0.3619,
|
| 246 |
+
"step": 825
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.55,
|
| 250 |
+
"learning_rate": 4.80225988700565e-05,
|
| 251 |
+
"loss": 0.3428,
|
| 252 |
+
"step": 850
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.68,
|
| 256 |
+
"learning_rate": 4.788135593220339e-05,
|
| 257 |
+
"loss": 0.3622,
|
| 258 |
+
"step": 875
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.81,
|
| 262 |
+
"learning_rate": 4.7740112994350286e-05,
|
| 263 |
+
"loss": 0.3451,
|
| 264 |
+
"step": 900
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.95,
|
| 268 |
+
"learning_rate": 4.759887005649718e-05,
|
| 269 |
+
"loss": 0.3714,
|
| 270 |
+
"step": 925
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 5.0,
|
| 274 |
+
"eval_accuracy": 0.7845645893520429,
|
| 275 |
+
"eval_loss": 0.4528542459011078,
|
| 276 |
+
"eval_macro_f1": 0.7833690643653564,
|
| 277 |
+
"eval_runtime": 16.4437,
|
| 278 |
+
"eval_samples_per_second": 294.702,
|
| 279 |
+
"eval_steps_per_second": 2.858,
|
| 280 |
+
"step": 935
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 5.08,
|
| 284 |
+
"learning_rate": 4.745762711864407e-05,
|
| 285 |
+
"loss": 0.339,
|
| 286 |
+
"step": 950
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 5.21,
|
| 290 |
+
"learning_rate": 4.7316384180790966e-05,
|
| 291 |
+
"loss": 0.2876,
|
| 292 |
+
"step": 975
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.35,
|
| 296 |
+
"learning_rate": 4.717514124293785e-05,
|
| 297 |
+
"loss": 0.355,
|
| 298 |
+
"step": 1000
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.48,
|
| 302 |
+
"learning_rate": 4.703389830508475e-05,
|
| 303 |
+
"loss": 0.3099,
|
| 304 |
+
"step": 1025
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.61,
|
| 308 |
+
"learning_rate": 4.689265536723164e-05,
|
| 309 |
+
"loss": 0.322,
|
| 310 |
+
"step": 1050
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.75,
|
| 314 |
+
"learning_rate": 4.675141242937853e-05,
|
| 315 |
+
"loss": 0.3246,
|
| 316 |
+
"step": 1075
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.88,
|
| 320 |
+
"learning_rate": 4.6610169491525425e-05,
|
| 321 |
+
"loss": 0.3719,
|
| 322 |
+
"step": 1100
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 6.0,
|
| 326 |
+
"eval_accuracy": 0.8186132893107718,
|
| 327 |
+
"eval_loss": 0.40387290716171265,
|
| 328 |
+
"eval_macro_f1": 0.8179837351502454,
|
| 329 |
+
"eval_runtime": 15.9198,
|
| 330 |
+
"eval_samples_per_second": 304.4,
|
| 331 |
+
"eval_steps_per_second": 2.952,
|
| 332 |
+
"step": 1122
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 6.02,
|
| 336 |
+
"learning_rate": 4.646892655367232e-05,
|
| 337 |
+
"loss": 0.2755,
|
| 338 |
+
"step": 1125
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 6.15,
|
| 342 |
+
"learning_rate": 4.632768361581921e-05,
|
| 343 |
+
"loss": 0.293,
|
| 344 |
+
"step": 1150
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 6.28,
|
| 348 |
+
"learning_rate": 4.6186440677966104e-05,
|
| 349 |
+
"loss": 0.3708,
|
| 350 |
+
"step": 1175
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 6.42,
|
| 354 |
+
"learning_rate": 4.6045197740113e-05,
|
| 355 |
+
"loss": 0.2885,
|
| 356 |
+
"step": 1200
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.55,
|
| 360 |
+
"learning_rate": 4.590395480225989e-05,
|
| 361 |
+
"loss": 0.2562,
|
| 362 |
+
"step": 1225
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.68,
|
| 366 |
+
"learning_rate": 4.5762711864406784e-05,
|
| 367 |
+
"loss": 0.2814,
|
| 368 |
+
"step": 1250
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.82,
|
| 372 |
+
"learning_rate": 4.562146892655367e-05,
|
| 373 |
+
"loss": 0.3036,
|
| 374 |
+
"step": 1275
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.95,
|
| 378 |
+
"learning_rate": 4.548022598870056e-05,
|
| 379 |
+
"loss": 0.278,
|
| 380 |
+
"step": 1300
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 7.0,
|
| 384 |
+
"eval_accuracy": 0.7992158481221626,
|
| 385 |
+
"eval_loss": 0.44179922342300415,
|
| 386 |
+
"eval_macro_f1": 0.7983638814289351,
|
| 387 |
+
"eval_runtime": 16.9377,
|
| 388 |
+
"eval_samples_per_second": 286.107,
|
| 389 |
+
"eval_steps_per_second": 2.775,
|
| 390 |
+
"step": 1309
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 7.09,
|
| 394 |
+
"learning_rate": 4.533898305084746e-05,
|
| 395 |
+
"loss": 0.2443,
|
| 396 |
+
"step": 1325
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 7.22,
|
| 400 |
+
"learning_rate": 4.519774011299435e-05,
|
| 401 |
+
"loss": 0.2577,
|
| 402 |
+
"step": 1350
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 7.35,
|
| 406 |
+
"learning_rate": 4.505649717514124e-05,
|
| 407 |
+
"loss": 0.2615,
|
| 408 |
+
"step": 1375
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 7.49,
|
| 412 |
+
"learning_rate": 4.491525423728814e-05,
|
| 413 |
+
"loss": 0.2527,
|
| 414 |
+
"step": 1400
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.62,
|
| 418 |
+
"learning_rate": 4.477401129943503e-05,
|
| 419 |
+
"loss": 0.3119,
|
| 420 |
+
"step": 1425
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.75,
|
| 424 |
+
"learning_rate": 4.463276836158192e-05,
|
| 425 |
+
"loss": 0.2463,
|
| 426 |
+
"step": 1450
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.89,
|
| 430 |
+
"learning_rate": 4.4491525423728816e-05,
|
| 431 |
+
"loss": 0.2449,
|
| 432 |
+
"step": 1475
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 8.0,
|
| 436 |
+
"eval_accuracy": 0.8196450680973999,
|
| 437 |
+
"eval_loss": 0.4183129370212555,
|
| 438 |
+
"eval_macro_f1": 0.8195520921951007,
|
| 439 |
+
"eval_runtime": 16.081,
|
| 440 |
+
"eval_samples_per_second": 301.35,
|
| 441 |
+
"eval_steps_per_second": 2.923,
|
| 442 |
+
"step": 1496
|
| 443 |
+
}
|
| 444 |
+
],
|
| 445 |
+
"max_steps": 9350,
|
| 446 |
+
"num_train_epochs": 50,
|
| 447 |
+
"total_flos": 6154099081543680.0,
|
| 448 |
+
"trial_name": null,
|
| 449 |
+
"trial_params": null
|
| 450 |
+
}
|
data_curation&frozen_layers/dataset1_all/F0/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1e2d0e3aa5d2da2620967534f5f2e2c2256b221f18f54b08d3a18b5b4ce4645e
|
| 3 |
+
size 82168378
|
data_curation&frozen_layers/dataset3_2048/F2/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a75e0236ae48e131d17c55edac95710df7cda4d1d3ae117d50c03e177291089f
|
| 3 |
+
size 15006
|
data_curation&frozen_layers/dataset3_2048/F2/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b5b52d243330319c0af9e8ee37eec3cd97862d9821d4bac5ea3efa228ba7e3a4
|
| 3 |
+
size 1064
|
data_curation&frozen_layers/dataset3_2048/F4/predictions.pickle
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:092e9fa3ba8aec8b46c4e4554cfbd6a55d1b7c2b48af1268d0a19547360a5793
|
| 3 |
+
size 78067
|
data_curation&frozen_layers/dataset3_2048/F4/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:495394a542cc7656b4fd829b32b9dab7bc9f591b59ad7220ce156679e0002696
|
| 3 |
+
size 4280
|
scaling_performance/.DS_Store
CHANGED
|
Binary files a/scaling_performance/.DS_Store and b/scaling_performance/.DS_Store differ
|
|
|
scaling_performance/1000/.DS_Store
CHANGED
|
Binary files a/scaling_performance/1000/.DS_Store and b/scaling_performance/1000/.DS_Store differ
|
|
|
scaling_performance/1000/L2/.DS_Store
CHANGED
|
Binary files a/scaling_performance/1000/L2/.DS_Store and b/scaling_performance/1000/L2/.DS_Store differ
|
|
|
scaling_performance/1000/L2/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.545,
|
| 3 |
+
"test_loss": 0.6915959715843201,
|
| 4 |
+
"test_macro_f1": 0.3625442191166684,
|
| 5 |
+
"test_runtime": 0.4924,
|
| 6 |
+
"test_samples_per_second": 406.157,
|
| 7 |
+
"test_steps_per_second": 4.062
|
| 8 |
+
}
|
scaling_performance/1000/L2/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.02,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.02,
|
| 9 |
+
"hidden_size": 256,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 512,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 2048,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 4,
|
| 16 |
+
"num_hidden_layers": 2,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"problem_type": "single_label_classification",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.28.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
scaling_performance/1000/L2/eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.545,
|
| 3 |
+
"test_loss": 0.6915959715843201,
|
| 4 |
+
"test_macro_f1": 0.3625442191166684,
|
| 5 |
+
"test_runtime": 0.4924,
|
| 6 |
+
"test_samples_per_second": 406.157,
|
| 7 |
+
"test_steps_per_second": 4.062
|
| 8 |
+
}
|
scaling_performance/1000/L2/trainer_state.json
ADDED
|
@@ -0,0 +1,712 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6915959715843201,
|
| 3 |
+
"best_model_checkpoint": "/vsphhome/fengguoqing/Geneformer/models/5folds_allmodels/1000samples/L2/fold4/checkpoint-80",
|
| 4 |
+
"epoch": 12.0,
|
| 5 |
+
"global_step": 96,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 1.0000000000000001e-07,
|
| 13 |
+
"loss": 0.7159,
|
| 14 |
+
"step": 1
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.25,
|
| 18 |
+
"learning_rate": 2.0000000000000002e-07,
|
| 19 |
+
"loss": 0.701,
|
| 20 |
+
"step": 2
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.38,
|
| 24 |
+
"learning_rate": 3.0000000000000004e-07,
|
| 25 |
+
"loss": 0.7445,
|
| 26 |
+
"step": 3
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.5,
|
| 30 |
+
"learning_rate": 4.0000000000000003e-07,
|
| 31 |
+
"loss": 0.7041,
|
| 32 |
+
"step": 4
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.62,
|
| 36 |
+
"learning_rate": 5.000000000000001e-07,
|
| 37 |
+
"loss": 0.6917,
|
| 38 |
+
"step": 5
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.75,
|
| 42 |
+
"learning_rate": 6.000000000000001e-07,
|
| 43 |
+
"loss": 0.7,
|
| 44 |
+
"step": 6
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.88,
|
| 48 |
+
"learning_rate": 7.000000000000001e-07,
|
| 49 |
+
"loss": 0.6976,
|
| 50 |
+
"step": 7
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"learning_rate": 8.000000000000001e-07,
|
| 55 |
+
"loss": 0.731,
|
| 56 |
+
"step": 8
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.0,
|
| 60 |
+
"eval_accuracy": 0.45,
|
| 61 |
+
"eval_loss": 0.7109375,
|
| 62 |
+
"eval_macro_f1": 0.3103448275862069,
|
| 63 |
+
"eval_runtime": 0.5076,
|
| 64 |
+
"eval_samples_per_second": 393.99,
|
| 65 |
+
"eval_steps_per_second": 3.94,
|
| 66 |
+
"step": 8
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.12,
|
| 70 |
+
"learning_rate": 9e-07,
|
| 71 |
+
"loss": 0.7144,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.25,
|
| 76 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 77 |
+
"loss": 0.7244,
|
| 78 |
+
"step": 10
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.38,
|
| 82 |
+
"learning_rate": 1.1e-06,
|
| 83 |
+
"loss": 0.7184,
|
| 84 |
+
"step": 11
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5,
|
| 88 |
+
"learning_rate": 1.2000000000000002e-06,
|
| 89 |
+
"loss": 0.7062,
|
| 90 |
+
"step": 12
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.62,
|
| 94 |
+
"learning_rate": 1.3e-06,
|
| 95 |
+
"loss": 0.6956,
|
| 96 |
+
"step": 13
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.75,
|
| 100 |
+
"learning_rate": 1.4000000000000001e-06,
|
| 101 |
+
"loss": 0.7072,
|
| 102 |
+
"step": 14
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.88,
|
| 106 |
+
"learning_rate": 1.5e-06,
|
| 107 |
+
"loss": 0.6984,
|
| 108 |
+
"step": 15
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 2.0,
|
| 112 |
+
"learning_rate": 1.6000000000000001e-06,
|
| 113 |
+
"loss": 0.7097,
|
| 114 |
+
"step": 16
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 2.0,
|
| 118 |
+
"eval_accuracy": 0.455,
|
| 119 |
+
"eval_loss": 0.7094874382019043,
|
| 120 |
+
"eval_macro_f1": 0.32780364466097256,
|
| 121 |
+
"eval_runtime": 0.499,
|
| 122 |
+
"eval_samples_per_second": 400.839,
|
| 123 |
+
"eval_steps_per_second": 4.008,
|
| 124 |
+
"step": 16
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.12,
|
| 128 |
+
"learning_rate": 1.7000000000000002e-06,
|
| 129 |
+
"loss": 0.7078,
|
| 130 |
+
"step": 17
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.25,
|
| 134 |
+
"learning_rate": 1.8e-06,
|
| 135 |
+
"loss": 0.7055,
|
| 136 |
+
"step": 18
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.38,
|
| 140 |
+
"learning_rate": 1.9e-06,
|
| 141 |
+
"loss": 0.698,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.5,
|
| 146 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 147 |
+
"loss": 0.715,
|
| 148 |
+
"step": 20
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.62,
|
| 152 |
+
"learning_rate": 2.1000000000000002e-06,
|
| 153 |
+
"loss": 0.7083,
|
| 154 |
+
"step": 21
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.75,
|
| 158 |
+
"learning_rate": 2.2e-06,
|
| 159 |
+
"loss": 0.7027,
|
| 160 |
+
"step": 22
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.88,
|
| 164 |
+
"learning_rate": 2.3e-06,
|
| 165 |
+
"loss": 0.7082,
|
| 166 |
+
"step": 23
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.0,
|
| 170 |
+
"learning_rate": 2.4000000000000003e-06,
|
| 171 |
+
"loss": 0.7034,
|
| 172 |
+
"step": 24
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.455,
|
| 177 |
+
"eval_loss": 0.7071405649185181,
|
| 178 |
+
"eval_macro_f1": 0.32780364466097256,
|
| 179 |
+
"eval_runtime": 0.5786,
|
| 180 |
+
"eval_samples_per_second": 345.685,
|
| 181 |
+
"eval_steps_per_second": 3.457,
|
| 182 |
+
"step": 24
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.12,
|
| 186 |
+
"learning_rate": 2.5e-06,
|
| 187 |
+
"loss": 0.7044,
|
| 188 |
+
"step": 25
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.25,
|
| 192 |
+
"learning_rate": 2.6e-06,
|
| 193 |
+
"loss": 0.7111,
|
| 194 |
+
"step": 26
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.38,
|
| 198 |
+
"learning_rate": 2.7e-06,
|
| 199 |
+
"loss": 0.6988,
|
| 200 |
+
"step": 27
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.5,
|
| 204 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 205 |
+
"loss": 0.6876,
|
| 206 |
+
"step": 28
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.62,
|
| 210 |
+
"learning_rate": 2.9e-06,
|
| 211 |
+
"loss": 0.7027,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.75,
|
| 216 |
+
"learning_rate": 3e-06,
|
| 217 |
+
"loss": 0.7028,
|
| 218 |
+
"step": 30
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.88,
|
| 222 |
+
"learning_rate": 3.1e-06,
|
| 223 |
+
"loss": 0.7086,
|
| 224 |
+
"step": 31
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.0,
|
| 228 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 229 |
+
"loss": 0.7079,
|
| 230 |
+
"step": 32
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.0,
|
| 234 |
+
"eval_accuracy": 0.44,
|
| 235 |
+
"eval_loss": 0.7041958570480347,
|
| 236 |
+
"eval_macro_f1": 0.3333333333333333,
|
| 237 |
+
"eval_runtime": 0.5218,
|
| 238 |
+
"eval_samples_per_second": 383.289,
|
| 239 |
+
"eval_steps_per_second": 3.833,
|
| 240 |
+
"step": 32
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.12,
|
| 244 |
+
"learning_rate": 3.3e-06,
|
| 245 |
+
"loss": 0.7005,
|
| 246 |
+
"step": 33
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.25,
|
| 250 |
+
"learning_rate": 3.4000000000000005e-06,
|
| 251 |
+
"loss": 0.6968,
|
| 252 |
+
"step": 34
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.38,
|
| 256 |
+
"learning_rate": 3.5000000000000004e-06,
|
| 257 |
+
"loss": 0.7009,
|
| 258 |
+
"step": 35
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.5,
|
| 262 |
+
"learning_rate": 3.6e-06,
|
| 263 |
+
"loss": 0.7043,
|
| 264 |
+
"step": 36
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.62,
|
| 268 |
+
"learning_rate": 3.7e-06,
|
| 269 |
+
"loss": 0.6945,
|
| 270 |
+
"step": 37
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 4.75,
|
| 274 |
+
"learning_rate": 3.8e-06,
|
| 275 |
+
"loss": 0.6995,
|
| 276 |
+
"step": 38
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 4.88,
|
| 280 |
+
"learning_rate": 3.9e-06,
|
| 281 |
+
"loss": 0.693,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 5.0,
|
| 286 |
+
"learning_rate": 4.000000000000001e-06,
|
| 287 |
+
"loss": 0.7036,
|
| 288 |
+
"step": 40
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.0,
|
| 292 |
+
"eval_accuracy": 0.46,
|
| 293 |
+
"eval_loss": 0.7008482217788696,
|
| 294 |
+
"eval_macro_f1": 0.3846153846153846,
|
| 295 |
+
"eval_runtime": 0.6101,
|
| 296 |
+
"eval_samples_per_second": 327.803,
|
| 297 |
+
"eval_steps_per_second": 3.278,
|
| 298 |
+
"step": 40
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.12,
|
| 302 |
+
"learning_rate": 4.1000000000000006e-06,
|
| 303 |
+
"loss": 0.6946,
|
| 304 |
+
"step": 41
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.25,
|
| 308 |
+
"learning_rate": 4.2000000000000004e-06,
|
| 309 |
+
"loss": 0.6955,
|
| 310 |
+
"step": 42
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.38,
|
| 314 |
+
"learning_rate": 4.2999999999999995e-06,
|
| 315 |
+
"loss": 0.697,
|
| 316 |
+
"step": 43
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.5,
|
| 320 |
+
"learning_rate": 4.4e-06,
|
| 321 |
+
"loss": 0.7048,
|
| 322 |
+
"step": 44
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 5.62,
|
| 326 |
+
"learning_rate": 4.5e-06,
|
| 327 |
+
"loss": 0.6956,
|
| 328 |
+
"step": 45
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.75,
|
| 332 |
+
"learning_rate": 4.6e-06,
|
| 333 |
+
"loss": 0.6956,
|
| 334 |
+
"step": 46
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 5.88,
|
| 338 |
+
"learning_rate": 4.7e-06,
|
| 339 |
+
"loss": 0.691,
|
| 340 |
+
"step": 47
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 6.0,
|
| 344 |
+
"learning_rate": 4.800000000000001e-06,
|
| 345 |
+
"loss": 0.6909,
|
| 346 |
+
"step": 48
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.0,
|
| 350 |
+
"eval_accuracy": 0.46,
|
| 351 |
+
"eval_loss": 0.6975666880607605,
|
| 352 |
+
"eval_macro_f1": 0.4375,
|
| 353 |
+
"eval_runtime": 0.5216,
|
| 354 |
+
"eval_samples_per_second": 383.451,
|
| 355 |
+
"eval_steps_per_second": 3.835,
|
| 356 |
+
"step": 48
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.12,
|
| 360 |
+
"learning_rate": 4.9000000000000005e-06,
|
| 361 |
+
"loss": 0.6932,
|
| 362 |
+
"step": 49
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.25,
|
| 366 |
+
"learning_rate": 5e-06,
|
| 367 |
+
"loss": 0.6907,
|
| 368 |
+
"step": 50
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.38,
|
| 372 |
+
"learning_rate": 5.1e-06,
|
| 373 |
+
"loss": 0.6925,
|
| 374 |
+
"step": 51
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.5,
|
| 378 |
+
"learning_rate": 5.2e-06,
|
| 379 |
+
"loss": 0.6979,
|
| 380 |
+
"step": 52
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.62,
|
| 384 |
+
"learning_rate": 5.3e-06,
|
| 385 |
+
"loss": 0.6935,
|
| 386 |
+
"step": 53
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 6.75,
|
| 390 |
+
"learning_rate": 5.4e-06,
|
| 391 |
+
"loss": 0.689,
|
| 392 |
+
"step": 54
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 6.88,
|
| 396 |
+
"learning_rate": 5.500000000000001e-06,
|
| 397 |
+
"loss": 0.6883,
|
| 398 |
+
"step": 55
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.0,
|
| 402 |
+
"learning_rate": 5.600000000000001e-06,
|
| 403 |
+
"loss": 0.6878,
|
| 404 |
+
"step": 56
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.0,
|
| 408 |
+
"eval_accuracy": 0.48,
|
| 409 |
+
"eval_loss": 0.6947869062423706,
|
| 410 |
+
"eval_macro_f1": 0.47743945332127424,
|
| 411 |
+
"eval_runtime": 0.5746,
|
| 412 |
+
"eval_samples_per_second": 348.056,
|
| 413 |
+
"eval_steps_per_second": 3.481,
|
| 414 |
+
"step": 56
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.12,
|
| 418 |
+
"learning_rate": 5.7000000000000005e-06,
|
| 419 |
+
"loss": 0.6902,
|
| 420 |
+
"step": 57
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.25,
|
| 424 |
+
"learning_rate": 5.8e-06,
|
| 425 |
+
"loss": 0.6918,
|
| 426 |
+
"step": 58
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.38,
|
| 430 |
+
"learning_rate": 5.9e-06,
|
| 431 |
+
"loss": 0.6876,
|
| 432 |
+
"step": 59
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.5,
|
| 436 |
+
"learning_rate": 6e-06,
|
| 437 |
+
"loss": 0.6932,
|
| 438 |
+
"step": 60
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 7.62,
|
| 442 |
+
"learning_rate": 6.1e-06,
|
| 443 |
+
"loss": 0.6856,
|
| 444 |
+
"step": 61
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.75,
|
| 448 |
+
"learning_rate": 6.2e-06,
|
| 449 |
+
"loss": 0.6851,
|
| 450 |
+
"step": 62
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 7.88,
|
| 454 |
+
"learning_rate": 6.300000000000001e-06,
|
| 455 |
+
"loss": 0.6837,
|
| 456 |
+
"step": 63
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 8.0,
|
| 460 |
+
"learning_rate": 6.4000000000000006e-06,
|
| 461 |
+
"loss": 0.687,
|
| 462 |
+
"step": 64
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.0,
|
| 466 |
+
"eval_accuracy": 0.54,
|
| 467 |
+
"eval_loss": 0.6931593418121338,
|
| 468 |
+
"eval_macro_f1": 0.4837840870833801,
|
| 469 |
+
"eval_runtime": 0.5564,
|
| 470 |
+
"eval_samples_per_second": 359.445,
|
| 471 |
+
"eval_steps_per_second": 3.594,
|
| 472 |
+
"step": 64
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.12,
|
| 476 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 477 |
+
"loss": 0.6901,
|
| 478 |
+
"step": 65
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 8.25,
|
| 482 |
+
"learning_rate": 6.6e-06,
|
| 483 |
+
"loss": 0.6886,
|
| 484 |
+
"step": 66
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 8.38,
|
| 488 |
+
"learning_rate": 6.700000000000001e-06,
|
| 489 |
+
"loss": 0.6802,
|
| 490 |
+
"step": 67
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.5,
|
| 494 |
+
"learning_rate": 6.800000000000001e-06,
|
| 495 |
+
"loss": 0.679,
|
| 496 |
+
"step": 68
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.62,
|
| 500 |
+
"learning_rate": 6.900000000000001e-06,
|
| 501 |
+
"loss": 0.6916,
|
| 502 |
+
"step": 69
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 8.75,
|
| 506 |
+
"learning_rate": 7.000000000000001e-06,
|
| 507 |
+
"loss": 0.6878,
|
| 508 |
+
"step": 70
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 8.88,
|
| 512 |
+
"learning_rate": 7.1e-06,
|
| 513 |
+
"loss": 0.684,
|
| 514 |
+
"step": 71
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 9.0,
|
| 518 |
+
"learning_rate": 7.2e-06,
|
| 519 |
+
"loss": 0.6892,
|
| 520 |
+
"step": 72
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 9.0,
|
| 524 |
+
"eval_accuracy": 0.55,
|
| 525 |
+
"eval_loss": 0.6919929385185242,
|
| 526 |
+
"eval_macro_f1": 0.39999999999999997,
|
| 527 |
+
"eval_runtime": 0.5264,
|
| 528 |
+
"eval_samples_per_second": 379.923,
|
| 529 |
+
"eval_steps_per_second": 3.799,
|
| 530 |
+
"step": 72
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 9.12,
|
| 534 |
+
"learning_rate": 7.2999999999999996e-06,
|
| 535 |
+
"loss": 0.6799,
|
| 536 |
+
"step": 73
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.25,
|
| 540 |
+
"learning_rate": 7.4e-06,
|
| 541 |
+
"loss": 0.6864,
|
| 542 |
+
"step": 74
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 9.38,
|
| 546 |
+
"learning_rate": 7.5e-06,
|
| 547 |
+
"loss": 0.682,
|
| 548 |
+
"step": 75
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 9.5,
|
| 552 |
+
"learning_rate": 7.6e-06,
|
| 553 |
+
"loss": 0.6829,
|
| 554 |
+
"step": 76
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 9.62,
|
| 558 |
+
"learning_rate": 7.7e-06,
|
| 559 |
+
"loss": 0.6891,
|
| 560 |
+
"step": 77
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 9.75,
|
| 564 |
+
"learning_rate": 7.8e-06,
|
| 565 |
+
"loss": 0.6714,
|
| 566 |
+
"step": 78
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 9.88,
|
| 570 |
+
"learning_rate": 7.9e-06,
|
| 571 |
+
"loss": 0.6854,
|
| 572 |
+
"step": 79
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 10.0,
|
| 576 |
+
"learning_rate": 8.000000000000001e-06,
|
| 577 |
+
"loss": 0.6885,
|
| 578 |
+
"step": 80
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.0,
|
| 582 |
+
"eval_accuracy": 0.545,
|
| 583 |
+
"eval_loss": 0.6915959715843201,
|
| 584 |
+
"eval_macro_f1": 0.3625442191166684,
|
| 585 |
+
"eval_runtime": 0.5616,
|
| 586 |
+
"eval_samples_per_second": 356.111,
|
| 587 |
+
"eval_steps_per_second": 3.561,
|
| 588 |
+
"step": 80
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 10.12,
|
| 592 |
+
"learning_rate": 8.1e-06,
|
| 593 |
+
"loss": 0.6902,
|
| 594 |
+
"step": 81
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 10.25,
|
| 598 |
+
"learning_rate": 8.200000000000001e-06,
|
| 599 |
+
"loss": 0.68,
|
| 600 |
+
"step": 82
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 10.38,
|
| 604 |
+
"learning_rate": 8.3e-06,
|
| 605 |
+
"loss": 0.6864,
|
| 606 |
+
"step": 83
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 10.5,
|
| 610 |
+
"learning_rate": 8.400000000000001e-06,
|
| 611 |
+
"loss": 0.687,
|
| 612 |
+
"step": 84
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 10.62,
|
| 616 |
+
"learning_rate": 8.500000000000002e-06,
|
| 617 |
+
"loss": 0.681,
|
| 618 |
+
"step": 85
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 10.75,
|
| 622 |
+
"learning_rate": 8.599999999999999e-06,
|
| 623 |
+
"loss": 0.6713,
|
| 624 |
+
"step": 86
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.88,
|
| 628 |
+
"learning_rate": 8.7e-06,
|
| 629 |
+
"loss": 0.6792,
|
| 630 |
+
"step": 87
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 11.0,
|
| 634 |
+
"learning_rate": 8.8e-06,
|
| 635 |
+
"loss": 0.6705,
|
| 636 |
+
"step": 88
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 11.0,
|
| 640 |
+
"eval_accuracy": 0.54,
|
| 641 |
+
"eval_loss": 0.6916151642799377,
|
| 642 |
+
"eval_macro_f1": 0.35064935064935066,
|
| 643 |
+
"eval_runtime": 0.4888,
|
| 644 |
+
"eval_samples_per_second": 409.169,
|
| 645 |
+
"eval_steps_per_second": 4.092,
|
| 646 |
+
"step": 88
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 11.12,
|
| 650 |
+
"learning_rate": 8.9e-06,
|
| 651 |
+
"loss": 0.6946,
|
| 652 |
+
"step": 89
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 11.25,
|
| 656 |
+
"learning_rate": 9e-06,
|
| 657 |
+
"loss": 0.6807,
|
| 658 |
+
"step": 90
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 11.38,
|
| 662 |
+
"learning_rate": 9.100000000000001e-06,
|
| 663 |
+
"loss": 0.6819,
|
| 664 |
+
"step": 91
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 11.5,
|
| 668 |
+
"learning_rate": 9.2e-06,
|
| 669 |
+
"loss": 0.6822,
|
| 670 |
+
"step": 92
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 11.62,
|
| 674 |
+
"learning_rate": 9.3e-06,
|
| 675 |
+
"loss": 0.6774,
|
| 676 |
+
"step": 93
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 11.75,
|
| 680 |
+
"learning_rate": 9.4e-06,
|
| 681 |
+
"loss": 0.6684,
|
| 682 |
+
"step": 94
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 11.88,
|
| 686 |
+
"learning_rate": 9.5e-06,
|
| 687 |
+
"loss": 0.6839,
|
| 688 |
+
"step": 95
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 12.0,
|
| 692 |
+
"learning_rate": 9.600000000000001e-06,
|
| 693 |
+
"loss": 0.6724,
|
| 694 |
+
"step": 96
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 12.0,
|
| 698 |
+
"eval_accuracy": 0.545,
|
| 699 |
+
"eval_loss": 0.691866934299469,
|
| 700 |
+
"eval_macro_f1": 0.35275080906148865,
|
| 701 |
+
"eval_runtime": 0.5673,
|
| 702 |
+
"eval_samples_per_second": 352.518,
|
| 703 |
+
"eval_steps_per_second": 3.525,
|
| 704 |
+
"step": 96
|
| 705 |
+
}
|
| 706 |
+
],
|
| 707 |
+
"max_steps": 400,
|
| 708 |
+
"num_train_epochs": 50,
|
| 709 |
+
"total_flos": 132241607884800.0,
|
| 710 |
+
"trial_name": null,
|
| 711 |
+
"trial_params": null
|
| 712 |
+
}
|
scaling_performance/1000/L4/.DS_Store
CHANGED
|
Binary files a/scaling_performance/1000/L4/.DS_Store and b/scaling_performance/1000/L4/.DS_Store differ
|
|
|
scaling_performance/1000/L4/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.545,
|
| 3 |
+
"test_loss": 0.6891449689865112,
|
| 4 |
+
"test_macro_f1": 0.35275080906148865,
|
| 5 |
+
"test_runtime": 0.5999,
|
| 6 |
+
"test_samples_per_second": 333.384,
|
| 7 |
+
"test_steps_per_second": 3.334
|
| 8 |
+
}
|
scaling_performance/1000/L4/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.02,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.02,
|
| 9 |
+
"hidden_size": 256,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 512,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 2048,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 4,
|
| 16 |
+
"num_hidden_layers": 4,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"problem_type": "single_label_classification",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.28.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
scaling_performance/1000/L4/eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.545,
|
| 3 |
+
"test_loss": 0.6891449689865112,
|
| 4 |
+
"test_macro_f1": 0.35275080906148865,
|
| 5 |
+
"test_runtime": 0.5999,
|
| 6 |
+
"test_samples_per_second": 333.384,
|
| 7 |
+
"test_steps_per_second": 3.334
|
| 8 |
+
}
|
scaling_performance/1000/L4/trainer_state.json
ADDED
|
@@ -0,0 +1,480 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6891449689865112,
|
| 3 |
+
"best_model_checkpoint": "/vsphhome/fengguoqing/Geneformer/models/5folds_allmodels/1000samples/L4/fold4/checkpoint-48",
|
| 4 |
+
"epoch": 8.0,
|
| 5 |
+
"global_step": 64,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 1.0000000000000001e-07,
|
| 13 |
+
"loss": 0.6931,
|
| 14 |
+
"step": 1
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.25,
|
| 18 |
+
"learning_rate": 2.0000000000000002e-07,
|
| 19 |
+
"loss": 0.6891,
|
| 20 |
+
"step": 2
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.38,
|
| 24 |
+
"learning_rate": 3.0000000000000004e-07,
|
| 25 |
+
"loss": 0.69,
|
| 26 |
+
"step": 3
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.5,
|
| 30 |
+
"learning_rate": 4.0000000000000003e-07,
|
| 31 |
+
"loss": 0.6921,
|
| 32 |
+
"step": 4
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.62,
|
| 36 |
+
"learning_rate": 5.000000000000001e-07,
|
| 37 |
+
"loss": 0.6961,
|
| 38 |
+
"step": 5
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.75,
|
| 42 |
+
"learning_rate": 6.000000000000001e-07,
|
| 43 |
+
"loss": 0.6871,
|
| 44 |
+
"step": 6
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.88,
|
| 48 |
+
"learning_rate": 7.000000000000001e-07,
|
| 49 |
+
"loss": 0.6956,
|
| 50 |
+
"step": 7
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"learning_rate": 8.000000000000001e-07,
|
| 55 |
+
"loss": 0.6891,
|
| 56 |
+
"step": 8
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.0,
|
| 60 |
+
"eval_accuracy": 0.535,
|
| 61 |
+
"eval_loss": 0.6912603974342346,
|
| 62 |
+
"eval_macro_f1": 0.4906760864207673,
|
| 63 |
+
"eval_runtime": 0.6002,
|
| 64 |
+
"eval_samples_per_second": 333.239,
|
| 65 |
+
"eval_steps_per_second": 3.332,
|
| 66 |
+
"step": 8
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.12,
|
| 70 |
+
"learning_rate": 9e-07,
|
| 71 |
+
"loss": 0.6912,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.25,
|
| 76 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 77 |
+
"loss": 0.6894,
|
| 78 |
+
"step": 10
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.38,
|
| 82 |
+
"learning_rate": 1.1e-06,
|
| 83 |
+
"loss": 0.6862,
|
| 84 |
+
"step": 11
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5,
|
| 88 |
+
"learning_rate": 1.2000000000000002e-06,
|
| 89 |
+
"loss": 0.6888,
|
| 90 |
+
"step": 12
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.62,
|
| 94 |
+
"learning_rate": 1.3e-06,
|
| 95 |
+
"loss": 0.6906,
|
| 96 |
+
"step": 13
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.75,
|
| 100 |
+
"learning_rate": 1.4000000000000001e-06,
|
| 101 |
+
"loss": 0.6974,
|
| 102 |
+
"step": 14
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.88,
|
| 106 |
+
"learning_rate": 1.5e-06,
|
| 107 |
+
"loss": 0.6964,
|
| 108 |
+
"step": 15
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 2.0,
|
| 112 |
+
"learning_rate": 1.6000000000000001e-06,
|
| 113 |
+
"loss": 0.689,
|
| 114 |
+
"step": 16
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 2.0,
|
| 118 |
+
"eval_accuracy": 0.555,
|
| 119 |
+
"eval_loss": 0.6907957196235657,
|
| 120 |
+
"eval_macro_f1": 0.5024458420684835,
|
| 121 |
+
"eval_runtime": 0.5971,
|
| 122 |
+
"eval_samples_per_second": 334.937,
|
| 123 |
+
"eval_steps_per_second": 3.349,
|
| 124 |
+
"step": 16
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.12,
|
| 128 |
+
"learning_rate": 1.7000000000000002e-06,
|
| 129 |
+
"loss": 0.6874,
|
| 130 |
+
"step": 17
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.25,
|
| 134 |
+
"learning_rate": 1.8e-06,
|
| 135 |
+
"loss": 0.6979,
|
| 136 |
+
"step": 18
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.38,
|
| 140 |
+
"learning_rate": 1.9e-06,
|
| 141 |
+
"loss": 0.6931,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.5,
|
| 146 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 147 |
+
"loss": 0.6823,
|
| 148 |
+
"step": 20
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.62,
|
| 152 |
+
"learning_rate": 2.1000000000000002e-06,
|
| 153 |
+
"loss": 0.6901,
|
| 154 |
+
"step": 21
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.75,
|
| 158 |
+
"learning_rate": 2.2e-06,
|
| 159 |
+
"loss": 0.6894,
|
| 160 |
+
"step": 22
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.88,
|
| 164 |
+
"learning_rate": 2.3e-06,
|
| 165 |
+
"loss": 0.6887,
|
| 166 |
+
"step": 23
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.0,
|
| 170 |
+
"learning_rate": 2.4000000000000003e-06,
|
| 171 |
+
"loss": 0.693,
|
| 172 |
+
"step": 24
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.565,
|
| 177 |
+
"eval_loss": 0.690175473690033,
|
| 178 |
+
"eval_macro_f1": 0.4634763035367396,
|
| 179 |
+
"eval_runtime": 0.5949,
|
| 180 |
+
"eval_samples_per_second": 336.173,
|
| 181 |
+
"eval_steps_per_second": 3.362,
|
| 182 |
+
"step": 24
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.12,
|
| 186 |
+
"learning_rate": 2.5e-06,
|
| 187 |
+
"loss": 0.6931,
|
| 188 |
+
"step": 25
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.25,
|
| 192 |
+
"learning_rate": 2.6e-06,
|
| 193 |
+
"loss": 0.6882,
|
| 194 |
+
"step": 26
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.38,
|
| 198 |
+
"learning_rate": 2.7e-06,
|
| 199 |
+
"loss": 0.694,
|
| 200 |
+
"step": 27
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.5,
|
| 204 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 205 |
+
"loss": 0.692,
|
| 206 |
+
"step": 28
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.62,
|
| 210 |
+
"learning_rate": 2.9e-06,
|
| 211 |
+
"loss": 0.6899,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.75,
|
| 216 |
+
"learning_rate": 3e-06,
|
| 217 |
+
"loss": 0.6909,
|
| 218 |
+
"step": 30
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.88,
|
| 222 |
+
"learning_rate": 3.1e-06,
|
| 223 |
+
"loss": 0.6849,
|
| 224 |
+
"step": 31
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.0,
|
| 228 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 229 |
+
"loss": 0.6811,
|
| 230 |
+
"step": 32
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.0,
|
| 234 |
+
"eval_accuracy": 0.565,
|
| 235 |
+
"eval_loss": 0.6895776987075806,
|
| 236 |
+
"eval_macro_f1": 0.4238219808602934,
|
| 237 |
+
"eval_runtime": 0.5921,
|
| 238 |
+
"eval_samples_per_second": 337.774,
|
| 239 |
+
"eval_steps_per_second": 3.378,
|
| 240 |
+
"step": 32
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.12,
|
| 244 |
+
"learning_rate": 3.3e-06,
|
| 245 |
+
"loss": 0.6888,
|
| 246 |
+
"step": 33
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.25,
|
| 250 |
+
"learning_rate": 3.4000000000000005e-06,
|
| 251 |
+
"loss": 0.6896,
|
| 252 |
+
"step": 34
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.38,
|
| 256 |
+
"learning_rate": 3.5000000000000004e-06,
|
| 257 |
+
"loss": 0.6842,
|
| 258 |
+
"step": 35
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.5,
|
| 262 |
+
"learning_rate": 3.6e-06,
|
| 263 |
+
"loss": 0.6798,
|
| 264 |
+
"step": 36
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.62,
|
| 268 |
+
"learning_rate": 3.7e-06,
|
| 269 |
+
"loss": 0.6918,
|
| 270 |
+
"step": 37
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 4.75,
|
| 274 |
+
"learning_rate": 3.8e-06,
|
| 275 |
+
"loss": 0.6909,
|
| 276 |
+
"step": 38
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 4.88,
|
| 280 |
+
"learning_rate": 3.9e-06,
|
| 281 |
+
"loss": 0.6875,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 5.0,
|
| 286 |
+
"learning_rate": 4.000000000000001e-06,
|
| 287 |
+
"loss": 0.6771,
|
| 288 |
+
"step": 40
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.0,
|
| 292 |
+
"eval_accuracy": 0.55,
|
| 293 |
+
"eval_loss": 0.6891635060310364,
|
| 294 |
+
"eval_macro_f1": 0.3647656691134952,
|
| 295 |
+
"eval_runtime": 0.6082,
|
| 296 |
+
"eval_samples_per_second": 328.828,
|
| 297 |
+
"eval_steps_per_second": 3.288,
|
| 298 |
+
"step": 40
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.12,
|
| 302 |
+
"learning_rate": 4.1000000000000006e-06,
|
| 303 |
+
"loss": 0.6995,
|
| 304 |
+
"step": 41
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.25,
|
| 308 |
+
"learning_rate": 4.2000000000000004e-06,
|
| 309 |
+
"loss": 0.6806,
|
| 310 |
+
"step": 42
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.38,
|
| 314 |
+
"learning_rate": 4.2999999999999995e-06,
|
| 315 |
+
"loss": 0.6824,
|
| 316 |
+
"step": 43
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.5,
|
| 320 |
+
"learning_rate": 4.4e-06,
|
| 321 |
+
"loss": 0.6884,
|
| 322 |
+
"step": 44
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 5.62,
|
| 326 |
+
"learning_rate": 4.5e-06,
|
| 327 |
+
"loss": 0.6792,
|
| 328 |
+
"step": 45
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.75,
|
| 332 |
+
"learning_rate": 4.6e-06,
|
| 333 |
+
"loss": 0.6931,
|
| 334 |
+
"step": 46
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 5.88,
|
| 338 |
+
"learning_rate": 4.7e-06,
|
| 339 |
+
"loss": 0.6867,
|
| 340 |
+
"step": 47
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 6.0,
|
| 344 |
+
"learning_rate": 4.800000000000001e-06,
|
| 345 |
+
"loss": 0.6819,
|
| 346 |
+
"step": 48
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.0,
|
| 350 |
+
"eval_accuracy": 0.545,
|
| 351 |
+
"eval_loss": 0.6891449689865112,
|
| 352 |
+
"eval_macro_f1": 0.35275080906148865,
|
| 353 |
+
"eval_runtime": 0.6013,
|
| 354 |
+
"eval_samples_per_second": 332.599,
|
| 355 |
+
"eval_steps_per_second": 3.326,
|
| 356 |
+
"step": 48
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.12,
|
| 360 |
+
"learning_rate": 4.9000000000000005e-06,
|
| 361 |
+
"loss": 0.6826,
|
| 362 |
+
"step": 49
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.25,
|
| 366 |
+
"learning_rate": 5e-06,
|
| 367 |
+
"loss": 0.6705,
|
| 368 |
+
"step": 50
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.38,
|
| 372 |
+
"learning_rate": 5.1e-06,
|
| 373 |
+
"loss": 0.6671,
|
| 374 |
+
"step": 51
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.5,
|
| 378 |
+
"learning_rate": 5.2e-06,
|
| 379 |
+
"loss": 0.7043,
|
| 380 |
+
"step": 52
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.62,
|
| 384 |
+
"learning_rate": 5.3e-06,
|
| 385 |
+
"loss": 0.6757,
|
| 386 |
+
"step": 53
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 6.75,
|
| 390 |
+
"learning_rate": 5.4e-06,
|
| 391 |
+
"loss": 0.6942,
|
| 392 |
+
"step": 54
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 6.88,
|
| 396 |
+
"learning_rate": 5.500000000000001e-06,
|
| 397 |
+
"loss": 0.6902,
|
| 398 |
+
"step": 55
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.0,
|
| 402 |
+
"learning_rate": 5.600000000000001e-06,
|
| 403 |
+
"loss": 0.7139,
|
| 404 |
+
"step": 56
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.0,
|
| 408 |
+
"eval_accuracy": 0.545,
|
| 409 |
+
"eval_loss": 0.6895077228546143,
|
| 410 |
+
"eval_macro_f1": 0.35275080906148865,
|
| 411 |
+
"eval_runtime": 0.5744,
|
| 412 |
+
"eval_samples_per_second": 348.173,
|
| 413 |
+
"eval_steps_per_second": 3.482,
|
| 414 |
+
"step": 56
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.12,
|
| 418 |
+
"learning_rate": 5.7000000000000005e-06,
|
| 419 |
+
"loss": 0.6857,
|
| 420 |
+
"step": 57
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.25,
|
| 424 |
+
"learning_rate": 5.8e-06,
|
| 425 |
+
"loss": 0.7153,
|
| 426 |
+
"step": 58
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.38,
|
| 430 |
+
"learning_rate": 5.9e-06,
|
| 431 |
+
"loss": 0.6892,
|
| 432 |
+
"step": 59
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.5,
|
| 436 |
+
"learning_rate": 6e-06,
|
| 437 |
+
"loss": 0.6819,
|
| 438 |
+
"step": 60
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 7.62,
|
| 442 |
+
"learning_rate": 6.1e-06,
|
| 443 |
+
"loss": 0.6698,
|
| 444 |
+
"step": 61
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.75,
|
| 448 |
+
"learning_rate": 6.2e-06,
|
| 449 |
+
"loss": 0.6628,
|
| 450 |
+
"step": 62
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 7.88,
|
| 454 |
+
"learning_rate": 6.300000000000001e-06,
|
| 455 |
+
"loss": 0.6882,
|
| 456 |
+
"step": 63
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 8.0,
|
| 460 |
+
"learning_rate": 6.4000000000000006e-06,
|
| 461 |
+
"loss": 0.6845,
|
| 462 |
+
"step": 64
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.0,
|
| 466 |
+
"eval_accuracy": 0.545,
|
| 467 |
+
"eval_loss": 0.6896821856498718,
|
| 468 |
+
"eval_macro_f1": 0.35275080906148865,
|
| 469 |
+
"eval_runtime": 0.5921,
|
| 470 |
+
"eval_samples_per_second": 337.784,
|
| 471 |
+
"eval_steps_per_second": 3.378,
|
| 472 |
+
"step": 64
|
| 473 |
+
}
|
| 474 |
+
],
|
| 475 |
+
"max_steps": 400,
|
| 476 |
+
"num_train_epochs": 50,
|
| 477 |
+
"total_flos": 171067362508800.0,
|
| 478 |
+
"trial_name": null,
|
| 479 |
+
"trial_params": null
|
| 480 |
+
}
|
scaling_performance/1000/fine-tuned/.DS_Store
CHANGED
|
Binary files a/scaling_performance/1000/fine-tuned/.DS_Store and b/scaling_performance/1000/fine-tuned/.DS_Store differ
|
|
|
scaling_performance/1000/fine-tuned/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.615,
|
| 3 |
+
"test_loss": 0.6587879657745361,
|
| 4 |
+
"test_macro_f1": 0.6042250263421655,
|
| 5 |
+
"test_runtime": 0.6844,
|
| 6 |
+
"test_samples_per_second": 292.222,
|
| 7 |
+
"test_steps_per_second": 2.922
|
| 8 |
+
}
|
scaling_performance/1000/fine-tuned/config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "/vsphhome/fengguoqing/Geneformer",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.02,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"gradient_checkpointing": false,
|
| 9 |
+
"hidden_act": "relu",
|
| 10 |
+
"hidden_dropout_prob": 0.02,
|
| 11 |
+
"hidden_size": 256,
|
| 12 |
+
"initializer_range": 0.02,
|
| 13 |
+
"intermediate_size": 512,
|
| 14 |
+
"layer_norm_eps": 1e-12,
|
| 15 |
+
"max_position_embeddings": 2048,
|
| 16 |
+
"model_type": "bert",
|
| 17 |
+
"num_attention_heads": 4,
|
| 18 |
+
"num_hidden_layers": 6,
|
| 19 |
+
"pad_token_id": 0,
|
| 20 |
+
"position_embedding_type": "absolute",
|
| 21 |
+
"problem_type": "single_label_classification",
|
| 22 |
+
"torch_dtype": "float32",
|
| 23 |
+
"transformers_version": "4.28.0",
|
| 24 |
+
"type_vocab_size": 2,
|
| 25 |
+
"use_cache": true,
|
| 26 |
+
"vocab_size": 25426
|
| 27 |
+
}
|
scaling_performance/1000/fine-tuned/eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.615,
|
| 3 |
+
"test_loss": 0.6587879657745361,
|
| 4 |
+
"test_macro_f1": 0.6042250263421655,
|
| 5 |
+
"test_runtime": 0.6844,
|
| 6 |
+
"test_samples_per_second": 292.222,
|
| 7 |
+
"test_steps_per_second": 2.922
|
| 8 |
+
}
|
scaling_performance/1000/fine-tuned/trainer_state.json
ADDED
|
@@ -0,0 +1,1756 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6587879657745361,
|
| 3 |
+
"best_model_checkpoint": "/vsphhome/fengguoqing/Geneformer/models/5folds_allmodels/1000samples/fine_tuned/fold0/checkpoint-224",
|
| 4 |
+
"epoch": 30.0,
|
| 5 |
+
"global_step": 240,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 1.0000000000000001e-07,
|
| 13 |
+
"loss": 0.6848,
|
| 14 |
+
"step": 1
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.25,
|
| 18 |
+
"learning_rate": 2.0000000000000002e-07,
|
| 19 |
+
"loss": 0.7017,
|
| 20 |
+
"step": 2
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.38,
|
| 24 |
+
"learning_rate": 3.0000000000000004e-07,
|
| 25 |
+
"loss": 0.7008,
|
| 26 |
+
"step": 3
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.5,
|
| 30 |
+
"learning_rate": 4.0000000000000003e-07,
|
| 31 |
+
"loss": 0.6882,
|
| 32 |
+
"step": 4
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.62,
|
| 36 |
+
"learning_rate": 5.000000000000001e-07,
|
| 37 |
+
"loss": 0.6897,
|
| 38 |
+
"step": 5
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.75,
|
| 42 |
+
"learning_rate": 6.000000000000001e-07,
|
| 43 |
+
"loss": 0.6837,
|
| 44 |
+
"step": 6
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.88,
|
| 48 |
+
"learning_rate": 7.000000000000001e-07,
|
| 49 |
+
"loss": 0.6949,
|
| 50 |
+
"step": 7
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"learning_rate": 8.000000000000001e-07,
|
| 55 |
+
"loss": 0.6921,
|
| 56 |
+
"step": 8
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.0,
|
| 60 |
+
"eval_accuracy": 0.455,
|
| 61 |
+
"eval_loss": 0.6985742449760437,
|
| 62 |
+
"eval_macro_f1": 0.4285864066472701,
|
| 63 |
+
"eval_runtime": 0.767,
|
| 64 |
+
"eval_samples_per_second": 260.748,
|
| 65 |
+
"eval_steps_per_second": 2.607,
|
| 66 |
+
"step": 8
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.12,
|
| 70 |
+
"learning_rate": 9e-07,
|
| 71 |
+
"loss": 0.687,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.25,
|
| 76 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 77 |
+
"loss": 0.6932,
|
| 78 |
+
"step": 10
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.38,
|
| 82 |
+
"learning_rate": 1.1e-06,
|
| 83 |
+
"loss": 0.6838,
|
| 84 |
+
"step": 11
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5,
|
| 88 |
+
"learning_rate": 1.2000000000000002e-06,
|
| 89 |
+
"loss": 0.6902,
|
| 90 |
+
"step": 12
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.62,
|
| 94 |
+
"learning_rate": 1.3e-06,
|
| 95 |
+
"loss": 0.6992,
|
| 96 |
+
"step": 13
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.75,
|
| 100 |
+
"learning_rate": 1.4000000000000001e-06,
|
| 101 |
+
"loss": 0.6965,
|
| 102 |
+
"step": 14
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.88,
|
| 106 |
+
"learning_rate": 1.5e-06,
|
| 107 |
+
"loss": 0.6888,
|
| 108 |
+
"step": 15
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 2.0,
|
| 112 |
+
"learning_rate": 1.6000000000000001e-06,
|
| 113 |
+
"loss": 0.6817,
|
| 114 |
+
"step": 16
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 2.0,
|
| 118 |
+
"eval_accuracy": 0.475,
|
| 119 |
+
"eval_loss": 0.6979620456695557,
|
| 120 |
+
"eval_macro_f1": 0.42495687176538244,
|
| 121 |
+
"eval_runtime": 0.7018,
|
| 122 |
+
"eval_samples_per_second": 284.975,
|
| 123 |
+
"eval_steps_per_second": 2.85,
|
| 124 |
+
"step": 16
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.12,
|
| 128 |
+
"learning_rate": 1.7000000000000002e-06,
|
| 129 |
+
"loss": 0.6905,
|
| 130 |
+
"step": 17
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.25,
|
| 134 |
+
"learning_rate": 1.8e-06,
|
| 135 |
+
"loss": 0.6945,
|
| 136 |
+
"step": 18
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.38,
|
| 140 |
+
"learning_rate": 1.9e-06,
|
| 141 |
+
"loss": 0.6895,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.5,
|
| 146 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 147 |
+
"loss": 0.6841,
|
| 148 |
+
"step": 20
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.62,
|
| 152 |
+
"learning_rate": 2.1000000000000002e-06,
|
| 153 |
+
"loss": 0.685,
|
| 154 |
+
"step": 21
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.75,
|
| 158 |
+
"learning_rate": 2.2e-06,
|
| 159 |
+
"loss": 0.6915,
|
| 160 |
+
"step": 22
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.88,
|
| 164 |
+
"learning_rate": 2.3e-06,
|
| 165 |
+
"loss": 0.6967,
|
| 166 |
+
"step": 23
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.0,
|
| 170 |
+
"learning_rate": 2.4000000000000003e-06,
|
| 171 |
+
"loss": 0.6912,
|
| 172 |
+
"step": 24
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.48,
|
| 177 |
+
"eval_loss": 0.6971560716629028,
|
| 178 |
+
"eval_macro_f1": 0.37492487077773773,
|
| 179 |
+
"eval_runtime": 0.7011,
|
| 180 |
+
"eval_samples_per_second": 285.285,
|
| 181 |
+
"eval_steps_per_second": 2.853,
|
| 182 |
+
"step": 24
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.12,
|
| 186 |
+
"learning_rate": 2.5e-06,
|
| 187 |
+
"loss": 0.684,
|
| 188 |
+
"step": 25
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.25,
|
| 192 |
+
"learning_rate": 2.6e-06,
|
| 193 |
+
"loss": 0.6841,
|
| 194 |
+
"step": 26
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.38,
|
| 198 |
+
"learning_rate": 2.7e-06,
|
| 199 |
+
"loss": 0.6929,
|
| 200 |
+
"step": 27
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.5,
|
| 204 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 205 |
+
"loss": 0.6904,
|
| 206 |
+
"step": 28
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.62,
|
| 210 |
+
"learning_rate": 2.9e-06,
|
| 211 |
+
"loss": 0.6837,
|
| 212 |
+
"step": 29
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.75,
|
| 216 |
+
"learning_rate": 3e-06,
|
| 217 |
+
"loss": 0.6952,
|
| 218 |
+
"step": 30
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.88,
|
| 222 |
+
"learning_rate": 3.1e-06,
|
| 223 |
+
"loss": 0.686,
|
| 224 |
+
"step": 31
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.0,
|
| 228 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 229 |
+
"loss": 0.6879,
|
| 230 |
+
"step": 32
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.0,
|
| 234 |
+
"eval_accuracy": 0.505,
|
| 235 |
+
"eval_loss": 0.6962434649467468,
|
| 236 |
+
"eval_macro_f1": 0.35275080906148865,
|
| 237 |
+
"eval_runtime": 0.6739,
|
| 238 |
+
"eval_samples_per_second": 296.767,
|
| 239 |
+
"eval_steps_per_second": 2.968,
|
| 240 |
+
"step": 32
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.12,
|
| 244 |
+
"learning_rate": 3.3e-06,
|
| 245 |
+
"loss": 0.6858,
|
| 246 |
+
"step": 33
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.25,
|
| 250 |
+
"learning_rate": 3.4000000000000005e-06,
|
| 251 |
+
"loss": 0.6956,
|
| 252 |
+
"step": 34
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.38,
|
| 256 |
+
"learning_rate": 3.5000000000000004e-06,
|
| 257 |
+
"loss": 0.678,
|
| 258 |
+
"step": 35
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.5,
|
| 262 |
+
"learning_rate": 3.6e-06,
|
| 263 |
+
"loss": 0.6862,
|
| 264 |
+
"step": 36
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.62,
|
| 268 |
+
"learning_rate": 3.7e-06,
|
| 269 |
+
"loss": 0.69,
|
| 270 |
+
"step": 37
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 4.75,
|
| 274 |
+
"learning_rate": 3.8e-06,
|
| 275 |
+
"loss": 0.6754,
|
| 276 |
+
"step": 38
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 4.88,
|
| 280 |
+
"learning_rate": 3.9e-06,
|
| 281 |
+
"loss": 0.6943,
|
| 282 |
+
"step": 39
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 5.0,
|
| 286 |
+
"learning_rate": 4.000000000000001e-06,
|
| 287 |
+
"loss": 0.6874,
|
| 288 |
+
"step": 40
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.0,
|
| 292 |
+
"eval_accuracy": 0.545,
|
| 293 |
+
"eval_loss": 0.6953359246253967,
|
| 294 |
+
"eval_macro_f1": 0.3625442191166684,
|
| 295 |
+
"eval_runtime": 0.682,
|
| 296 |
+
"eval_samples_per_second": 293.25,
|
| 297 |
+
"eval_steps_per_second": 2.933,
|
| 298 |
+
"step": 40
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.12,
|
| 302 |
+
"learning_rate": 4.1000000000000006e-06,
|
| 303 |
+
"loss": 0.6939,
|
| 304 |
+
"step": 41
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.25,
|
| 308 |
+
"learning_rate": 4.2000000000000004e-06,
|
| 309 |
+
"loss": 0.6865,
|
| 310 |
+
"step": 42
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.38,
|
| 314 |
+
"learning_rate": 4.2999999999999995e-06,
|
| 315 |
+
"loss": 0.6706,
|
| 316 |
+
"step": 43
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.5,
|
| 320 |
+
"learning_rate": 4.4e-06,
|
| 321 |
+
"loss": 0.6749,
|
| 322 |
+
"step": 44
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 5.62,
|
| 326 |
+
"learning_rate": 4.5e-06,
|
| 327 |
+
"loss": 0.6929,
|
| 328 |
+
"step": 45
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.75,
|
| 332 |
+
"learning_rate": 4.6e-06,
|
| 333 |
+
"loss": 0.684,
|
| 334 |
+
"step": 46
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 5.88,
|
| 338 |
+
"learning_rate": 4.7e-06,
|
| 339 |
+
"loss": 0.6802,
|
| 340 |
+
"step": 47
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 6.0,
|
| 344 |
+
"learning_rate": 4.800000000000001e-06,
|
| 345 |
+
"loss": 0.7163,
|
| 346 |
+
"step": 48
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.0,
|
| 350 |
+
"eval_accuracy": 0.555,
|
| 351 |
+
"eval_loss": 0.6945347785949707,
|
| 352 |
+
"eval_macro_f1": 0.3669760660051922,
|
| 353 |
+
"eval_runtime": 0.7337,
|
| 354 |
+
"eval_samples_per_second": 272.601,
|
| 355 |
+
"eval_steps_per_second": 2.726,
|
| 356 |
+
"step": 48
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.12,
|
| 360 |
+
"learning_rate": 4.9000000000000005e-06,
|
| 361 |
+
"loss": 0.6936,
|
| 362 |
+
"step": 49
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.25,
|
| 366 |
+
"learning_rate": 5e-06,
|
| 367 |
+
"loss": 0.6854,
|
| 368 |
+
"step": 50
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.38,
|
| 372 |
+
"learning_rate": 5.1e-06,
|
| 373 |
+
"loss": 0.6865,
|
| 374 |
+
"step": 51
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.5,
|
| 378 |
+
"learning_rate": 5.2e-06,
|
| 379 |
+
"loss": 0.6784,
|
| 380 |
+
"step": 52
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.62,
|
| 384 |
+
"learning_rate": 5.3e-06,
|
| 385 |
+
"loss": 0.681,
|
| 386 |
+
"step": 53
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 6.75,
|
| 390 |
+
"learning_rate": 5.4e-06,
|
| 391 |
+
"loss": 0.6769,
|
| 392 |
+
"step": 54
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 6.88,
|
| 396 |
+
"learning_rate": 5.500000000000001e-06,
|
| 397 |
+
"loss": 0.6971,
|
| 398 |
+
"step": 55
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.0,
|
| 402 |
+
"learning_rate": 5.600000000000001e-06,
|
| 403 |
+
"loss": 0.6755,
|
| 404 |
+
"step": 56
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.0,
|
| 408 |
+
"eval_accuracy": 0.55,
|
| 409 |
+
"eval_loss": 0.6938488483428955,
|
| 410 |
+
"eval_macro_f1": 0.3548387096774194,
|
| 411 |
+
"eval_runtime": 0.6956,
|
| 412 |
+
"eval_samples_per_second": 287.509,
|
| 413 |
+
"eval_steps_per_second": 2.875,
|
| 414 |
+
"step": 56
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.12,
|
| 418 |
+
"learning_rate": 5.7000000000000005e-06,
|
| 419 |
+
"loss": 0.6804,
|
| 420 |
+
"step": 57
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.25,
|
| 424 |
+
"learning_rate": 5.8e-06,
|
| 425 |
+
"loss": 0.6875,
|
| 426 |
+
"step": 58
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.38,
|
| 430 |
+
"learning_rate": 5.9e-06,
|
| 431 |
+
"loss": 0.6561,
|
| 432 |
+
"step": 59
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.5,
|
| 436 |
+
"learning_rate": 6e-06,
|
| 437 |
+
"loss": 0.6871,
|
| 438 |
+
"step": 60
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 7.62,
|
| 442 |
+
"learning_rate": 6.1e-06,
|
| 443 |
+
"loss": 0.7011,
|
| 444 |
+
"step": 61
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.75,
|
| 448 |
+
"learning_rate": 6.2e-06,
|
| 449 |
+
"loss": 0.697,
|
| 450 |
+
"step": 62
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 7.88,
|
| 454 |
+
"learning_rate": 6.300000000000001e-06,
|
| 455 |
+
"loss": 0.6829,
|
| 456 |
+
"step": 63
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 8.0,
|
| 460 |
+
"learning_rate": 6.4000000000000006e-06,
|
| 461 |
+
"loss": 0.6857,
|
| 462 |
+
"step": 64
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.0,
|
| 466 |
+
"eval_accuracy": 0.55,
|
| 467 |
+
"eval_loss": 0.6929141283035278,
|
| 468 |
+
"eval_macro_f1": 0.3548387096774194,
|
| 469 |
+
"eval_runtime": 0.679,
|
| 470 |
+
"eval_samples_per_second": 294.572,
|
| 471 |
+
"eval_steps_per_second": 2.946,
|
| 472 |
+
"step": 64
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.12,
|
| 476 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 477 |
+
"loss": 0.6794,
|
| 478 |
+
"step": 65
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 8.25,
|
| 482 |
+
"learning_rate": 6.6e-06,
|
| 483 |
+
"loss": 0.6768,
|
| 484 |
+
"step": 66
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 8.38,
|
| 488 |
+
"learning_rate": 6.700000000000001e-06,
|
| 489 |
+
"loss": 0.6813,
|
| 490 |
+
"step": 67
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.5,
|
| 494 |
+
"learning_rate": 6.800000000000001e-06,
|
| 495 |
+
"loss": 0.6926,
|
| 496 |
+
"step": 68
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.62,
|
| 500 |
+
"learning_rate": 6.900000000000001e-06,
|
| 501 |
+
"loss": 0.6782,
|
| 502 |
+
"step": 69
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 8.75,
|
| 506 |
+
"learning_rate": 7.000000000000001e-06,
|
| 507 |
+
"loss": 0.69,
|
| 508 |
+
"step": 70
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 8.88,
|
| 512 |
+
"learning_rate": 7.1e-06,
|
| 513 |
+
"loss": 0.6683,
|
| 514 |
+
"step": 71
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 9.0,
|
| 518 |
+
"learning_rate": 7.2e-06,
|
| 519 |
+
"loss": 0.703,
|
| 520 |
+
"step": 72
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 9.0,
|
| 524 |
+
"eval_accuracy": 0.555,
|
| 525 |
+
"eval_loss": 0.6919710040092468,
|
| 526 |
+
"eval_macro_f1": 0.3669760660051922,
|
| 527 |
+
"eval_runtime": 0.6934,
|
| 528 |
+
"eval_samples_per_second": 288.454,
|
| 529 |
+
"eval_steps_per_second": 2.885,
|
| 530 |
+
"step": 72
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 9.12,
|
| 534 |
+
"learning_rate": 7.2999999999999996e-06,
|
| 535 |
+
"loss": 0.7012,
|
| 536 |
+
"step": 73
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.25,
|
| 540 |
+
"learning_rate": 7.4e-06,
|
| 541 |
+
"loss": 0.6872,
|
| 542 |
+
"step": 74
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 9.38,
|
| 546 |
+
"learning_rate": 7.5e-06,
|
| 547 |
+
"loss": 0.6639,
|
| 548 |
+
"step": 75
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 9.5,
|
| 552 |
+
"learning_rate": 7.6e-06,
|
| 553 |
+
"loss": 0.692,
|
| 554 |
+
"step": 76
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 9.62,
|
| 558 |
+
"learning_rate": 7.7e-06,
|
| 559 |
+
"loss": 0.6915,
|
| 560 |
+
"step": 77
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 9.75,
|
| 564 |
+
"learning_rate": 7.8e-06,
|
| 565 |
+
"loss": 0.6845,
|
| 566 |
+
"step": 78
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 9.88,
|
| 570 |
+
"learning_rate": 7.9e-06,
|
| 571 |
+
"loss": 0.6628,
|
| 572 |
+
"step": 79
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 10.0,
|
| 576 |
+
"learning_rate": 8.000000000000001e-06,
|
| 577 |
+
"loss": 0.6792,
|
| 578 |
+
"step": 80
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.0,
|
| 582 |
+
"eval_accuracy": 0.555,
|
| 583 |
+
"eval_loss": 0.6907283067703247,
|
| 584 |
+
"eval_macro_f1": 0.3856773080241587,
|
| 585 |
+
"eval_runtime": 0.6816,
|
| 586 |
+
"eval_samples_per_second": 293.434,
|
| 587 |
+
"eval_steps_per_second": 2.934,
|
| 588 |
+
"step": 80
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 10.12,
|
| 592 |
+
"learning_rate": 8.1e-06,
|
| 593 |
+
"loss": 0.6768,
|
| 594 |
+
"step": 81
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 10.25,
|
| 598 |
+
"learning_rate": 8.200000000000001e-06,
|
| 599 |
+
"loss": 0.6778,
|
| 600 |
+
"step": 82
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 10.38,
|
| 604 |
+
"learning_rate": 8.3e-06,
|
| 605 |
+
"loss": 0.6719,
|
| 606 |
+
"step": 83
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 10.5,
|
| 610 |
+
"learning_rate": 8.400000000000001e-06,
|
| 611 |
+
"loss": 0.6957,
|
| 612 |
+
"step": 84
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 10.62,
|
| 616 |
+
"learning_rate": 8.500000000000002e-06,
|
| 617 |
+
"loss": 0.685,
|
| 618 |
+
"step": 85
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 10.75,
|
| 622 |
+
"learning_rate": 8.599999999999999e-06,
|
| 623 |
+
"loss": 0.6855,
|
| 624 |
+
"step": 86
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.88,
|
| 628 |
+
"learning_rate": 8.7e-06,
|
| 629 |
+
"loss": 0.6756,
|
| 630 |
+
"step": 87
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 11.0,
|
| 634 |
+
"learning_rate": 8.8e-06,
|
| 635 |
+
"loss": 0.6906,
|
| 636 |
+
"step": 88
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 11.0,
|
| 640 |
+
"eval_accuracy": 0.57,
|
| 641 |
+
"eval_loss": 0.6894752383232117,
|
| 642 |
+
"eval_macro_f1": 0.4188403838356535,
|
| 643 |
+
"eval_runtime": 0.689,
|
| 644 |
+
"eval_samples_per_second": 290.257,
|
| 645 |
+
"eval_steps_per_second": 2.903,
|
| 646 |
+
"step": 88
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 11.12,
|
| 650 |
+
"learning_rate": 8.9e-06,
|
| 651 |
+
"loss": 0.683,
|
| 652 |
+
"step": 89
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 11.25,
|
| 656 |
+
"learning_rate": 9e-06,
|
| 657 |
+
"loss": 0.675,
|
| 658 |
+
"step": 90
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 11.38,
|
| 662 |
+
"learning_rate": 9.100000000000001e-06,
|
| 663 |
+
"loss": 0.7001,
|
| 664 |
+
"step": 91
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 11.5,
|
| 668 |
+
"learning_rate": 9.2e-06,
|
| 669 |
+
"loss": 0.6714,
|
| 670 |
+
"step": 92
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 11.62,
|
| 674 |
+
"learning_rate": 9.3e-06,
|
| 675 |
+
"loss": 0.6846,
|
| 676 |
+
"step": 93
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 11.75,
|
| 680 |
+
"learning_rate": 9.4e-06,
|
| 681 |
+
"loss": 0.6772,
|
| 682 |
+
"step": 94
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 11.88,
|
| 686 |
+
"learning_rate": 9.5e-06,
|
| 687 |
+
"loss": 0.6588,
|
| 688 |
+
"step": 95
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 12.0,
|
| 692 |
+
"learning_rate": 9.600000000000001e-06,
|
| 693 |
+
"loss": 0.6864,
|
| 694 |
+
"step": 96
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 12.0,
|
| 698 |
+
"eval_accuracy": 0.56,
|
| 699 |
+
"eval_loss": 0.6882457733154297,
|
| 700 |
+
"eval_macro_f1": 0.43524579643178024,
|
| 701 |
+
"eval_runtime": 0.6737,
|
| 702 |
+
"eval_samples_per_second": 296.852,
|
| 703 |
+
"eval_steps_per_second": 2.969,
|
| 704 |
+
"step": 96
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 12.12,
|
| 708 |
+
"learning_rate": 9.7e-06,
|
| 709 |
+
"loss": 0.6651,
|
| 710 |
+
"step": 97
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 12.25,
|
| 714 |
+
"learning_rate": 9.800000000000001e-06,
|
| 715 |
+
"loss": 0.6887,
|
| 716 |
+
"step": 98
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 12.38,
|
| 720 |
+
"learning_rate": 9.900000000000002e-06,
|
| 721 |
+
"loss": 0.6746,
|
| 722 |
+
"step": 99
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 12.5,
|
| 726 |
+
"learning_rate": 1e-05,
|
| 727 |
+
"loss": 0.6694,
|
| 728 |
+
"step": 100
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 12.62,
|
| 732 |
+
"learning_rate": 1.0100000000000002e-05,
|
| 733 |
+
"loss": 0.6768,
|
| 734 |
+
"step": 101
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 12.75,
|
| 738 |
+
"learning_rate": 1.02e-05,
|
| 739 |
+
"loss": 0.6619,
|
| 740 |
+
"step": 102
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 12.88,
|
| 744 |
+
"learning_rate": 1.03e-05,
|
| 745 |
+
"loss": 0.7057,
|
| 746 |
+
"step": 103
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 13.0,
|
| 750 |
+
"learning_rate": 1.04e-05,
|
| 751 |
+
"loss": 0.6648,
|
| 752 |
+
"step": 104
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 13.0,
|
| 756 |
+
"eval_accuracy": 0.565,
|
| 757 |
+
"eval_loss": 0.6870580911636353,
|
| 758 |
+
"eval_macro_f1": 0.43825665859564167,
|
| 759 |
+
"eval_runtime": 0.7161,
|
| 760 |
+
"eval_samples_per_second": 279.296,
|
| 761 |
+
"eval_steps_per_second": 2.793,
|
| 762 |
+
"step": 104
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 13.12,
|
| 766 |
+
"learning_rate": 1.05e-05,
|
| 767 |
+
"loss": 0.6833,
|
| 768 |
+
"step": 105
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 13.25,
|
| 772 |
+
"learning_rate": 1.06e-05,
|
| 773 |
+
"loss": 0.6797,
|
| 774 |
+
"step": 106
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 13.38,
|
| 778 |
+
"learning_rate": 1.0700000000000001e-05,
|
| 779 |
+
"loss": 0.66,
|
| 780 |
+
"step": 107
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 13.5,
|
| 784 |
+
"learning_rate": 1.08e-05,
|
| 785 |
+
"loss": 0.6844,
|
| 786 |
+
"step": 108
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 13.62,
|
| 790 |
+
"learning_rate": 1.09e-05,
|
| 791 |
+
"loss": 0.6803,
|
| 792 |
+
"step": 109
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 13.75,
|
| 796 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 797 |
+
"loss": 0.6712,
|
| 798 |
+
"step": 110
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 13.88,
|
| 802 |
+
"learning_rate": 1.11e-05,
|
| 803 |
+
"loss": 0.6683,
|
| 804 |
+
"step": 111
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 14.0,
|
| 808 |
+
"learning_rate": 1.1200000000000001e-05,
|
| 809 |
+
"loss": 0.6821,
|
| 810 |
+
"step": 112
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 14.0,
|
| 814 |
+
"eval_accuracy": 0.555,
|
| 815 |
+
"eval_loss": 0.6860873699188232,
|
| 816 |
+
"eval_macro_f1": 0.4388221570667423,
|
| 817 |
+
"eval_runtime": 0.6892,
|
| 818 |
+
"eval_samples_per_second": 290.191,
|
| 819 |
+
"eval_steps_per_second": 2.902,
|
| 820 |
+
"step": 112
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 14.12,
|
| 824 |
+
"learning_rate": 1.13e-05,
|
| 825 |
+
"loss": 0.6755,
|
| 826 |
+
"step": 113
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 14.25,
|
| 830 |
+
"learning_rate": 1.1400000000000001e-05,
|
| 831 |
+
"loss": 0.6666,
|
| 832 |
+
"step": 114
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 14.38,
|
| 836 |
+
"learning_rate": 1.1500000000000002e-05,
|
| 837 |
+
"loss": 0.6747,
|
| 838 |
+
"step": 115
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 14.5,
|
| 842 |
+
"learning_rate": 1.16e-05,
|
| 843 |
+
"loss": 0.6628,
|
| 844 |
+
"step": 116
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 14.62,
|
| 848 |
+
"learning_rate": 1.1700000000000001e-05,
|
| 849 |
+
"loss": 0.6699,
|
| 850 |
+
"step": 117
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 14.75,
|
| 854 |
+
"learning_rate": 1.18e-05,
|
| 855 |
+
"loss": 0.6717,
|
| 856 |
+
"step": 118
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 14.88,
|
| 860 |
+
"learning_rate": 1.19e-05,
|
| 861 |
+
"loss": 0.6764,
|
| 862 |
+
"step": 119
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 15.0,
|
| 866 |
+
"learning_rate": 1.2e-05,
|
| 867 |
+
"loss": 0.6721,
|
| 868 |
+
"step": 120
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 15.0,
|
| 872 |
+
"eval_accuracy": 0.555,
|
| 873 |
+
"eval_loss": 0.6848286390304565,
|
| 874 |
+
"eval_macro_f1": 0.4624143034037027,
|
| 875 |
+
"eval_runtime": 0.6968,
|
| 876 |
+
"eval_samples_per_second": 287.018,
|
| 877 |
+
"eval_steps_per_second": 2.87,
|
| 878 |
+
"step": 120
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 15.12,
|
| 882 |
+
"learning_rate": 1.2100000000000001e-05,
|
| 883 |
+
"loss": 0.6756,
|
| 884 |
+
"step": 121
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 15.25,
|
| 888 |
+
"learning_rate": 1.22e-05,
|
| 889 |
+
"loss": 0.6814,
|
| 890 |
+
"step": 122
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 15.38,
|
| 894 |
+
"learning_rate": 1.23e-05,
|
| 895 |
+
"loss": 0.6693,
|
| 896 |
+
"step": 123
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 15.5,
|
| 900 |
+
"learning_rate": 1.24e-05,
|
| 901 |
+
"loss": 0.6723,
|
| 902 |
+
"step": 124
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 15.62,
|
| 906 |
+
"learning_rate": 1.25e-05,
|
| 907 |
+
"loss": 0.6506,
|
| 908 |
+
"step": 125
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 15.75,
|
| 912 |
+
"learning_rate": 1.2600000000000001e-05,
|
| 913 |
+
"loss": 0.6818,
|
| 914 |
+
"step": 126
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 15.88,
|
| 918 |
+
"learning_rate": 1.27e-05,
|
| 919 |
+
"loss": 0.6681,
|
| 920 |
+
"step": 127
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 16.0,
|
| 924 |
+
"learning_rate": 1.2800000000000001e-05,
|
| 925 |
+
"loss": 0.676,
|
| 926 |
+
"step": 128
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 16.0,
|
| 930 |
+
"eval_accuracy": 0.54,
|
| 931 |
+
"eval_loss": 0.6835575699806213,
|
| 932 |
+
"eval_macro_f1": 0.4875222816399287,
|
| 933 |
+
"eval_runtime": 0.6965,
|
| 934 |
+
"eval_samples_per_second": 287.142,
|
| 935 |
+
"eval_steps_per_second": 2.871,
|
| 936 |
+
"step": 128
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 16.12,
|
| 940 |
+
"learning_rate": 1.29e-05,
|
| 941 |
+
"loss": 0.662,
|
| 942 |
+
"step": 129
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 16.25,
|
| 946 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 947 |
+
"loss": 0.6677,
|
| 948 |
+
"step": 130
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 16.38,
|
| 952 |
+
"learning_rate": 1.3100000000000002e-05,
|
| 953 |
+
"loss": 0.6641,
|
| 954 |
+
"step": 131
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 16.5,
|
| 958 |
+
"learning_rate": 1.32e-05,
|
| 959 |
+
"loss": 0.6832,
|
| 960 |
+
"step": 132
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 16.62,
|
| 964 |
+
"learning_rate": 1.3300000000000001e-05,
|
| 965 |
+
"loss": 0.6764,
|
| 966 |
+
"step": 133
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 16.75,
|
| 970 |
+
"learning_rate": 1.3400000000000002e-05,
|
| 971 |
+
"loss": 0.6583,
|
| 972 |
+
"step": 134
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 16.88,
|
| 976 |
+
"learning_rate": 1.3500000000000001e-05,
|
| 977 |
+
"loss": 0.659,
|
| 978 |
+
"step": 135
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 17.0,
|
| 982 |
+
"learning_rate": 1.3600000000000002e-05,
|
| 983 |
+
"loss": 0.6883,
|
| 984 |
+
"step": 136
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 17.0,
|
| 988 |
+
"eval_accuracy": 0.555,
|
| 989 |
+
"eval_loss": 0.6816694736480713,
|
| 990 |
+
"eval_macro_f1": 0.4676874308442238,
|
| 991 |
+
"eval_runtime": 0.6967,
|
| 992 |
+
"eval_samples_per_second": 287.071,
|
| 993 |
+
"eval_steps_per_second": 2.871,
|
| 994 |
+
"step": 136
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 17.12,
|
| 998 |
+
"learning_rate": 1.3700000000000001e-05,
|
| 999 |
+
"loss": 0.6548,
|
| 1000 |
+
"step": 137
|
| 1001 |
+
},
|
| 1002 |
+
{
|
| 1003 |
+
"epoch": 17.25,
|
| 1004 |
+
"learning_rate": 1.3800000000000002e-05,
|
| 1005 |
+
"loss": 0.6767,
|
| 1006 |
+
"step": 138
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 17.38,
|
| 1010 |
+
"learning_rate": 1.3900000000000002e-05,
|
| 1011 |
+
"loss": 0.6723,
|
| 1012 |
+
"step": 139
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 17.5,
|
| 1016 |
+
"learning_rate": 1.4000000000000001e-05,
|
| 1017 |
+
"loss": 0.6818,
|
| 1018 |
+
"step": 140
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 17.62,
|
| 1022 |
+
"learning_rate": 1.4099999999999999e-05,
|
| 1023 |
+
"loss": 0.6456,
|
| 1024 |
+
"step": 141
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 17.75,
|
| 1028 |
+
"learning_rate": 1.42e-05,
|
| 1029 |
+
"loss": 0.677,
|
| 1030 |
+
"step": 142
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 17.88,
|
| 1034 |
+
"learning_rate": 1.43e-05,
|
| 1035 |
+
"loss": 0.6573,
|
| 1036 |
+
"step": 143
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 18.0,
|
| 1040 |
+
"learning_rate": 1.44e-05,
|
| 1041 |
+
"loss": 0.6332,
|
| 1042 |
+
"step": 144
|
| 1043 |
+
},
|
| 1044 |
+
{
|
| 1045 |
+
"epoch": 18.0,
|
| 1046 |
+
"eval_accuracy": 0.57,
|
| 1047 |
+
"eval_loss": 0.6794130802154541,
|
| 1048 |
+
"eval_macro_f1": 0.5174503422735944,
|
| 1049 |
+
"eval_runtime": 0.692,
|
| 1050 |
+
"eval_samples_per_second": 289.012,
|
| 1051 |
+
"eval_steps_per_second": 2.89,
|
| 1052 |
+
"step": 144
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 18.12,
|
| 1056 |
+
"learning_rate": 1.45e-05,
|
| 1057 |
+
"loss": 0.6403,
|
| 1058 |
+
"step": 145
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 18.25,
|
| 1062 |
+
"learning_rate": 1.4599999999999999e-05,
|
| 1063 |
+
"loss": 0.6612,
|
| 1064 |
+
"step": 146
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 18.38,
|
| 1068 |
+
"learning_rate": 1.47e-05,
|
| 1069 |
+
"loss": 0.6482,
|
| 1070 |
+
"step": 147
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 18.5,
|
| 1074 |
+
"learning_rate": 1.48e-05,
|
| 1075 |
+
"loss": 0.6469,
|
| 1076 |
+
"step": 148
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 18.62,
|
| 1080 |
+
"learning_rate": 1.49e-05,
|
| 1081 |
+
"loss": 0.6756,
|
| 1082 |
+
"step": 149
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 18.75,
|
| 1086 |
+
"learning_rate": 1.5e-05,
|
| 1087 |
+
"loss": 0.6543,
|
| 1088 |
+
"step": 150
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 18.88,
|
| 1092 |
+
"learning_rate": 1.51e-05,
|
| 1093 |
+
"loss": 0.6734,
|
| 1094 |
+
"step": 151
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 19.0,
|
| 1098 |
+
"learning_rate": 1.52e-05,
|
| 1099 |
+
"loss": 0.6401,
|
| 1100 |
+
"step": 152
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 19.0,
|
| 1104 |
+
"eval_accuracy": 0.55,
|
| 1105 |
+
"eval_loss": 0.6782782077789307,
|
| 1106 |
+
"eval_macro_f1": 0.5146154675870995,
|
| 1107 |
+
"eval_runtime": 0.6724,
|
| 1108 |
+
"eval_samples_per_second": 297.433,
|
| 1109 |
+
"eval_steps_per_second": 2.974,
|
| 1110 |
+
"step": 152
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 19.12,
|
| 1114 |
+
"learning_rate": 1.53e-05,
|
| 1115 |
+
"loss": 0.6368,
|
| 1116 |
+
"step": 153
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 19.25,
|
| 1120 |
+
"learning_rate": 1.54e-05,
|
| 1121 |
+
"loss": 0.618,
|
| 1122 |
+
"step": 154
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 19.38,
|
| 1126 |
+
"learning_rate": 1.55e-05,
|
| 1127 |
+
"loss": 0.6665,
|
| 1128 |
+
"step": 155
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 19.5,
|
| 1132 |
+
"learning_rate": 1.56e-05,
|
| 1133 |
+
"loss": 0.6675,
|
| 1134 |
+
"step": 156
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 19.62,
|
| 1138 |
+
"learning_rate": 1.5700000000000002e-05,
|
| 1139 |
+
"loss": 0.6414,
|
| 1140 |
+
"step": 157
|
| 1141 |
+
},
|
| 1142 |
+
{
|
| 1143 |
+
"epoch": 19.75,
|
| 1144 |
+
"learning_rate": 1.58e-05,
|
| 1145 |
+
"loss": 0.6627,
|
| 1146 |
+
"step": 158
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 19.88,
|
| 1150 |
+
"learning_rate": 1.59e-05,
|
| 1151 |
+
"loss": 0.6375,
|
| 1152 |
+
"step": 159
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 20.0,
|
| 1156 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 1157 |
+
"loss": 0.6576,
|
| 1158 |
+
"step": 160
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 20.0,
|
| 1162 |
+
"eval_accuracy": 0.56,
|
| 1163 |
+
"eval_loss": 0.6768765449523926,
|
| 1164 |
+
"eval_macro_f1": 0.5331069609507639,
|
| 1165 |
+
"eval_runtime": 0.6907,
|
| 1166 |
+
"eval_samples_per_second": 289.541,
|
| 1167 |
+
"eval_steps_per_second": 2.895,
|
| 1168 |
+
"step": 160
|
| 1169 |
+
},
|
| 1170 |
+
{
|
| 1171 |
+
"epoch": 20.12,
|
| 1172 |
+
"learning_rate": 1.6100000000000002e-05,
|
| 1173 |
+
"loss": 0.6339,
|
| 1174 |
+
"step": 161
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 20.25,
|
| 1178 |
+
"learning_rate": 1.62e-05,
|
| 1179 |
+
"loss": 0.6623,
|
| 1180 |
+
"step": 162
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 20.38,
|
| 1184 |
+
"learning_rate": 1.63e-05,
|
| 1185 |
+
"loss": 0.6738,
|
| 1186 |
+
"step": 163
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 20.5,
|
| 1190 |
+
"learning_rate": 1.6400000000000002e-05,
|
| 1191 |
+
"loss": 0.6148,
|
| 1192 |
+
"step": 164
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 20.62,
|
| 1196 |
+
"learning_rate": 1.65e-05,
|
| 1197 |
+
"loss": 0.6195,
|
| 1198 |
+
"step": 165
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 20.75,
|
| 1202 |
+
"learning_rate": 1.66e-05,
|
| 1203 |
+
"loss": 0.675,
|
| 1204 |
+
"step": 166
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 20.88,
|
| 1208 |
+
"learning_rate": 1.6700000000000003e-05,
|
| 1209 |
+
"loss": 0.6345,
|
| 1210 |
+
"step": 167
|
| 1211 |
+
},
|
| 1212 |
+
{
|
| 1213 |
+
"epoch": 21.0,
|
| 1214 |
+
"learning_rate": 1.6800000000000002e-05,
|
| 1215 |
+
"loss": 0.6266,
|
| 1216 |
+
"step": 168
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 21.0,
|
| 1220 |
+
"eval_accuracy": 0.565,
|
| 1221 |
+
"eval_loss": 0.6750273108482361,
|
| 1222 |
+
"eval_macro_f1": 0.5372217346206016,
|
| 1223 |
+
"eval_runtime": 0.6888,
|
| 1224 |
+
"eval_samples_per_second": 290.355,
|
| 1225 |
+
"eval_steps_per_second": 2.904,
|
| 1226 |
+
"step": 168
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 21.12,
|
| 1230 |
+
"learning_rate": 1.69e-05,
|
| 1231 |
+
"loss": 0.622,
|
| 1232 |
+
"step": 169
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 21.25,
|
| 1236 |
+
"learning_rate": 1.7000000000000003e-05,
|
| 1237 |
+
"loss": 0.6214,
|
| 1238 |
+
"step": 170
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 21.38,
|
| 1242 |
+
"learning_rate": 1.7100000000000002e-05,
|
| 1243 |
+
"loss": 0.6917,
|
| 1244 |
+
"step": 171
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 21.5,
|
| 1248 |
+
"learning_rate": 1.7199999999999998e-05,
|
| 1249 |
+
"loss": 0.6134,
|
| 1250 |
+
"step": 172
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 21.62,
|
| 1254 |
+
"learning_rate": 1.73e-05,
|
| 1255 |
+
"loss": 0.6315,
|
| 1256 |
+
"step": 173
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 21.75,
|
| 1260 |
+
"learning_rate": 1.74e-05,
|
| 1261 |
+
"loss": 0.6232,
|
| 1262 |
+
"step": 174
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 21.88,
|
| 1266 |
+
"learning_rate": 1.75e-05,
|
| 1267 |
+
"loss": 0.6393,
|
| 1268 |
+
"step": 175
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 22.0,
|
| 1272 |
+
"learning_rate": 1.76e-05,
|
| 1273 |
+
"loss": 0.6437,
|
| 1274 |
+
"step": 176
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 22.0,
|
| 1278 |
+
"eval_accuracy": 0.575,
|
| 1279 |
+
"eval_loss": 0.677147626876831,
|
| 1280 |
+
"eval_macro_f1": 0.5523370638578011,
|
| 1281 |
+
"eval_runtime": 0.6959,
|
| 1282 |
+
"eval_samples_per_second": 287.392,
|
| 1283 |
+
"eval_steps_per_second": 2.874,
|
| 1284 |
+
"step": 176
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 22.12,
|
| 1288 |
+
"learning_rate": 1.77e-05,
|
| 1289 |
+
"loss": 0.5981,
|
| 1290 |
+
"step": 177
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 22.25,
|
| 1294 |
+
"learning_rate": 1.78e-05,
|
| 1295 |
+
"loss": 0.6213,
|
| 1296 |
+
"step": 178
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 22.38,
|
| 1300 |
+
"learning_rate": 1.79e-05,
|
| 1301 |
+
"loss": 0.6262,
|
| 1302 |
+
"step": 179
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 22.5,
|
| 1306 |
+
"learning_rate": 1.8e-05,
|
| 1307 |
+
"loss": 0.6094,
|
| 1308 |
+
"step": 180
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 22.62,
|
| 1312 |
+
"learning_rate": 1.81e-05,
|
| 1313 |
+
"loss": 0.6472,
|
| 1314 |
+
"step": 181
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 22.75,
|
| 1318 |
+
"learning_rate": 1.8200000000000002e-05,
|
| 1319 |
+
"loss": 0.6196,
|
| 1320 |
+
"step": 182
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 22.88,
|
| 1324 |
+
"learning_rate": 1.83e-05,
|
| 1325 |
+
"loss": 0.6156,
|
| 1326 |
+
"step": 183
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 23.0,
|
| 1330 |
+
"learning_rate": 1.84e-05,
|
| 1331 |
+
"loss": 0.6671,
|
| 1332 |
+
"step": 184
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 23.0,
|
| 1336 |
+
"eval_accuracy": 0.565,
|
| 1337 |
+
"eval_loss": 0.6740648150444031,
|
| 1338 |
+
"eval_macro_f1": 0.5294117647058824,
|
| 1339 |
+
"eval_runtime": 0.7084,
|
| 1340 |
+
"eval_samples_per_second": 282.312,
|
| 1341 |
+
"eval_steps_per_second": 2.823,
|
| 1342 |
+
"step": 184
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 23.12,
|
| 1346 |
+
"learning_rate": 1.85e-05,
|
| 1347 |
+
"loss": 0.6085,
|
| 1348 |
+
"step": 185
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 23.25,
|
| 1352 |
+
"learning_rate": 1.86e-05,
|
| 1353 |
+
"loss": 0.6269,
|
| 1354 |
+
"step": 186
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 23.38,
|
| 1358 |
+
"learning_rate": 1.87e-05,
|
| 1359 |
+
"loss": 0.6125,
|
| 1360 |
+
"step": 187
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 23.5,
|
| 1364 |
+
"learning_rate": 1.88e-05,
|
| 1365 |
+
"loss": 0.6173,
|
| 1366 |
+
"step": 188
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 23.62,
|
| 1370 |
+
"learning_rate": 1.8900000000000002e-05,
|
| 1371 |
+
"loss": 0.6425,
|
| 1372 |
+
"step": 189
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"epoch": 23.75,
|
| 1376 |
+
"learning_rate": 1.9e-05,
|
| 1377 |
+
"loss": 0.609,
|
| 1378 |
+
"step": 190
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 23.88,
|
| 1382 |
+
"learning_rate": 1.91e-05,
|
| 1383 |
+
"loss": 0.6031,
|
| 1384 |
+
"step": 191
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 24.0,
|
| 1388 |
+
"learning_rate": 1.9200000000000003e-05,
|
| 1389 |
+
"loss": 0.6229,
|
| 1390 |
+
"step": 192
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 24.0,
|
| 1394 |
+
"eval_accuracy": 0.585,
|
| 1395 |
+
"eval_loss": 0.6694185137748718,
|
| 1396 |
+
"eval_macro_f1": 0.5718891038039975,
|
| 1397 |
+
"eval_runtime": 0.695,
|
| 1398 |
+
"eval_samples_per_second": 287.787,
|
| 1399 |
+
"eval_steps_per_second": 2.878,
|
| 1400 |
+
"step": 192
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 24.12,
|
| 1404 |
+
"learning_rate": 1.93e-05,
|
| 1405 |
+
"loss": 0.612,
|
| 1406 |
+
"step": 193
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 24.25,
|
| 1410 |
+
"learning_rate": 1.94e-05,
|
| 1411 |
+
"loss": 0.5779,
|
| 1412 |
+
"step": 194
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 24.38,
|
| 1416 |
+
"learning_rate": 1.9500000000000003e-05,
|
| 1417 |
+
"loss": 0.6329,
|
| 1418 |
+
"step": 195
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 24.5,
|
| 1422 |
+
"learning_rate": 1.9600000000000002e-05,
|
| 1423 |
+
"loss": 0.6185,
|
| 1424 |
+
"step": 196
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 24.62,
|
| 1428 |
+
"learning_rate": 1.97e-05,
|
| 1429 |
+
"loss": 0.606,
|
| 1430 |
+
"step": 197
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 24.75,
|
| 1434 |
+
"learning_rate": 1.9800000000000004e-05,
|
| 1435 |
+
"loss": 0.6113,
|
| 1436 |
+
"step": 198
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 24.88,
|
| 1440 |
+
"learning_rate": 1.9900000000000003e-05,
|
| 1441 |
+
"loss": 0.6261,
|
| 1442 |
+
"step": 199
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 25.0,
|
| 1446 |
+
"learning_rate": 2e-05,
|
| 1447 |
+
"loss": 0.6475,
|
| 1448 |
+
"step": 200
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 25.0,
|
| 1452 |
+
"eval_accuracy": 0.585,
|
| 1453 |
+
"eval_loss": 0.6710052490234375,
|
| 1454 |
+
"eval_macro_f1": 0.5359888190076869,
|
| 1455 |
+
"eval_runtime": 0.7024,
|
| 1456 |
+
"eval_samples_per_second": 284.724,
|
| 1457 |
+
"eval_steps_per_second": 2.847,
|
| 1458 |
+
"step": 200
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 25.12,
|
| 1462 |
+
"learning_rate": 2.01e-05,
|
| 1463 |
+
"loss": 0.605,
|
| 1464 |
+
"step": 201
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 25.25,
|
| 1468 |
+
"learning_rate": 2.0200000000000003e-05,
|
| 1469 |
+
"loss": 0.6224,
|
| 1470 |
+
"step": 202
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 25.38,
|
| 1474 |
+
"learning_rate": 2.0300000000000002e-05,
|
| 1475 |
+
"loss": 0.6304,
|
| 1476 |
+
"step": 203
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 25.5,
|
| 1480 |
+
"learning_rate": 2.04e-05,
|
| 1481 |
+
"loss": 0.5808,
|
| 1482 |
+
"step": 204
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 25.62,
|
| 1486 |
+
"learning_rate": 2.05e-05,
|
| 1487 |
+
"loss": 0.5741,
|
| 1488 |
+
"step": 205
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 25.75,
|
| 1492 |
+
"learning_rate": 2.06e-05,
|
| 1493 |
+
"loss": 0.61,
|
| 1494 |
+
"step": 206
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 25.88,
|
| 1498 |
+
"learning_rate": 2.07e-05,
|
| 1499 |
+
"loss": 0.6129,
|
| 1500 |
+
"step": 207
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 26.0,
|
| 1504 |
+
"learning_rate": 2.08e-05,
|
| 1505 |
+
"loss": 0.5683,
|
| 1506 |
+
"step": 208
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 26.0,
|
| 1510 |
+
"eval_accuracy": 0.625,
|
| 1511 |
+
"eval_loss": 0.6624136567115784,
|
| 1512 |
+
"eval_macro_f1": 0.6180387563341907,
|
| 1513 |
+
"eval_runtime": 0.688,
|
| 1514 |
+
"eval_samples_per_second": 290.698,
|
| 1515 |
+
"eval_steps_per_second": 2.907,
|
| 1516 |
+
"step": 208
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 26.12,
|
| 1520 |
+
"learning_rate": 2.09e-05,
|
| 1521 |
+
"loss": 0.5728,
|
| 1522 |
+
"step": 209
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 26.25,
|
| 1526 |
+
"learning_rate": 2.1e-05,
|
| 1527 |
+
"loss": 0.6,
|
| 1528 |
+
"step": 210
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 26.38,
|
| 1532 |
+
"learning_rate": 2.11e-05,
|
| 1533 |
+
"loss": 0.5585,
|
| 1534 |
+
"step": 211
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 26.5,
|
| 1538 |
+
"learning_rate": 2.12e-05,
|
| 1539 |
+
"loss": 0.583,
|
| 1540 |
+
"step": 212
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 26.62,
|
| 1544 |
+
"learning_rate": 2.13e-05,
|
| 1545 |
+
"loss": 0.5753,
|
| 1546 |
+
"step": 213
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 26.75,
|
| 1550 |
+
"learning_rate": 2.1400000000000002e-05,
|
| 1551 |
+
"loss": 0.6253,
|
| 1552 |
+
"step": 214
|
| 1553 |
+
},
|
| 1554 |
+
{
|
| 1555 |
+
"epoch": 26.88,
|
| 1556 |
+
"learning_rate": 2.15e-05,
|
| 1557 |
+
"loss": 0.5611,
|
| 1558 |
+
"step": 215
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 27.0,
|
| 1562 |
+
"learning_rate": 2.16e-05,
|
| 1563 |
+
"loss": 0.6105,
|
| 1564 |
+
"step": 216
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 27.0,
|
| 1568 |
+
"eval_accuracy": 0.585,
|
| 1569 |
+
"eval_loss": 0.6604620218276978,
|
| 1570 |
+
"eval_macro_f1": 0.5536554542765723,
|
| 1571 |
+
"eval_runtime": 0.6984,
|
| 1572 |
+
"eval_samples_per_second": 286.362,
|
| 1573 |
+
"eval_steps_per_second": 2.864,
|
| 1574 |
+
"step": 216
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 27.12,
|
| 1578 |
+
"learning_rate": 2.1700000000000002e-05,
|
| 1579 |
+
"loss": 0.5816,
|
| 1580 |
+
"step": 217
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 27.25,
|
| 1584 |
+
"learning_rate": 2.18e-05,
|
| 1585 |
+
"loss": 0.5959,
|
| 1586 |
+
"step": 218
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 27.38,
|
| 1590 |
+
"learning_rate": 2.19e-05,
|
| 1591 |
+
"loss": 0.6061,
|
| 1592 |
+
"step": 219
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 27.5,
|
| 1596 |
+
"learning_rate": 2.2000000000000003e-05,
|
| 1597 |
+
"loss": 0.5869,
|
| 1598 |
+
"step": 220
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 27.62,
|
| 1602 |
+
"learning_rate": 2.2100000000000002e-05,
|
| 1603 |
+
"loss": 0.6022,
|
| 1604 |
+
"step": 221
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 27.75,
|
| 1608 |
+
"learning_rate": 2.22e-05,
|
| 1609 |
+
"loss": 0.5463,
|
| 1610 |
+
"step": 222
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 27.88,
|
| 1614 |
+
"learning_rate": 2.23e-05,
|
| 1615 |
+
"loss": 0.5492,
|
| 1616 |
+
"step": 223
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 28.0,
|
| 1620 |
+
"learning_rate": 2.2400000000000002e-05,
|
| 1621 |
+
"loss": 0.5077,
|
| 1622 |
+
"step": 224
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 28.0,
|
| 1626 |
+
"eval_accuracy": 0.615,
|
| 1627 |
+
"eval_loss": 0.6587879657745361,
|
| 1628 |
+
"eval_macro_f1": 0.6042250263421655,
|
| 1629 |
+
"eval_runtime": 0.6976,
|
| 1630 |
+
"eval_samples_per_second": 286.693,
|
| 1631 |
+
"eval_steps_per_second": 2.867,
|
| 1632 |
+
"step": 224
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 28.12,
|
| 1636 |
+
"learning_rate": 2.25e-05,
|
| 1637 |
+
"loss": 0.5857,
|
| 1638 |
+
"step": 225
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 28.25,
|
| 1642 |
+
"learning_rate": 2.26e-05,
|
| 1643 |
+
"loss": 0.5795,
|
| 1644 |
+
"step": 226
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 28.38,
|
| 1648 |
+
"learning_rate": 2.2700000000000003e-05,
|
| 1649 |
+
"loss": 0.563,
|
| 1650 |
+
"step": 227
|
| 1651 |
+
},
|
| 1652 |
+
{
|
| 1653 |
+
"epoch": 28.5,
|
| 1654 |
+
"learning_rate": 2.2800000000000002e-05,
|
| 1655 |
+
"loss": 0.5889,
|
| 1656 |
+
"step": 228
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 28.62,
|
| 1660 |
+
"learning_rate": 2.29e-05,
|
| 1661 |
+
"loss": 0.5617,
|
| 1662 |
+
"step": 229
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 28.75,
|
| 1666 |
+
"learning_rate": 2.3000000000000003e-05,
|
| 1667 |
+
"loss": 0.5756,
|
| 1668 |
+
"step": 230
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 28.88,
|
| 1672 |
+
"learning_rate": 2.3100000000000002e-05,
|
| 1673 |
+
"loss": 0.5581,
|
| 1674 |
+
"step": 231
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 29.0,
|
| 1678 |
+
"learning_rate": 2.32e-05,
|
| 1679 |
+
"loss": 0.5557,
|
| 1680 |
+
"step": 232
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 29.0,
|
| 1684 |
+
"eval_accuracy": 0.595,
|
| 1685 |
+
"eval_loss": 0.660551130771637,
|
| 1686 |
+
"eval_macro_f1": 0.5644107445349681,
|
| 1687 |
+
"eval_runtime": 0.6821,
|
| 1688 |
+
"eval_samples_per_second": 293.231,
|
| 1689 |
+
"eval_steps_per_second": 2.932,
|
| 1690 |
+
"step": 232
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 29.12,
|
| 1694 |
+
"learning_rate": 2.3300000000000004e-05,
|
| 1695 |
+
"loss": 0.6115,
|
| 1696 |
+
"step": 233
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 29.25,
|
| 1700 |
+
"learning_rate": 2.3400000000000003e-05,
|
| 1701 |
+
"loss": 0.5437,
|
| 1702 |
+
"step": 234
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 29.38,
|
| 1706 |
+
"learning_rate": 2.35e-05,
|
| 1707 |
+
"loss": 0.5806,
|
| 1708 |
+
"step": 235
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 29.5,
|
| 1712 |
+
"learning_rate": 2.36e-05,
|
| 1713 |
+
"loss": 0.5529,
|
| 1714 |
+
"step": 236
|
| 1715 |
+
},
|
| 1716 |
+
{
|
| 1717 |
+
"epoch": 29.62,
|
| 1718 |
+
"learning_rate": 2.37e-05,
|
| 1719 |
+
"loss": 0.5296,
|
| 1720 |
+
"step": 237
|
| 1721 |
+
},
|
| 1722 |
+
{
|
| 1723 |
+
"epoch": 29.75,
|
| 1724 |
+
"learning_rate": 2.38e-05,
|
| 1725 |
+
"loss": 0.5341,
|
| 1726 |
+
"step": 238
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 29.88,
|
| 1730 |
+
"learning_rate": 2.39e-05,
|
| 1731 |
+
"loss": 0.5454,
|
| 1732 |
+
"step": 239
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 30.0,
|
| 1736 |
+
"learning_rate": 2.4e-05,
|
| 1737 |
+
"loss": 0.5498,
|
| 1738 |
+
"step": 240
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 30.0,
|
| 1742 |
+
"eval_accuracy": 0.63,
|
| 1743 |
+
"eval_loss": 0.6594185829162598,
|
| 1744 |
+
"eval_macro_f1": 0.6269785260610949,
|
| 1745 |
+
"eval_runtime": 0.6994,
|
| 1746 |
+
"eval_samples_per_second": 285.941,
|
| 1747 |
+
"eval_steps_per_second": 2.859,
|
| 1748 |
+
"step": 240
|
| 1749 |
+
}
|
| 1750 |
+
],
|
| 1751 |
+
"max_steps": 400,
|
| 1752 |
+
"num_train_epochs": 50,
|
| 1753 |
+
"total_flos": 952401199104000.0,
|
| 1754 |
+
"trial_name": null,
|
| 1755 |
+
"trial_params": null
|
| 1756 |
+
}
|
scaling_performance/2000/.DS_Store
CHANGED
|
Binary files a/scaling_performance/2000/.DS_Store and b/scaling_performance/2000/.DS_Store differ
|
|
|
scaling_performance/2000/L1/.DS_Store
CHANGED
|
Binary files a/scaling_performance/2000/L1/.DS_Store and b/scaling_performance/2000/L1/.DS_Store differ
|
|
|
scaling_performance/2000/L1/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.5775,
|
| 3 |
+
"test_loss": 0.6810634136199951,
|
| 4 |
+
"test_macro_f1": 0.36608557844690964,
|
| 5 |
+
"test_runtime": 0.9129,
|
| 6 |
+
"test_samples_per_second": 438.178,
|
| 7 |
+
"test_steps_per_second": 4.382
|
| 8 |
+
}
|
scaling_performance/2000/L1/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.02,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.02,
|
| 9 |
+
"hidden_size": 256,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 512,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 2048,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 4,
|
| 16 |
+
"num_hidden_layers": 1,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"problem_type": "single_label_classification",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.28.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
scaling_performance/2000/L1/eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.5775,
|
| 3 |
+
"test_loss": 0.6810634136199951,
|
| 4 |
+
"test_macro_f1": 0.36608557844690964,
|
| 5 |
+
"test_runtime": 0.9129,
|
| 6 |
+
"test_samples_per_second": 438.178,
|
| 7 |
+
"test_steps_per_second": 4.382
|
| 8 |
+
}
|
scaling_performance/2000/L1/trainer_state.json
ADDED
|
@@ -0,0 +1,886 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6810634136199951,
|
| 3 |
+
"best_model_checkpoint": "./models/240626_geneformer_CellClassifier_PM25_Layers1_L2048_B26_LR5e-05_LSlinear_WU600_E20_Oadamw_F0_fold4/checkpoint-208",
|
| 4 |
+
"epoch": 15.0,
|
| 5 |
+
"global_step": 240,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 1.6666666666666668e-07,
|
| 13 |
+
"loss": 0.6939,
|
| 14 |
+
"step": 2
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.25,
|
| 18 |
+
"learning_rate": 3.3333333333333335e-07,
|
| 19 |
+
"loss": 0.6887,
|
| 20 |
+
"step": 4
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.38,
|
| 24 |
+
"learning_rate": 5.000000000000001e-07,
|
| 25 |
+
"loss": 0.6921,
|
| 26 |
+
"step": 6
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.5,
|
| 30 |
+
"learning_rate": 6.666666666666667e-07,
|
| 31 |
+
"loss": 0.6899,
|
| 32 |
+
"step": 8
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.62,
|
| 36 |
+
"learning_rate": 8.333333333333333e-07,
|
| 37 |
+
"loss": 0.6913,
|
| 38 |
+
"step": 10
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.75,
|
| 42 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 43 |
+
"loss": 0.691,
|
| 44 |
+
"step": 12
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.88,
|
| 48 |
+
"learning_rate": 1.1666666666666668e-06,
|
| 49 |
+
"loss": 0.6912,
|
| 50 |
+
"step": 14
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 55 |
+
"loss": 0.689,
|
| 56 |
+
"step": 16
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.0,
|
| 60 |
+
"eval_accuracy": 0.505,
|
| 61 |
+
"eval_loss": 0.6925567388534546,
|
| 62 |
+
"eval_macro_f1": 0.49510403916768664,
|
| 63 |
+
"eval_runtime": 0.8698,
|
| 64 |
+
"eval_samples_per_second": 459.879,
|
| 65 |
+
"eval_steps_per_second": 4.599,
|
| 66 |
+
"step": 16
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.12,
|
| 70 |
+
"learning_rate": 1.5e-06,
|
| 71 |
+
"loss": 0.6906,
|
| 72 |
+
"step": 18
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.25,
|
| 76 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 77 |
+
"loss": 0.6896,
|
| 78 |
+
"step": 20
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.38,
|
| 82 |
+
"learning_rate": 1.8333333333333335e-06,
|
| 83 |
+
"loss": 0.6892,
|
| 84 |
+
"step": 22
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5,
|
| 88 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 89 |
+
"loss": 0.6984,
|
| 90 |
+
"step": 24
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.62,
|
| 94 |
+
"learning_rate": 2.166666666666667e-06,
|
| 95 |
+
"loss": 0.6923,
|
| 96 |
+
"step": 26
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.75,
|
| 100 |
+
"learning_rate": 2.3333333333333336e-06,
|
| 101 |
+
"loss": 0.6879,
|
| 102 |
+
"step": 28
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.88,
|
| 106 |
+
"learning_rate": 2.5e-06,
|
| 107 |
+
"loss": 0.6918,
|
| 108 |
+
"step": 30
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 2.0,
|
| 112 |
+
"learning_rate": 2.666666666666667e-06,
|
| 113 |
+
"loss": 0.6832,
|
| 114 |
+
"step": 32
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 2.0,
|
| 118 |
+
"eval_accuracy": 0.5375,
|
| 119 |
+
"eval_loss": 0.6915823221206665,
|
| 120 |
+
"eval_macro_f1": 0.513410793074652,
|
| 121 |
+
"eval_runtime": 1.0103,
|
| 122 |
+
"eval_samples_per_second": 395.928,
|
| 123 |
+
"eval_steps_per_second": 3.959,
|
| 124 |
+
"step": 32
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.12,
|
| 128 |
+
"learning_rate": 2.8333333333333335e-06,
|
| 129 |
+
"loss": 0.6928,
|
| 130 |
+
"step": 34
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.25,
|
| 134 |
+
"learning_rate": 3e-06,
|
| 135 |
+
"loss": 0.6909,
|
| 136 |
+
"step": 36
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.38,
|
| 140 |
+
"learning_rate": 3.166666666666667e-06,
|
| 141 |
+
"loss": 0.6901,
|
| 142 |
+
"step": 38
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.5,
|
| 146 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 147 |
+
"loss": 0.6891,
|
| 148 |
+
"step": 40
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.62,
|
| 152 |
+
"learning_rate": 3.5000000000000004e-06,
|
| 153 |
+
"loss": 0.6851,
|
| 154 |
+
"step": 42
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.75,
|
| 158 |
+
"learning_rate": 3.666666666666667e-06,
|
| 159 |
+
"loss": 0.6853,
|
| 160 |
+
"step": 44
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.88,
|
| 164 |
+
"learning_rate": 3.833333333333334e-06,
|
| 165 |
+
"loss": 0.6949,
|
| 166 |
+
"step": 46
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.0,
|
| 170 |
+
"learning_rate": 4.000000000000001e-06,
|
| 171 |
+
"loss": 0.6825,
|
| 172 |
+
"step": 48
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.5475,
|
| 177 |
+
"eval_loss": 0.689637303352356,
|
| 178 |
+
"eval_macro_f1": 0.49125506812544356,
|
| 179 |
+
"eval_runtime": 0.8008,
|
| 180 |
+
"eval_samples_per_second": 499.483,
|
| 181 |
+
"eval_steps_per_second": 4.995,
|
| 182 |
+
"step": 48
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.12,
|
| 186 |
+
"learning_rate": 4.166666666666667e-06,
|
| 187 |
+
"loss": 0.6895,
|
| 188 |
+
"step": 50
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.25,
|
| 192 |
+
"learning_rate": 4.333333333333334e-06,
|
| 193 |
+
"loss": 0.6837,
|
| 194 |
+
"step": 52
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.38,
|
| 198 |
+
"learning_rate": 4.5e-06,
|
| 199 |
+
"loss": 0.689,
|
| 200 |
+
"step": 54
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.5,
|
| 204 |
+
"learning_rate": 4.666666666666667e-06,
|
| 205 |
+
"loss": 0.6954,
|
| 206 |
+
"step": 56
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.62,
|
| 210 |
+
"learning_rate": 4.833333333333333e-06,
|
| 211 |
+
"loss": 0.6826,
|
| 212 |
+
"step": 58
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.75,
|
| 216 |
+
"learning_rate": 5e-06,
|
| 217 |
+
"loss": 0.6898,
|
| 218 |
+
"step": 60
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.88,
|
| 222 |
+
"learning_rate": 5.166666666666667e-06,
|
| 223 |
+
"loss": 0.6897,
|
| 224 |
+
"step": 62
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.0,
|
| 228 |
+
"learning_rate": 5.333333333333334e-06,
|
| 229 |
+
"loss": 0.6791,
|
| 230 |
+
"step": 64
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.0,
|
| 234 |
+
"eval_accuracy": 0.55,
|
| 235 |
+
"eval_loss": 0.6878230571746826,
|
| 236 |
+
"eval_macro_f1": 0.4642857142857143,
|
| 237 |
+
"eval_runtime": 0.7978,
|
| 238 |
+
"eval_samples_per_second": 501.403,
|
| 239 |
+
"eval_steps_per_second": 5.014,
|
| 240 |
+
"step": 64
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.12,
|
| 244 |
+
"learning_rate": 5.500000000000001e-06,
|
| 245 |
+
"loss": 0.6911,
|
| 246 |
+
"step": 66
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.25,
|
| 250 |
+
"learning_rate": 5.666666666666667e-06,
|
| 251 |
+
"loss": 0.685,
|
| 252 |
+
"step": 68
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.38,
|
| 256 |
+
"learning_rate": 5.833333333333334e-06,
|
| 257 |
+
"loss": 0.6865,
|
| 258 |
+
"step": 70
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.5,
|
| 262 |
+
"learning_rate": 6e-06,
|
| 263 |
+
"loss": 0.6938,
|
| 264 |
+
"step": 72
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.62,
|
| 268 |
+
"learning_rate": 6.166666666666667e-06,
|
| 269 |
+
"loss": 0.6845,
|
| 270 |
+
"step": 74
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 4.75,
|
| 274 |
+
"learning_rate": 6.333333333333334e-06,
|
| 275 |
+
"loss": 0.6838,
|
| 276 |
+
"step": 76
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 4.88,
|
| 280 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 281 |
+
"loss": 0.6883,
|
| 282 |
+
"step": 78
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 5.0,
|
| 286 |
+
"learning_rate": 6.666666666666667e-06,
|
| 287 |
+
"loss": 0.6804,
|
| 288 |
+
"step": 80
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.0,
|
| 292 |
+
"eval_accuracy": 0.57,
|
| 293 |
+
"eval_loss": 0.6861330270767212,
|
| 294 |
+
"eval_macro_f1": 0.4377431270635154,
|
| 295 |
+
"eval_runtime": 0.9894,
|
| 296 |
+
"eval_samples_per_second": 404.267,
|
| 297 |
+
"eval_steps_per_second": 4.043,
|
| 298 |
+
"step": 80
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.12,
|
| 302 |
+
"learning_rate": 6.833333333333333e-06,
|
| 303 |
+
"loss": 0.6921,
|
| 304 |
+
"step": 82
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.25,
|
| 308 |
+
"learning_rate": 7.000000000000001e-06,
|
| 309 |
+
"loss": 0.6867,
|
| 310 |
+
"step": 84
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.38,
|
| 314 |
+
"learning_rate": 7.166666666666667e-06,
|
| 315 |
+
"loss": 0.6789,
|
| 316 |
+
"step": 86
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.5,
|
| 320 |
+
"learning_rate": 7.333333333333334e-06,
|
| 321 |
+
"loss": 0.6912,
|
| 322 |
+
"step": 88
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 5.62,
|
| 326 |
+
"learning_rate": 7.5e-06,
|
| 327 |
+
"loss": 0.6885,
|
| 328 |
+
"step": 90
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.75,
|
| 332 |
+
"learning_rate": 7.666666666666667e-06,
|
| 333 |
+
"loss": 0.6824,
|
| 334 |
+
"step": 92
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 5.88,
|
| 338 |
+
"learning_rate": 7.833333333333333e-06,
|
| 339 |
+
"loss": 0.6915,
|
| 340 |
+
"step": 94
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 6.0,
|
| 344 |
+
"learning_rate": 8.000000000000001e-06,
|
| 345 |
+
"loss": 0.6669,
|
| 346 |
+
"step": 96
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.0,
|
| 350 |
+
"eval_accuracy": 0.5625,
|
| 351 |
+
"eval_loss": 0.685230016708374,
|
| 352 |
+
"eval_macro_f1": 0.3893449415952055,
|
| 353 |
+
"eval_runtime": 1.0307,
|
| 354 |
+
"eval_samples_per_second": 388.076,
|
| 355 |
+
"eval_steps_per_second": 3.881,
|
| 356 |
+
"step": 96
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.12,
|
| 360 |
+
"learning_rate": 8.166666666666668e-06,
|
| 361 |
+
"loss": 0.6902,
|
| 362 |
+
"step": 98
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.25,
|
| 366 |
+
"learning_rate": 8.333333333333334e-06,
|
| 367 |
+
"loss": 0.6888,
|
| 368 |
+
"step": 100
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.38,
|
| 372 |
+
"learning_rate": 8.500000000000002e-06,
|
| 373 |
+
"loss": 0.6718,
|
| 374 |
+
"step": 102
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.5,
|
| 378 |
+
"learning_rate": 8.666666666666668e-06,
|
| 379 |
+
"loss": 0.6953,
|
| 380 |
+
"step": 104
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.62,
|
| 384 |
+
"learning_rate": 8.833333333333334e-06,
|
| 385 |
+
"loss": 0.679,
|
| 386 |
+
"step": 106
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 6.75,
|
| 390 |
+
"learning_rate": 9e-06,
|
| 391 |
+
"loss": 0.6763,
|
| 392 |
+
"step": 108
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 6.88,
|
| 396 |
+
"learning_rate": 9.166666666666666e-06,
|
| 397 |
+
"loss": 0.6932,
|
| 398 |
+
"step": 110
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.0,
|
| 402 |
+
"learning_rate": 9.333333333333334e-06,
|
| 403 |
+
"loss": 0.672,
|
| 404 |
+
"step": 112
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.0,
|
| 408 |
+
"eval_accuracy": 0.5725,
|
| 409 |
+
"eval_loss": 0.6839529275894165,
|
| 410 |
+
"eval_macro_f1": 0.3640699523052464,
|
| 411 |
+
"eval_runtime": 0.8433,
|
| 412 |
+
"eval_samples_per_second": 474.35,
|
| 413 |
+
"eval_steps_per_second": 4.744,
|
| 414 |
+
"step": 112
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.12,
|
| 418 |
+
"learning_rate": 9.5e-06,
|
| 419 |
+
"loss": 0.6954,
|
| 420 |
+
"step": 114
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.25,
|
| 424 |
+
"learning_rate": 9.666666666666667e-06,
|
| 425 |
+
"loss": 0.6823,
|
| 426 |
+
"step": 116
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.38,
|
| 430 |
+
"learning_rate": 9.833333333333333e-06,
|
| 431 |
+
"loss": 0.6712,
|
| 432 |
+
"step": 118
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.5,
|
| 436 |
+
"learning_rate": 1e-05,
|
| 437 |
+
"loss": 0.696,
|
| 438 |
+
"step": 120
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 7.62,
|
| 442 |
+
"learning_rate": 1.0166666666666667e-05,
|
| 443 |
+
"loss": 0.6741,
|
| 444 |
+
"step": 122
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.75,
|
| 448 |
+
"learning_rate": 1.0333333333333333e-05,
|
| 449 |
+
"loss": 0.6594,
|
| 450 |
+
"step": 124
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 7.88,
|
| 454 |
+
"learning_rate": 1.05e-05,
|
| 455 |
+
"loss": 0.7076,
|
| 456 |
+
"step": 126
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 8.0,
|
| 460 |
+
"learning_rate": 1.0666666666666667e-05,
|
| 461 |
+
"loss": 0.6754,
|
| 462 |
+
"step": 128
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.0,
|
| 466 |
+
"eval_accuracy": 0.575,
|
| 467 |
+
"eval_loss": 0.6830737590789795,
|
| 468 |
+
"eval_macro_f1": 0.36507936507936506,
|
| 469 |
+
"eval_runtime": 0.8854,
|
| 470 |
+
"eval_samples_per_second": 451.794,
|
| 471 |
+
"eval_steps_per_second": 4.518,
|
| 472 |
+
"step": 128
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.12,
|
| 476 |
+
"learning_rate": 1.0833333333333334e-05,
|
| 477 |
+
"loss": 0.6895,
|
| 478 |
+
"step": 130
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 8.25,
|
| 482 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 483 |
+
"loss": 0.6817,
|
| 484 |
+
"step": 132
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 8.38,
|
| 488 |
+
"learning_rate": 1.1166666666666668e-05,
|
| 489 |
+
"loss": 0.6737,
|
| 490 |
+
"step": 134
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.5,
|
| 494 |
+
"learning_rate": 1.1333333333333334e-05,
|
| 495 |
+
"loss": 0.6945,
|
| 496 |
+
"step": 136
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.62,
|
| 500 |
+
"learning_rate": 1.1500000000000002e-05,
|
| 501 |
+
"loss": 0.6698,
|
| 502 |
+
"step": 138
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 8.75,
|
| 506 |
+
"learning_rate": 1.1666666666666668e-05,
|
| 507 |
+
"loss": 0.6699,
|
| 508 |
+
"step": 140
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 8.88,
|
| 512 |
+
"learning_rate": 1.1833333333333334e-05,
|
| 513 |
+
"loss": 0.6993,
|
| 514 |
+
"step": 142
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 9.0,
|
| 518 |
+
"learning_rate": 1.2e-05,
|
| 519 |
+
"loss": 0.6833,
|
| 520 |
+
"step": 144
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 9.0,
|
| 524 |
+
"eval_accuracy": 0.575,
|
| 525 |
+
"eval_loss": 0.6830700039863586,
|
| 526 |
+
"eval_macro_f1": 0.36507936507936506,
|
| 527 |
+
"eval_runtime": 0.9198,
|
| 528 |
+
"eval_samples_per_second": 434.883,
|
| 529 |
+
"eval_steps_per_second": 4.349,
|
| 530 |
+
"step": 144
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 9.12,
|
| 534 |
+
"learning_rate": 1.2166666666666668e-05,
|
| 535 |
+
"loss": 0.6917,
|
| 536 |
+
"step": 146
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.25,
|
| 540 |
+
"learning_rate": 1.2333333333333334e-05,
|
| 541 |
+
"loss": 0.6684,
|
| 542 |
+
"step": 148
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 9.38,
|
| 546 |
+
"learning_rate": 1.25e-05,
|
| 547 |
+
"loss": 0.6738,
|
| 548 |
+
"step": 150
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 9.5,
|
| 552 |
+
"learning_rate": 1.2666666666666668e-05,
|
| 553 |
+
"loss": 0.7013,
|
| 554 |
+
"step": 152
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 9.62,
|
| 558 |
+
"learning_rate": 1.2833333333333333e-05,
|
| 559 |
+
"loss": 0.6761,
|
| 560 |
+
"step": 154
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 9.75,
|
| 564 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 565 |
+
"loss": 0.6761,
|
| 566 |
+
"step": 156
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 9.88,
|
| 570 |
+
"learning_rate": 1.3166666666666665e-05,
|
| 571 |
+
"loss": 0.6897,
|
| 572 |
+
"step": 158
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 10.0,
|
| 576 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 577 |
+
"loss": 0.6625,
|
| 578 |
+
"step": 160
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.0,
|
| 582 |
+
"eval_accuracy": 0.57,
|
| 583 |
+
"eval_loss": 0.6834887862205505,
|
| 584 |
+
"eval_macro_f1": 0.36836693474349086,
|
| 585 |
+
"eval_runtime": 0.842,
|
| 586 |
+
"eval_samples_per_second": 475.084,
|
| 587 |
+
"eval_steps_per_second": 4.751,
|
| 588 |
+
"step": 160
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 10.12,
|
| 592 |
+
"learning_rate": 1.3500000000000001e-05,
|
| 593 |
+
"loss": 0.689,
|
| 594 |
+
"step": 162
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 10.25,
|
| 598 |
+
"learning_rate": 1.3666666666666666e-05,
|
| 599 |
+
"loss": 0.6762,
|
| 600 |
+
"step": 164
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 10.38,
|
| 604 |
+
"learning_rate": 1.3833333333333334e-05,
|
| 605 |
+
"loss": 0.6749,
|
| 606 |
+
"step": 166
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 10.5,
|
| 610 |
+
"learning_rate": 1.4000000000000001e-05,
|
| 611 |
+
"loss": 0.6952,
|
| 612 |
+
"step": 168
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 10.62,
|
| 616 |
+
"learning_rate": 1.4166666666666668e-05,
|
| 617 |
+
"loss": 0.671,
|
| 618 |
+
"step": 170
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 10.75,
|
| 622 |
+
"learning_rate": 1.4333333333333334e-05,
|
| 623 |
+
"loss": 0.6755,
|
| 624 |
+
"step": 172
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.88,
|
| 628 |
+
"learning_rate": 1.45e-05,
|
| 629 |
+
"loss": 0.6946,
|
| 630 |
+
"step": 174
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 11.0,
|
| 634 |
+
"learning_rate": 1.4666666666666668e-05,
|
| 635 |
+
"loss": 0.6535,
|
| 636 |
+
"step": 176
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 11.0,
|
| 640 |
+
"eval_accuracy": 0.575,
|
| 641 |
+
"eval_loss": 0.6825265288352966,
|
| 642 |
+
"eval_macro_f1": 0.36507936507936506,
|
| 643 |
+
"eval_runtime": 0.7992,
|
| 644 |
+
"eval_samples_per_second": 500.47,
|
| 645 |
+
"eval_steps_per_second": 5.005,
|
| 646 |
+
"step": 176
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 11.12,
|
| 650 |
+
"learning_rate": 1.4833333333333336e-05,
|
| 651 |
+
"loss": 0.6953,
|
| 652 |
+
"step": 178
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 11.25,
|
| 656 |
+
"learning_rate": 1.5e-05,
|
| 657 |
+
"loss": 0.6763,
|
| 658 |
+
"step": 180
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 11.38,
|
| 662 |
+
"learning_rate": 1.5166666666666668e-05,
|
| 663 |
+
"loss": 0.6666,
|
| 664 |
+
"step": 182
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 11.5,
|
| 668 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 669 |
+
"loss": 0.6878,
|
| 670 |
+
"step": 184
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 11.62,
|
| 674 |
+
"learning_rate": 1.55e-05,
|
| 675 |
+
"loss": 0.6762,
|
| 676 |
+
"step": 186
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 11.75,
|
| 680 |
+
"learning_rate": 1.5666666666666667e-05,
|
| 681 |
+
"loss": 0.6701,
|
| 682 |
+
"step": 188
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 11.88,
|
| 686 |
+
"learning_rate": 1.5833333333333333e-05,
|
| 687 |
+
"loss": 0.683,
|
| 688 |
+
"step": 190
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 12.0,
|
| 692 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 693 |
+
"loss": 0.6597,
|
| 694 |
+
"step": 192
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 12.0,
|
| 698 |
+
"eval_accuracy": 0.5775,
|
| 699 |
+
"eval_loss": 0.6820663213729858,
|
| 700 |
+
"eval_macro_f1": 0.36608557844690964,
|
| 701 |
+
"eval_runtime": 0.8333,
|
| 702 |
+
"eval_samples_per_second": 480.024,
|
| 703 |
+
"eval_steps_per_second": 4.8,
|
| 704 |
+
"step": 192
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 12.12,
|
| 708 |
+
"learning_rate": 1.6166666666666665e-05,
|
| 709 |
+
"loss": 0.6837,
|
| 710 |
+
"step": 194
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 12.25,
|
| 714 |
+
"learning_rate": 1.6333333333333335e-05,
|
| 715 |
+
"loss": 0.6707,
|
| 716 |
+
"step": 196
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 12.38,
|
| 720 |
+
"learning_rate": 1.65e-05,
|
| 721 |
+
"loss": 0.6663,
|
| 722 |
+
"step": 198
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 12.5,
|
| 726 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 727 |
+
"loss": 0.7006,
|
| 728 |
+
"step": 200
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 12.62,
|
| 732 |
+
"learning_rate": 1.6833333333333334e-05,
|
| 733 |
+
"loss": 0.6666,
|
| 734 |
+
"step": 202
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 12.75,
|
| 738 |
+
"learning_rate": 1.7000000000000003e-05,
|
| 739 |
+
"loss": 0.6649,
|
| 740 |
+
"step": 204
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 12.88,
|
| 744 |
+
"learning_rate": 1.7166666666666666e-05,
|
| 745 |
+
"loss": 0.6918,
|
| 746 |
+
"step": 206
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 13.0,
|
| 750 |
+
"learning_rate": 1.7333333333333336e-05,
|
| 751 |
+
"loss": 0.6747,
|
| 752 |
+
"step": 208
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 13.0,
|
| 756 |
+
"eval_accuracy": 0.5775,
|
| 757 |
+
"eval_loss": 0.6810634136199951,
|
| 758 |
+
"eval_macro_f1": 0.36608557844690964,
|
| 759 |
+
"eval_runtime": 0.8039,
|
| 760 |
+
"eval_samples_per_second": 497.582,
|
| 761 |
+
"eval_steps_per_second": 4.976,
|
| 762 |
+
"step": 208
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 13.12,
|
| 766 |
+
"learning_rate": 1.75e-05,
|
| 767 |
+
"loss": 0.6878,
|
| 768 |
+
"step": 210
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 13.25,
|
| 772 |
+
"learning_rate": 1.7666666666666668e-05,
|
| 773 |
+
"loss": 0.6655,
|
| 774 |
+
"step": 212
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 13.38,
|
| 778 |
+
"learning_rate": 1.7833333333333334e-05,
|
| 779 |
+
"loss": 0.6616,
|
| 780 |
+
"step": 214
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 13.5,
|
| 784 |
+
"learning_rate": 1.8e-05,
|
| 785 |
+
"loss": 0.6951,
|
| 786 |
+
"step": 216
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 13.62,
|
| 790 |
+
"learning_rate": 1.8166666666666667e-05,
|
| 791 |
+
"loss": 0.6714,
|
| 792 |
+
"step": 218
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 13.75,
|
| 796 |
+
"learning_rate": 1.8333333333333333e-05,
|
| 797 |
+
"loss": 0.6741,
|
| 798 |
+
"step": 220
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 13.88,
|
| 802 |
+
"learning_rate": 1.85e-05,
|
| 803 |
+
"loss": 0.6853,
|
| 804 |
+
"step": 222
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 14.0,
|
| 808 |
+
"learning_rate": 1.866666666666667e-05,
|
| 809 |
+
"loss": 0.6661,
|
| 810 |
+
"step": 224
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 14.0,
|
| 814 |
+
"eval_accuracy": 0.5775,
|
| 815 |
+
"eval_loss": 0.681349515914917,
|
| 816 |
+
"eval_macro_f1": 0.38696484116404134,
|
| 817 |
+
"eval_runtime": 0.8726,
|
| 818 |
+
"eval_samples_per_second": 458.379,
|
| 819 |
+
"eval_steps_per_second": 4.584,
|
| 820 |
+
"step": 224
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 14.12,
|
| 824 |
+
"learning_rate": 1.8833333333333335e-05,
|
| 825 |
+
"loss": 0.6902,
|
| 826 |
+
"step": 226
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 14.25,
|
| 830 |
+
"learning_rate": 1.9e-05,
|
| 831 |
+
"loss": 0.674,
|
| 832 |
+
"step": 228
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 14.38,
|
| 836 |
+
"learning_rate": 1.9166666666666667e-05,
|
| 837 |
+
"loss": 0.6707,
|
| 838 |
+
"step": 230
|
| 839 |
+
},
|
| 840 |
+
{
|
| 841 |
+
"epoch": 14.5,
|
| 842 |
+
"learning_rate": 1.9333333333333333e-05,
|
| 843 |
+
"loss": 0.6799,
|
| 844 |
+
"step": 232
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 14.62,
|
| 848 |
+
"learning_rate": 1.9500000000000003e-05,
|
| 849 |
+
"loss": 0.6607,
|
| 850 |
+
"step": 234
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 14.75,
|
| 854 |
+
"learning_rate": 1.9666666666666666e-05,
|
| 855 |
+
"loss": 0.6748,
|
| 856 |
+
"step": 236
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 14.88,
|
| 860 |
+
"learning_rate": 1.9833333333333335e-05,
|
| 861 |
+
"loss": 0.6771,
|
| 862 |
+
"step": 238
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 15.0,
|
| 866 |
+
"learning_rate": 2e-05,
|
| 867 |
+
"loss": 0.6543,
|
| 868 |
+
"step": 240
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 15.0,
|
| 872 |
+
"eval_accuracy": 0.5725,
|
| 873 |
+
"eval_loss": 0.6816413998603821,
|
| 874 |
+
"eval_macro_f1": 0.4076845140674928,
|
| 875 |
+
"eval_runtime": 0.8528,
|
| 876 |
+
"eval_samples_per_second": 469.044,
|
| 877 |
+
"eval_steps_per_second": 4.69,
|
| 878 |
+
"step": 240
|
| 879 |
+
}
|
| 880 |
+
],
|
| 881 |
+
"max_steps": 320,
|
| 882 |
+
"num_train_epochs": 20,
|
| 883 |
+
"total_flos": 175154724864000.0,
|
| 884 |
+
"trial_name": null,
|
| 885 |
+
"trial_params": null
|
| 886 |
+
}
|
scaling_performance/2000/L2/.DS_Store
CHANGED
|
Binary files a/scaling_performance/2000/L2/.DS_Store and b/scaling_performance/2000/L2/.DS_Store differ
|
|
|
scaling_performance/2000/L2/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.5775,
|
| 3 |
+
"test_loss": 0.6842468976974487,
|
| 4 |
+
"test_macro_f1": 0.38194285714285714,
|
| 5 |
+
"test_runtime": 1.0481,
|
| 6 |
+
"test_samples_per_second": 381.652,
|
| 7 |
+
"test_steps_per_second": 3.817
|
| 8 |
+
}
|
scaling_performance/2000/L2/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.02,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.02,
|
| 9 |
+
"hidden_size": 256,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 512,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 2048,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 4,
|
| 16 |
+
"num_hidden_layers": 2,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"problem_type": "single_label_classification",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.28.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
scaling_performance/2000/L2/eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.5775,
|
| 3 |
+
"test_loss": 0.6842468976974487,
|
| 4 |
+
"test_macro_f1": 0.38194285714285714,
|
| 5 |
+
"test_runtime": 1.0481,
|
| 6 |
+
"test_samples_per_second": 381.652,
|
| 7 |
+
"test_steps_per_second": 3.817
|
| 8 |
+
}
|
scaling_performance/2000/L2/trainer_state.json
ADDED
|
@@ -0,0 +1,596 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6842468976974487,
|
| 3 |
+
"best_model_checkpoint": "./models/240626_geneformer_CellClassifier_PM25_Layers2_L2048_B26_LR5e-05_LSlinear_WU600_E20_Oadamw_F0_fold4/checkpoint-128",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"global_step": 160,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 1.6666666666666668e-07,
|
| 13 |
+
"loss": 0.6937,
|
| 14 |
+
"step": 2
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.25,
|
| 18 |
+
"learning_rate": 3.3333333333333335e-07,
|
| 19 |
+
"loss": 0.7018,
|
| 20 |
+
"step": 4
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.38,
|
| 24 |
+
"learning_rate": 5.000000000000001e-07,
|
| 25 |
+
"loss": 0.7031,
|
| 26 |
+
"step": 6
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.5,
|
| 30 |
+
"learning_rate": 6.666666666666667e-07,
|
| 31 |
+
"loss": 0.6949,
|
| 32 |
+
"step": 8
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.62,
|
| 36 |
+
"learning_rate": 8.333333333333333e-07,
|
| 37 |
+
"loss": 0.694,
|
| 38 |
+
"step": 10
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.75,
|
| 42 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 43 |
+
"loss": 0.7025,
|
| 44 |
+
"step": 12
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.88,
|
| 48 |
+
"learning_rate": 1.1666666666666668e-06,
|
| 49 |
+
"loss": 0.6975,
|
| 50 |
+
"step": 14
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 55 |
+
"loss": 0.7029,
|
| 56 |
+
"step": 16
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.0,
|
| 60 |
+
"eval_accuracy": 0.4575,
|
| 61 |
+
"eval_loss": 0.6974406242370605,
|
| 62 |
+
"eval_macro_f1": 0.4533144386710754,
|
| 63 |
+
"eval_runtime": 0.9243,
|
| 64 |
+
"eval_samples_per_second": 432.744,
|
| 65 |
+
"eval_steps_per_second": 4.327,
|
| 66 |
+
"step": 16
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.12,
|
| 70 |
+
"learning_rate": 1.5e-06,
|
| 71 |
+
"loss": 0.6934,
|
| 72 |
+
"step": 18
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.25,
|
| 76 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 77 |
+
"loss": 0.7034,
|
| 78 |
+
"step": 20
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.38,
|
| 82 |
+
"learning_rate": 1.8333333333333335e-06,
|
| 83 |
+
"loss": 0.6945,
|
| 84 |
+
"step": 22
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5,
|
| 88 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 89 |
+
"loss": 0.6894,
|
| 90 |
+
"step": 24
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.62,
|
| 94 |
+
"learning_rate": 2.166666666666667e-06,
|
| 95 |
+
"loss": 0.7029,
|
| 96 |
+
"step": 26
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.75,
|
| 100 |
+
"learning_rate": 2.3333333333333336e-06,
|
| 101 |
+
"loss": 0.7002,
|
| 102 |
+
"step": 28
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.88,
|
| 106 |
+
"learning_rate": 2.5e-06,
|
| 107 |
+
"loss": 0.6943,
|
| 108 |
+
"step": 30
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 2.0,
|
| 112 |
+
"learning_rate": 2.666666666666667e-06,
|
| 113 |
+
"loss": 0.6983,
|
| 114 |
+
"step": 32
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 2.0,
|
| 118 |
+
"eval_accuracy": 0.485,
|
| 119 |
+
"eval_loss": 0.695452094078064,
|
| 120 |
+
"eval_macro_f1": 0.4841746794871795,
|
| 121 |
+
"eval_runtime": 0.9144,
|
| 122 |
+
"eval_samples_per_second": 437.423,
|
| 123 |
+
"eval_steps_per_second": 4.374,
|
| 124 |
+
"step": 32
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.12,
|
| 128 |
+
"learning_rate": 2.8333333333333335e-06,
|
| 129 |
+
"loss": 0.6959,
|
| 130 |
+
"step": 34
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.25,
|
| 134 |
+
"learning_rate": 3e-06,
|
| 135 |
+
"loss": 0.6983,
|
| 136 |
+
"step": 36
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.38,
|
| 140 |
+
"learning_rate": 3.166666666666667e-06,
|
| 141 |
+
"loss": 0.6905,
|
| 142 |
+
"step": 38
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.5,
|
| 146 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 147 |
+
"loss": 0.6929,
|
| 148 |
+
"step": 40
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.62,
|
| 152 |
+
"learning_rate": 3.5000000000000004e-06,
|
| 153 |
+
"loss": 0.694,
|
| 154 |
+
"step": 42
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.75,
|
| 158 |
+
"learning_rate": 3.666666666666667e-06,
|
| 159 |
+
"loss": 0.6945,
|
| 160 |
+
"step": 44
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.88,
|
| 164 |
+
"learning_rate": 3.833333333333334e-06,
|
| 165 |
+
"loss": 0.6967,
|
| 166 |
+
"step": 46
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.0,
|
| 170 |
+
"learning_rate": 4.000000000000001e-06,
|
| 171 |
+
"loss": 0.6968,
|
| 172 |
+
"step": 48
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.495,
|
| 177 |
+
"eval_loss": 0.6920776963233948,
|
| 178 |
+
"eval_macro_f1": 0.47608673098869175,
|
| 179 |
+
"eval_runtime": 0.9068,
|
| 180 |
+
"eval_samples_per_second": 441.113,
|
| 181 |
+
"eval_steps_per_second": 4.411,
|
| 182 |
+
"step": 48
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.12,
|
| 186 |
+
"learning_rate": 4.166666666666667e-06,
|
| 187 |
+
"loss": 0.6908,
|
| 188 |
+
"step": 50
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.25,
|
| 192 |
+
"learning_rate": 4.333333333333334e-06,
|
| 193 |
+
"loss": 0.6944,
|
| 194 |
+
"step": 52
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.38,
|
| 198 |
+
"learning_rate": 4.5e-06,
|
| 199 |
+
"loss": 0.6914,
|
| 200 |
+
"step": 54
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.5,
|
| 204 |
+
"learning_rate": 4.666666666666667e-06,
|
| 205 |
+
"loss": 0.6911,
|
| 206 |
+
"step": 56
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.62,
|
| 210 |
+
"learning_rate": 4.833333333333333e-06,
|
| 211 |
+
"loss": 0.6893,
|
| 212 |
+
"step": 58
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.75,
|
| 216 |
+
"learning_rate": 5e-06,
|
| 217 |
+
"loss": 0.6949,
|
| 218 |
+
"step": 60
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.88,
|
| 222 |
+
"learning_rate": 5.166666666666667e-06,
|
| 223 |
+
"loss": 0.6949,
|
| 224 |
+
"step": 62
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.0,
|
| 228 |
+
"learning_rate": 5.333333333333334e-06,
|
| 229 |
+
"loss": 0.6868,
|
| 230 |
+
"step": 64
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.0,
|
| 234 |
+
"eval_accuracy": 0.5325,
|
| 235 |
+
"eval_loss": 0.6893474459648132,
|
| 236 |
+
"eval_macro_f1": 0.46399524188289587,
|
| 237 |
+
"eval_runtime": 0.9384,
|
| 238 |
+
"eval_samples_per_second": 426.254,
|
| 239 |
+
"eval_steps_per_second": 4.263,
|
| 240 |
+
"step": 64
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.12,
|
| 244 |
+
"learning_rate": 5.500000000000001e-06,
|
| 245 |
+
"loss": 0.6984,
|
| 246 |
+
"step": 66
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.25,
|
| 250 |
+
"learning_rate": 5.666666666666667e-06,
|
| 251 |
+
"loss": 0.6822,
|
| 252 |
+
"step": 68
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.38,
|
| 256 |
+
"learning_rate": 5.833333333333334e-06,
|
| 257 |
+
"loss": 0.6864,
|
| 258 |
+
"step": 70
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.5,
|
| 262 |
+
"learning_rate": 6e-06,
|
| 263 |
+
"loss": 0.7002,
|
| 264 |
+
"step": 72
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.62,
|
| 268 |
+
"learning_rate": 6.166666666666667e-06,
|
| 269 |
+
"loss": 0.6897,
|
| 270 |
+
"step": 74
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 4.75,
|
| 274 |
+
"learning_rate": 6.333333333333334e-06,
|
| 275 |
+
"loss": 0.6903,
|
| 276 |
+
"step": 76
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 4.88,
|
| 280 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 281 |
+
"loss": 0.6941,
|
| 282 |
+
"step": 78
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 5.0,
|
| 286 |
+
"learning_rate": 6.666666666666667e-06,
|
| 287 |
+
"loss": 0.6779,
|
| 288 |
+
"step": 80
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.0,
|
| 292 |
+
"eval_accuracy": 0.575,
|
| 293 |
+
"eval_loss": 0.6872670650482178,
|
| 294 |
+
"eval_macro_f1": 0.41748903508771934,
|
| 295 |
+
"eval_runtime": 0.928,
|
| 296 |
+
"eval_samples_per_second": 431.013,
|
| 297 |
+
"eval_steps_per_second": 4.31,
|
| 298 |
+
"step": 80
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.12,
|
| 302 |
+
"learning_rate": 6.833333333333333e-06,
|
| 303 |
+
"loss": 0.7031,
|
| 304 |
+
"step": 82
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.25,
|
| 308 |
+
"learning_rate": 7.000000000000001e-06,
|
| 309 |
+
"loss": 0.6892,
|
| 310 |
+
"step": 84
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.38,
|
| 314 |
+
"learning_rate": 7.166666666666667e-06,
|
| 315 |
+
"loss": 0.6823,
|
| 316 |
+
"step": 86
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.5,
|
| 320 |
+
"learning_rate": 7.333333333333334e-06,
|
| 321 |
+
"loss": 0.6936,
|
| 322 |
+
"step": 88
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 5.62,
|
| 326 |
+
"learning_rate": 7.5e-06,
|
| 327 |
+
"loss": 0.6907,
|
| 328 |
+
"step": 90
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.75,
|
| 332 |
+
"learning_rate": 7.666666666666667e-06,
|
| 333 |
+
"loss": 0.681,
|
| 334 |
+
"step": 92
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 5.88,
|
| 338 |
+
"learning_rate": 7.833333333333333e-06,
|
| 339 |
+
"loss": 0.6965,
|
| 340 |
+
"step": 94
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 6.0,
|
| 344 |
+
"learning_rate": 8.000000000000001e-06,
|
| 345 |
+
"loss": 0.6675,
|
| 346 |
+
"step": 96
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.0,
|
| 350 |
+
"eval_accuracy": 0.58,
|
| 351 |
+
"eval_loss": 0.6864795088768005,
|
| 352 |
+
"eval_macro_f1": 0.4159365874009178,
|
| 353 |
+
"eval_runtime": 0.9734,
|
| 354 |
+
"eval_samples_per_second": 410.925,
|
| 355 |
+
"eval_steps_per_second": 4.109,
|
| 356 |
+
"step": 96
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.12,
|
| 360 |
+
"learning_rate": 8.166666666666668e-06,
|
| 361 |
+
"loss": 0.6914,
|
| 362 |
+
"step": 98
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.25,
|
| 366 |
+
"learning_rate": 8.333333333333334e-06,
|
| 367 |
+
"loss": 0.6865,
|
| 368 |
+
"step": 100
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.38,
|
| 372 |
+
"learning_rate": 8.500000000000002e-06,
|
| 373 |
+
"loss": 0.6801,
|
| 374 |
+
"step": 102
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.5,
|
| 378 |
+
"learning_rate": 8.666666666666668e-06,
|
| 379 |
+
"loss": 0.699,
|
| 380 |
+
"step": 104
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.62,
|
| 384 |
+
"learning_rate": 8.833333333333334e-06,
|
| 385 |
+
"loss": 0.6816,
|
| 386 |
+
"step": 106
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 6.75,
|
| 390 |
+
"learning_rate": 9e-06,
|
| 391 |
+
"loss": 0.6803,
|
| 392 |
+
"step": 108
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 6.88,
|
| 396 |
+
"learning_rate": 9.166666666666666e-06,
|
| 397 |
+
"loss": 0.6934,
|
| 398 |
+
"step": 110
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.0,
|
| 402 |
+
"learning_rate": 9.333333333333334e-06,
|
| 403 |
+
"loss": 0.6765,
|
| 404 |
+
"step": 112
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.0,
|
| 408 |
+
"eval_accuracy": 0.5775,
|
| 409 |
+
"eval_loss": 0.6850780248641968,
|
| 410 |
+
"eval_macro_f1": 0.38696484116404134,
|
| 411 |
+
"eval_runtime": 0.9277,
|
| 412 |
+
"eval_samples_per_second": 431.159,
|
| 413 |
+
"eval_steps_per_second": 4.312,
|
| 414 |
+
"step": 112
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.12,
|
| 418 |
+
"learning_rate": 9.5e-06,
|
| 419 |
+
"loss": 0.6959,
|
| 420 |
+
"step": 114
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.25,
|
| 424 |
+
"learning_rate": 9.666666666666667e-06,
|
| 425 |
+
"loss": 0.6815,
|
| 426 |
+
"step": 116
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.38,
|
| 430 |
+
"learning_rate": 9.833333333333333e-06,
|
| 431 |
+
"loss": 0.6763,
|
| 432 |
+
"step": 118
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.5,
|
| 436 |
+
"learning_rate": 1e-05,
|
| 437 |
+
"loss": 0.698,
|
| 438 |
+
"step": 120
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 7.62,
|
| 442 |
+
"learning_rate": 1.0166666666666667e-05,
|
| 443 |
+
"loss": 0.6784,
|
| 444 |
+
"step": 122
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.75,
|
| 448 |
+
"learning_rate": 1.0333333333333333e-05,
|
| 449 |
+
"loss": 0.6672,
|
| 450 |
+
"step": 124
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 7.88,
|
| 454 |
+
"learning_rate": 1.05e-05,
|
| 455 |
+
"loss": 0.7066,
|
| 456 |
+
"step": 126
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 8.0,
|
| 460 |
+
"learning_rate": 1.0666666666666667e-05,
|
| 461 |
+
"loss": 0.6902,
|
| 462 |
+
"step": 128
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.0,
|
| 466 |
+
"eval_accuracy": 0.5775,
|
| 467 |
+
"eval_loss": 0.6842468976974487,
|
| 468 |
+
"eval_macro_f1": 0.38194285714285714,
|
| 469 |
+
"eval_runtime": 0.884,
|
| 470 |
+
"eval_samples_per_second": 452.506,
|
| 471 |
+
"eval_steps_per_second": 4.525,
|
| 472 |
+
"step": 128
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.12,
|
| 476 |
+
"learning_rate": 1.0833333333333334e-05,
|
| 477 |
+
"loss": 0.6862,
|
| 478 |
+
"step": 130
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 8.25,
|
| 482 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 483 |
+
"loss": 0.6784,
|
| 484 |
+
"step": 132
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 8.38,
|
| 488 |
+
"learning_rate": 1.1166666666666668e-05,
|
| 489 |
+
"loss": 0.6774,
|
| 490 |
+
"step": 134
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.5,
|
| 494 |
+
"learning_rate": 1.1333333333333334e-05,
|
| 495 |
+
"loss": 0.7004,
|
| 496 |
+
"step": 136
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.62,
|
| 500 |
+
"learning_rate": 1.1500000000000002e-05,
|
| 501 |
+
"loss": 0.6755,
|
| 502 |
+
"step": 138
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 8.75,
|
| 506 |
+
"learning_rate": 1.1666666666666668e-05,
|
| 507 |
+
"loss": 0.68,
|
| 508 |
+
"step": 140
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 8.88,
|
| 512 |
+
"learning_rate": 1.1833333333333334e-05,
|
| 513 |
+
"loss": 0.7023,
|
| 514 |
+
"step": 142
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 9.0,
|
| 518 |
+
"learning_rate": 1.2e-05,
|
| 519 |
+
"loss": 0.6802,
|
| 520 |
+
"step": 144
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 9.0,
|
| 524 |
+
"eval_accuracy": 0.5775,
|
| 525 |
+
"eval_loss": 0.6846153140068054,
|
| 526 |
+
"eval_macro_f1": 0.38696484116404134,
|
| 527 |
+
"eval_runtime": 0.9145,
|
| 528 |
+
"eval_samples_per_second": 437.397,
|
| 529 |
+
"eval_steps_per_second": 4.374,
|
| 530 |
+
"step": 144
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 9.12,
|
| 534 |
+
"learning_rate": 1.2166666666666668e-05,
|
| 535 |
+
"loss": 0.6962,
|
| 536 |
+
"step": 146
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.25,
|
| 540 |
+
"learning_rate": 1.2333333333333334e-05,
|
| 541 |
+
"loss": 0.6717,
|
| 542 |
+
"step": 148
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 9.38,
|
| 546 |
+
"learning_rate": 1.25e-05,
|
| 547 |
+
"loss": 0.6783,
|
| 548 |
+
"step": 150
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 9.5,
|
| 552 |
+
"learning_rate": 1.2666666666666668e-05,
|
| 553 |
+
"loss": 0.7028,
|
| 554 |
+
"step": 152
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 9.62,
|
| 558 |
+
"learning_rate": 1.2833333333333333e-05,
|
| 559 |
+
"loss": 0.683,
|
| 560 |
+
"step": 154
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 9.75,
|
| 564 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 565 |
+
"loss": 0.6835,
|
| 566 |
+
"step": 156
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 9.88,
|
| 570 |
+
"learning_rate": 1.3166666666666665e-05,
|
| 571 |
+
"loss": 0.687,
|
| 572 |
+
"step": 158
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 10.0,
|
| 576 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 577 |
+
"loss": 0.6693,
|
| 578 |
+
"step": 160
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.0,
|
| 582 |
+
"eval_accuracy": 0.5875,
|
| 583 |
+
"eval_loss": 0.6855114698410034,
|
| 584 |
+
"eval_macro_f1": 0.4199077125906394,
|
| 585 |
+
"eval_runtime": 1.013,
|
| 586 |
+
"eval_samples_per_second": 394.875,
|
| 587 |
+
"eval_steps_per_second": 3.949,
|
| 588 |
+
"step": 160
|
| 589 |
+
}
|
| 590 |
+
],
|
| 591 |
+
"max_steps": 320,
|
| 592 |
+
"num_train_epochs": 20,
|
| 593 |
+
"total_flos": 220402679808000.0,
|
| 594 |
+
"trial_name": null,
|
| 595 |
+
"trial_params": null
|
| 596 |
+
}
|
scaling_performance/2000/L4/.DS_Store
CHANGED
|
Binary files a/scaling_performance/2000/L4/.DS_Store and b/scaling_performance/2000/L4/.DS_Store differ
|
|
|
scaling_performance/2000/L4/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.6975,
|
| 3 |
+
"test_loss": 0.5687375068664551,
|
| 4 |
+
"test_macro_f1": 0.6942881145029971,
|
| 5 |
+
"test_runtime": 1.6145,
|
| 6 |
+
"test_samples_per_second": 247.752,
|
| 7 |
+
"test_steps_per_second": 2.478
|
| 8 |
+
}
|
scaling_performance/2000/L4/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.02,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.02,
|
| 9 |
+
"hidden_size": 256,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 512,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 2048,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 4,
|
| 16 |
+
"num_hidden_layers": 4,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"problem_type": "single_label_classification",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.28.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
scaling_performance/2000/L4/eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.6975,
|
| 3 |
+
"test_loss": 0.5687375068664551,
|
| 4 |
+
"test_macro_f1": 0.6942881145029971,
|
| 5 |
+
"test_runtime": 1.6145,
|
| 6 |
+
"test_samples_per_second": 247.752,
|
| 7 |
+
"test_steps_per_second": 2.478
|
| 8 |
+
}
|
scaling_performance/2000/L4/trainer_state.json
ADDED
|
@@ -0,0 +1,596 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6825920343399048,
|
| 3 |
+
"best_model_checkpoint": "./models/240626_geneformer_CellClassifier_PM25_Layers4_L2048_B26_LR5e-05_LSlinear_WU600_E20_Oadamw_F0_fold4/checkpoint-128",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"global_step": 160,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 1.6666666666666668e-07,
|
| 13 |
+
"loss": 0.6905,
|
| 14 |
+
"step": 2
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.25,
|
| 18 |
+
"learning_rate": 3.3333333333333335e-07,
|
| 19 |
+
"loss": 0.7029,
|
| 20 |
+
"step": 4
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.38,
|
| 24 |
+
"learning_rate": 5.000000000000001e-07,
|
| 25 |
+
"loss": 0.6978,
|
| 26 |
+
"step": 6
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.5,
|
| 30 |
+
"learning_rate": 6.666666666666667e-07,
|
| 31 |
+
"loss": 0.6968,
|
| 32 |
+
"step": 8
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.62,
|
| 36 |
+
"learning_rate": 8.333333333333333e-07,
|
| 37 |
+
"loss": 0.6954,
|
| 38 |
+
"step": 10
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.75,
|
| 42 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 43 |
+
"loss": 0.7,
|
| 44 |
+
"step": 12
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.88,
|
| 48 |
+
"learning_rate": 1.1666666666666668e-06,
|
| 49 |
+
"loss": 0.697,
|
| 50 |
+
"step": 14
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 55 |
+
"loss": 0.6996,
|
| 56 |
+
"step": 16
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.0,
|
| 60 |
+
"eval_accuracy": 0.5,
|
| 61 |
+
"eval_loss": 0.6958171129226685,
|
| 62 |
+
"eval_macro_f1": 0.493298877657014,
|
| 63 |
+
"eval_runtime": 1.1509,
|
| 64 |
+
"eval_samples_per_second": 347.552,
|
| 65 |
+
"eval_steps_per_second": 3.476,
|
| 66 |
+
"step": 16
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.12,
|
| 70 |
+
"learning_rate": 1.5e-06,
|
| 71 |
+
"loss": 0.69,
|
| 72 |
+
"step": 18
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.25,
|
| 76 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 77 |
+
"loss": 0.7017,
|
| 78 |
+
"step": 20
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.38,
|
| 82 |
+
"learning_rate": 1.8333333333333335e-06,
|
| 83 |
+
"loss": 0.6939,
|
| 84 |
+
"step": 22
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5,
|
| 88 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 89 |
+
"loss": 0.6928,
|
| 90 |
+
"step": 24
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.62,
|
| 94 |
+
"learning_rate": 2.166666666666667e-06,
|
| 95 |
+
"loss": 0.699,
|
| 96 |
+
"step": 26
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.75,
|
| 100 |
+
"learning_rate": 2.3333333333333336e-06,
|
| 101 |
+
"loss": 0.6927,
|
| 102 |
+
"step": 28
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.88,
|
| 106 |
+
"learning_rate": 2.5e-06,
|
| 107 |
+
"loss": 0.6986,
|
| 108 |
+
"step": 30
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 2.0,
|
| 112 |
+
"learning_rate": 2.666666666666667e-06,
|
| 113 |
+
"loss": 0.6999,
|
| 114 |
+
"step": 32
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 2.0,
|
| 118 |
+
"eval_accuracy": 0.51,
|
| 119 |
+
"eval_loss": 0.6933969259262085,
|
| 120 |
+
"eval_macro_f1": 0.5090057366166487,
|
| 121 |
+
"eval_runtime": 1.199,
|
| 122 |
+
"eval_samples_per_second": 333.601,
|
| 123 |
+
"eval_steps_per_second": 3.336,
|
| 124 |
+
"step": 32
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.12,
|
| 128 |
+
"learning_rate": 2.8333333333333335e-06,
|
| 129 |
+
"loss": 0.6941,
|
| 130 |
+
"step": 34
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.25,
|
| 134 |
+
"learning_rate": 3e-06,
|
| 135 |
+
"loss": 0.6918,
|
| 136 |
+
"step": 36
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.38,
|
| 140 |
+
"learning_rate": 3.166666666666667e-06,
|
| 141 |
+
"loss": 0.6899,
|
| 142 |
+
"step": 38
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.5,
|
| 146 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 147 |
+
"loss": 0.6906,
|
| 148 |
+
"step": 40
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.62,
|
| 152 |
+
"learning_rate": 3.5000000000000004e-06,
|
| 153 |
+
"loss": 0.6914,
|
| 154 |
+
"step": 42
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.75,
|
| 158 |
+
"learning_rate": 3.666666666666667e-06,
|
| 159 |
+
"loss": 0.6919,
|
| 160 |
+
"step": 44
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.88,
|
| 164 |
+
"learning_rate": 3.833333333333334e-06,
|
| 165 |
+
"loss": 0.7009,
|
| 166 |
+
"step": 46
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.0,
|
| 170 |
+
"learning_rate": 4.000000000000001e-06,
|
| 171 |
+
"loss": 0.6881,
|
| 172 |
+
"step": 48
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.545,
|
| 177 |
+
"eval_loss": 0.6895008087158203,
|
| 178 |
+
"eval_macro_f1": 0.4814814814814815,
|
| 179 |
+
"eval_runtime": 1.1808,
|
| 180 |
+
"eval_samples_per_second": 338.766,
|
| 181 |
+
"eval_steps_per_second": 3.388,
|
| 182 |
+
"step": 48
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.12,
|
| 186 |
+
"learning_rate": 4.166666666666667e-06,
|
| 187 |
+
"loss": 0.6938,
|
| 188 |
+
"step": 50
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.25,
|
| 192 |
+
"learning_rate": 4.333333333333334e-06,
|
| 193 |
+
"loss": 0.6876,
|
| 194 |
+
"step": 52
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.38,
|
| 198 |
+
"learning_rate": 4.5e-06,
|
| 199 |
+
"loss": 0.6855,
|
| 200 |
+
"step": 54
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.5,
|
| 204 |
+
"learning_rate": 4.666666666666667e-06,
|
| 205 |
+
"loss": 0.6973,
|
| 206 |
+
"step": 56
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.62,
|
| 210 |
+
"learning_rate": 4.833333333333333e-06,
|
| 211 |
+
"loss": 0.69,
|
| 212 |
+
"step": 58
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.75,
|
| 216 |
+
"learning_rate": 5e-06,
|
| 217 |
+
"loss": 0.6902,
|
| 218 |
+
"step": 60
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.88,
|
| 222 |
+
"learning_rate": 5.166666666666667e-06,
|
| 223 |
+
"loss": 0.6913,
|
| 224 |
+
"step": 62
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.0,
|
| 228 |
+
"learning_rate": 5.333333333333334e-06,
|
| 229 |
+
"loss": 0.6799,
|
| 230 |
+
"step": 64
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.0,
|
| 234 |
+
"eval_accuracy": 0.5625,
|
| 235 |
+
"eval_loss": 0.6866650581359863,
|
| 236 |
+
"eval_macro_f1": 0.41072976909025094,
|
| 237 |
+
"eval_runtime": 1.1114,
|
| 238 |
+
"eval_samples_per_second": 359.916,
|
| 239 |
+
"eval_steps_per_second": 3.599,
|
| 240 |
+
"step": 64
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.12,
|
| 244 |
+
"learning_rate": 5.500000000000001e-06,
|
| 245 |
+
"loss": 0.6976,
|
| 246 |
+
"step": 66
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.25,
|
| 250 |
+
"learning_rate": 5.666666666666667e-06,
|
| 251 |
+
"loss": 0.6823,
|
| 252 |
+
"step": 68
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.38,
|
| 256 |
+
"learning_rate": 5.833333333333334e-06,
|
| 257 |
+
"loss": 0.6863,
|
| 258 |
+
"step": 70
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.5,
|
| 262 |
+
"learning_rate": 6e-06,
|
| 263 |
+
"loss": 0.6955,
|
| 264 |
+
"step": 72
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.62,
|
| 268 |
+
"learning_rate": 6.166666666666667e-06,
|
| 269 |
+
"loss": 0.6842,
|
| 270 |
+
"step": 74
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 4.75,
|
| 274 |
+
"learning_rate": 6.333333333333334e-06,
|
| 275 |
+
"loss": 0.6873,
|
| 276 |
+
"step": 76
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 4.88,
|
| 280 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 281 |
+
"loss": 0.6987,
|
| 282 |
+
"step": 78
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 5.0,
|
| 286 |
+
"learning_rate": 6.666666666666667e-06,
|
| 287 |
+
"loss": 0.6779,
|
| 288 |
+
"step": 80
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.0,
|
| 292 |
+
"eval_accuracy": 0.5675,
|
| 293 |
+
"eval_loss": 0.6847481727600098,
|
| 294 |
+
"eval_macro_f1": 0.3673142857142857,
|
| 295 |
+
"eval_runtime": 1.2206,
|
| 296 |
+
"eval_samples_per_second": 327.716,
|
| 297 |
+
"eval_steps_per_second": 3.277,
|
| 298 |
+
"step": 80
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.12,
|
| 302 |
+
"learning_rate": 6.833333333333333e-06,
|
| 303 |
+
"loss": 0.695,
|
| 304 |
+
"step": 82
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.25,
|
| 308 |
+
"learning_rate": 7.000000000000001e-06,
|
| 309 |
+
"loss": 0.689,
|
| 310 |
+
"step": 84
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.38,
|
| 314 |
+
"learning_rate": 7.166666666666667e-06,
|
| 315 |
+
"loss": 0.6814,
|
| 316 |
+
"step": 86
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.5,
|
| 320 |
+
"learning_rate": 7.333333333333334e-06,
|
| 321 |
+
"loss": 0.693,
|
| 322 |
+
"step": 88
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 5.62,
|
| 326 |
+
"learning_rate": 7.5e-06,
|
| 327 |
+
"loss": 0.6876,
|
| 328 |
+
"step": 90
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.75,
|
| 332 |
+
"learning_rate": 7.666666666666667e-06,
|
| 333 |
+
"loss": 0.6822,
|
| 334 |
+
"step": 92
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 5.88,
|
| 338 |
+
"learning_rate": 7.833333333333333e-06,
|
| 339 |
+
"loss": 0.6933,
|
| 340 |
+
"step": 94
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 6.0,
|
| 344 |
+
"learning_rate": 8.000000000000001e-06,
|
| 345 |
+
"loss": 0.6579,
|
| 346 |
+
"step": 96
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.0,
|
| 350 |
+
"eval_accuracy": 0.57,
|
| 351 |
+
"eval_loss": 0.6842648386955261,
|
| 352 |
+
"eval_macro_f1": 0.3630573248407643,
|
| 353 |
+
"eval_runtime": 1.1578,
|
| 354 |
+
"eval_samples_per_second": 345.493,
|
| 355 |
+
"eval_steps_per_second": 3.455,
|
| 356 |
+
"step": 96
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.12,
|
| 360 |
+
"learning_rate": 8.166666666666668e-06,
|
| 361 |
+
"loss": 0.6891,
|
| 362 |
+
"step": 98
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.25,
|
| 366 |
+
"learning_rate": 8.333333333333334e-06,
|
| 367 |
+
"loss": 0.6925,
|
| 368 |
+
"step": 100
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.38,
|
| 372 |
+
"learning_rate": 8.500000000000002e-06,
|
| 373 |
+
"loss": 0.6739,
|
| 374 |
+
"step": 102
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.5,
|
| 378 |
+
"learning_rate": 8.666666666666668e-06,
|
| 379 |
+
"loss": 0.6974,
|
| 380 |
+
"step": 104
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.62,
|
| 384 |
+
"learning_rate": 8.833333333333334e-06,
|
| 385 |
+
"loss": 0.6804,
|
| 386 |
+
"step": 106
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 6.75,
|
| 390 |
+
"learning_rate": 9e-06,
|
| 391 |
+
"loss": 0.6762,
|
| 392 |
+
"step": 108
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 6.88,
|
| 396 |
+
"learning_rate": 9.166666666666666e-06,
|
| 397 |
+
"loss": 0.7002,
|
| 398 |
+
"step": 110
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.0,
|
| 402 |
+
"learning_rate": 9.333333333333334e-06,
|
| 403 |
+
"loss": 0.669,
|
| 404 |
+
"step": 112
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.0,
|
| 408 |
+
"eval_accuracy": 0.5725,
|
| 409 |
+
"eval_loss": 0.683117687702179,
|
| 410 |
+
"eval_macro_f1": 0.3640699523052464,
|
| 411 |
+
"eval_runtime": 1.117,
|
| 412 |
+
"eval_samples_per_second": 358.09,
|
| 413 |
+
"eval_steps_per_second": 3.581,
|
| 414 |
+
"step": 112
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.12,
|
| 418 |
+
"learning_rate": 9.5e-06,
|
| 419 |
+
"loss": 0.6975,
|
| 420 |
+
"step": 114
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.25,
|
| 424 |
+
"learning_rate": 9.666666666666667e-06,
|
| 425 |
+
"loss": 0.6847,
|
| 426 |
+
"step": 116
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.38,
|
| 430 |
+
"learning_rate": 9.833333333333333e-06,
|
| 431 |
+
"loss": 0.6715,
|
| 432 |
+
"step": 118
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.5,
|
| 436 |
+
"learning_rate": 1e-05,
|
| 437 |
+
"loss": 0.7029,
|
| 438 |
+
"step": 120
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 7.62,
|
| 442 |
+
"learning_rate": 1.0166666666666667e-05,
|
| 443 |
+
"loss": 0.6696,
|
| 444 |
+
"step": 122
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.75,
|
| 448 |
+
"learning_rate": 1.0333333333333333e-05,
|
| 449 |
+
"loss": 0.6602,
|
| 450 |
+
"step": 124
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 7.88,
|
| 454 |
+
"learning_rate": 1.05e-05,
|
| 455 |
+
"loss": 0.7126,
|
| 456 |
+
"step": 126
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 8.0,
|
| 460 |
+
"learning_rate": 1.0666666666666667e-05,
|
| 461 |
+
"loss": 0.6814,
|
| 462 |
+
"step": 128
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.0,
|
| 466 |
+
"eval_accuracy": 0.5775,
|
| 467 |
+
"eval_loss": 0.6825920343399048,
|
| 468 |
+
"eval_macro_f1": 0.36608557844690964,
|
| 469 |
+
"eval_runtime": 1.1048,
|
| 470 |
+
"eval_samples_per_second": 362.052,
|
| 471 |
+
"eval_steps_per_second": 3.621,
|
| 472 |
+
"step": 128
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.12,
|
| 476 |
+
"learning_rate": 1.0833333333333334e-05,
|
| 477 |
+
"loss": 0.6947,
|
| 478 |
+
"step": 130
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 8.25,
|
| 482 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 483 |
+
"loss": 0.6809,
|
| 484 |
+
"step": 132
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 8.38,
|
| 488 |
+
"learning_rate": 1.1166666666666668e-05,
|
| 489 |
+
"loss": 0.6765,
|
| 490 |
+
"step": 134
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.5,
|
| 494 |
+
"learning_rate": 1.1333333333333334e-05,
|
| 495 |
+
"loss": 0.6964,
|
| 496 |
+
"step": 136
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.62,
|
| 500 |
+
"learning_rate": 1.1500000000000002e-05,
|
| 501 |
+
"loss": 0.6771,
|
| 502 |
+
"step": 138
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 8.75,
|
| 506 |
+
"learning_rate": 1.1666666666666668e-05,
|
| 507 |
+
"loss": 0.667,
|
| 508 |
+
"step": 140
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 8.88,
|
| 512 |
+
"learning_rate": 1.1833333333333334e-05,
|
| 513 |
+
"loss": 0.7035,
|
| 514 |
+
"step": 142
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 9.0,
|
| 518 |
+
"learning_rate": 1.2e-05,
|
| 519 |
+
"loss": 0.6773,
|
| 520 |
+
"step": 144
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 9.0,
|
| 524 |
+
"eval_accuracy": 0.5725,
|
| 525 |
+
"eval_loss": 0.6831978559494019,
|
| 526 |
+
"eval_macro_f1": 0.3640699523052464,
|
| 527 |
+
"eval_runtime": 1.1552,
|
| 528 |
+
"eval_samples_per_second": 346.251,
|
| 529 |
+
"eval_steps_per_second": 3.463,
|
| 530 |
+
"step": 144
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 9.12,
|
| 534 |
+
"learning_rate": 1.2166666666666668e-05,
|
| 535 |
+
"loss": 0.6942,
|
| 536 |
+
"step": 146
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.25,
|
| 540 |
+
"learning_rate": 1.2333333333333334e-05,
|
| 541 |
+
"loss": 0.6726,
|
| 542 |
+
"step": 148
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 9.38,
|
| 546 |
+
"learning_rate": 1.25e-05,
|
| 547 |
+
"loss": 0.6732,
|
| 548 |
+
"step": 150
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 9.5,
|
| 552 |
+
"learning_rate": 1.2666666666666668e-05,
|
| 553 |
+
"loss": 0.7068,
|
| 554 |
+
"step": 152
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 9.62,
|
| 558 |
+
"learning_rate": 1.2833333333333333e-05,
|
| 559 |
+
"loss": 0.6755,
|
| 560 |
+
"step": 154
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 9.75,
|
| 564 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 565 |
+
"loss": 0.6788,
|
| 566 |
+
"step": 156
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 9.88,
|
| 570 |
+
"learning_rate": 1.3166666666666665e-05,
|
| 571 |
+
"loss": 0.691,
|
| 572 |
+
"step": 158
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 10.0,
|
| 576 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 577 |
+
"loss": 0.669,
|
| 578 |
+
"step": 160
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.0,
|
| 582 |
+
"eval_accuracy": 0.565,
|
| 583 |
+
"eval_loss": 0.6841627359390259,
|
| 584 |
+
"eval_macro_f1": 0.3950771798080934,
|
| 585 |
+
"eval_runtime": 1.123,
|
| 586 |
+
"eval_samples_per_second": 356.196,
|
| 587 |
+
"eval_steps_per_second": 3.562,
|
| 588 |
+
"step": 160
|
| 589 |
+
}
|
| 590 |
+
],
|
| 591 |
+
"max_steps": 320,
|
| 592 |
+
"num_train_epochs": 20,
|
| 593 |
+
"total_flos": 427668406272000.0,
|
| 594 |
+
"trial_name": null,
|
| 595 |
+
"trial_params": null
|
| 596 |
+
}
|
scaling_performance/2000/L6/.DS_Store
CHANGED
|
Binary files a/scaling_performance/2000/L6/.DS_Store and b/scaling_performance/2000/L6/.DS_Store differ
|
|
|
scaling_performance/2000/L6/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.5725,
|
| 3 |
+
"test_loss": 0.6847579479217529,
|
| 4 |
+
"test_macro_f1": 0.3746285714285715,
|
| 5 |
+
"test_runtime": 1.2641,
|
| 6 |
+
"test_samples_per_second": 316.42,
|
| 7 |
+
"test_steps_per_second": 3.164
|
| 8 |
+
}
|
scaling_performance/2000/L6/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.02,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "gelu",
|
| 8 |
+
"hidden_dropout_prob": 0.02,
|
| 9 |
+
"hidden_size": 256,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 512,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 2048,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 4,
|
| 16 |
+
"num_hidden_layers": 6,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"problem_type": "single_label_classification",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.28.0",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 30522
|
| 25 |
+
}
|
scaling_performance/2000/L6/eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"test_accuracy": 0.5725,
|
| 3 |
+
"test_loss": 0.6847579479217529,
|
| 4 |
+
"test_macro_f1": 0.3746285714285715,
|
| 5 |
+
"test_runtime": 1.2641,
|
| 6 |
+
"test_samples_per_second": 316.42,
|
| 7 |
+
"test_steps_per_second": 3.164
|
| 8 |
+
}
|
scaling_performance/2000/L6/trainer_state.json
ADDED
|
@@ -0,0 +1,596 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6847579479217529,
|
| 3 |
+
"best_model_checkpoint": "./models/240626_geneformer_CellClassifier_PM25_Layers6_L2048_B26_LR5e-05_LSlinear_WU600_E20_Oadamw_F0_fold4/checkpoint-128",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"global_step": 160,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.12,
|
| 12 |
+
"learning_rate": 1.6666666666666668e-07,
|
| 13 |
+
"loss": 0.6906,
|
| 14 |
+
"step": 2
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.25,
|
| 18 |
+
"learning_rate": 3.3333333333333335e-07,
|
| 19 |
+
"loss": 0.702,
|
| 20 |
+
"step": 4
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.38,
|
| 24 |
+
"learning_rate": 5.000000000000001e-07,
|
| 25 |
+
"loss": 0.7029,
|
| 26 |
+
"step": 6
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.5,
|
| 30 |
+
"learning_rate": 6.666666666666667e-07,
|
| 31 |
+
"loss": 0.6979,
|
| 32 |
+
"step": 8
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.62,
|
| 36 |
+
"learning_rate": 8.333333333333333e-07,
|
| 37 |
+
"loss": 0.7027,
|
| 38 |
+
"step": 10
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.75,
|
| 42 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 43 |
+
"loss": 0.7037,
|
| 44 |
+
"step": 12
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.88,
|
| 48 |
+
"learning_rate": 1.1666666666666668e-06,
|
| 49 |
+
"loss": 0.6946,
|
| 50 |
+
"step": 14
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.0,
|
| 54 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 55 |
+
"loss": 0.6957,
|
| 56 |
+
"step": 16
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.0,
|
| 60 |
+
"eval_accuracy": 0.44,
|
| 61 |
+
"eval_loss": 0.6994468569755554,
|
| 62 |
+
"eval_macro_f1": 0.4141646615754786,
|
| 63 |
+
"eval_runtime": 1.3238,
|
| 64 |
+
"eval_samples_per_second": 302.171,
|
| 65 |
+
"eval_steps_per_second": 3.022,
|
| 66 |
+
"step": 16
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.12,
|
| 70 |
+
"learning_rate": 1.5e-06,
|
| 71 |
+
"loss": 0.6918,
|
| 72 |
+
"step": 18
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 1.25,
|
| 76 |
+
"learning_rate": 1.6666666666666667e-06,
|
| 77 |
+
"loss": 0.7014,
|
| 78 |
+
"step": 20
|
| 79 |
+
},
|
| 80 |
+
{
|
| 81 |
+
"epoch": 1.38,
|
| 82 |
+
"learning_rate": 1.8333333333333335e-06,
|
| 83 |
+
"loss": 0.6956,
|
| 84 |
+
"step": 22
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5,
|
| 88 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 89 |
+
"loss": 0.6929,
|
| 90 |
+
"step": 24
|
| 91 |
+
},
|
| 92 |
+
{
|
| 93 |
+
"epoch": 1.62,
|
| 94 |
+
"learning_rate": 2.166666666666667e-06,
|
| 95 |
+
"loss": 0.698,
|
| 96 |
+
"step": 26
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 1.75,
|
| 100 |
+
"learning_rate": 2.3333333333333336e-06,
|
| 101 |
+
"loss": 0.6998,
|
| 102 |
+
"step": 28
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.88,
|
| 106 |
+
"learning_rate": 2.5e-06,
|
| 107 |
+
"loss": 0.6942,
|
| 108 |
+
"step": 30
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 2.0,
|
| 112 |
+
"learning_rate": 2.666666666666667e-06,
|
| 113 |
+
"loss": 0.6985,
|
| 114 |
+
"step": 32
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 2.0,
|
| 118 |
+
"eval_accuracy": 0.475,
|
| 119 |
+
"eval_loss": 0.6957302093505859,
|
| 120 |
+
"eval_macro_f1": 0.4745270743669302,
|
| 121 |
+
"eval_runtime": 1.4915,
|
| 122 |
+
"eval_samples_per_second": 268.182,
|
| 123 |
+
"eval_steps_per_second": 2.682,
|
| 124 |
+
"step": 32
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 2.12,
|
| 128 |
+
"learning_rate": 2.8333333333333335e-06,
|
| 129 |
+
"loss": 0.6932,
|
| 130 |
+
"step": 34
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.25,
|
| 134 |
+
"learning_rate": 3e-06,
|
| 135 |
+
"loss": 0.6924,
|
| 136 |
+
"step": 36
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 2.38,
|
| 140 |
+
"learning_rate": 3.166666666666667e-06,
|
| 141 |
+
"loss": 0.694,
|
| 142 |
+
"step": 38
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 2.5,
|
| 146 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 147 |
+
"loss": 0.6913,
|
| 148 |
+
"step": 40
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 2.62,
|
| 152 |
+
"learning_rate": 3.5000000000000004e-06,
|
| 153 |
+
"loss": 0.6875,
|
| 154 |
+
"step": 42
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 2.75,
|
| 158 |
+
"learning_rate": 3.666666666666667e-06,
|
| 159 |
+
"loss": 0.6913,
|
| 160 |
+
"step": 44
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.88,
|
| 164 |
+
"learning_rate": 3.833333333333334e-06,
|
| 165 |
+
"loss": 0.6968,
|
| 166 |
+
"step": 46
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.0,
|
| 170 |
+
"learning_rate": 4.000000000000001e-06,
|
| 171 |
+
"loss": 0.6905,
|
| 172 |
+
"step": 48
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.57,
|
| 177 |
+
"eval_loss": 0.6904626488685608,
|
| 178 |
+
"eval_macro_f1": 0.4724573671942093,
|
| 179 |
+
"eval_runtime": 1.349,
|
| 180 |
+
"eval_samples_per_second": 296.513,
|
| 181 |
+
"eval_steps_per_second": 2.965,
|
| 182 |
+
"step": 48
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 3.12,
|
| 186 |
+
"learning_rate": 4.166666666666667e-06,
|
| 187 |
+
"loss": 0.6902,
|
| 188 |
+
"step": 50
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 3.25,
|
| 192 |
+
"learning_rate": 4.333333333333334e-06,
|
| 193 |
+
"loss": 0.6873,
|
| 194 |
+
"step": 52
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 3.38,
|
| 198 |
+
"learning_rate": 4.5e-06,
|
| 199 |
+
"loss": 0.6838,
|
| 200 |
+
"step": 54
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 3.5,
|
| 204 |
+
"learning_rate": 4.666666666666667e-06,
|
| 205 |
+
"loss": 0.6996,
|
| 206 |
+
"step": 56
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 3.62,
|
| 210 |
+
"learning_rate": 4.833333333333333e-06,
|
| 211 |
+
"loss": 0.6857,
|
| 212 |
+
"step": 58
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 3.75,
|
| 216 |
+
"learning_rate": 5e-06,
|
| 217 |
+
"loss": 0.69,
|
| 218 |
+
"step": 60
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.88,
|
| 222 |
+
"learning_rate": 5.166666666666667e-06,
|
| 223 |
+
"loss": 0.6969,
|
| 224 |
+
"step": 62
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 4.0,
|
| 228 |
+
"learning_rate": 5.333333333333334e-06,
|
| 229 |
+
"loss": 0.6771,
|
| 230 |
+
"step": 64
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 4.0,
|
| 234 |
+
"eval_accuracy": 0.5725,
|
| 235 |
+
"eval_loss": 0.6874645352363586,
|
| 236 |
+
"eval_macro_f1": 0.38949829969921185,
|
| 237 |
+
"eval_runtime": 1.2979,
|
| 238 |
+
"eval_samples_per_second": 308.193,
|
| 239 |
+
"eval_steps_per_second": 3.082,
|
| 240 |
+
"step": 64
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 4.12,
|
| 244 |
+
"learning_rate": 5.500000000000001e-06,
|
| 245 |
+
"loss": 0.6962,
|
| 246 |
+
"step": 66
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.25,
|
| 250 |
+
"learning_rate": 5.666666666666667e-06,
|
| 251 |
+
"loss": 0.6745,
|
| 252 |
+
"step": 68
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 4.38,
|
| 256 |
+
"learning_rate": 5.833333333333334e-06,
|
| 257 |
+
"loss": 0.6831,
|
| 258 |
+
"step": 70
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.5,
|
| 262 |
+
"learning_rate": 6e-06,
|
| 263 |
+
"loss": 0.6989,
|
| 264 |
+
"step": 72
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.62,
|
| 268 |
+
"learning_rate": 6.166666666666667e-06,
|
| 269 |
+
"loss": 0.684,
|
| 270 |
+
"step": 74
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 4.75,
|
| 274 |
+
"learning_rate": 6.333333333333334e-06,
|
| 275 |
+
"loss": 0.6797,
|
| 276 |
+
"step": 76
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 4.88,
|
| 280 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 281 |
+
"loss": 0.6998,
|
| 282 |
+
"step": 78
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 5.0,
|
| 286 |
+
"learning_rate": 6.666666666666667e-06,
|
| 287 |
+
"loss": 0.6734,
|
| 288 |
+
"step": 80
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.0,
|
| 292 |
+
"eval_accuracy": 0.5725,
|
| 293 |
+
"eval_loss": 0.6859826445579529,
|
| 294 |
+
"eval_macro_f1": 0.37970998721331994,
|
| 295 |
+
"eval_runtime": 1.544,
|
| 296 |
+
"eval_samples_per_second": 259.064,
|
| 297 |
+
"eval_steps_per_second": 2.591,
|
| 298 |
+
"step": 80
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 5.12,
|
| 302 |
+
"learning_rate": 6.833333333333333e-06,
|
| 303 |
+
"loss": 0.6978,
|
| 304 |
+
"step": 82
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 5.25,
|
| 308 |
+
"learning_rate": 7.000000000000001e-06,
|
| 309 |
+
"loss": 0.6867,
|
| 310 |
+
"step": 84
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.38,
|
| 314 |
+
"learning_rate": 7.166666666666667e-06,
|
| 315 |
+
"loss": 0.68,
|
| 316 |
+
"step": 86
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 5.5,
|
| 320 |
+
"learning_rate": 7.333333333333334e-06,
|
| 321 |
+
"loss": 0.6922,
|
| 322 |
+
"step": 88
|
| 323 |
+
},
|
| 324 |
+
{
|
| 325 |
+
"epoch": 5.62,
|
| 326 |
+
"learning_rate": 7.5e-06,
|
| 327 |
+
"loss": 0.6833,
|
| 328 |
+
"step": 90
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 5.75,
|
| 332 |
+
"learning_rate": 7.666666666666667e-06,
|
| 333 |
+
"loss": 0.6832,
|
| 334 |
+
"step": 92
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 5.88,
|
| 338 |
+
"learning_rate": 7.833333333333333e-06,
|
| 339 |
+
"loss": 0.6961,
|
| 340 |
+
"step": 94
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 6.0,
|
| 344 |
+
"learning_rate": 8.000000000000001e-06,
|
| 345 |
+
"loss": 0.6562,
|
| 346 |
+
"step": 96
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 6.0,
|
| 350 |
+
"eval_accuracy": 0.5725,
|
| 351 |
+
"eval_loss": 0.6861733794212341,
|
| 352 |
+
"eval_macro_f1": 0.37970998721331994,
|
| 353 |
+
"eval_runtime": 1.424,
|
| 354 |
+
"eval_samples_per_second": 280.898,
|
| 355 |
+
"eval_steps_per_second": 2.809,
|
| 356 |
+
"step": 96
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.12,
|
| 360 |
+
"learning_rate": 8.166666666666668e-06,
|
| 361 |
+
"loss": 0.6936,
|
| 362 |
+
"step": 98
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 6.25,
|
| 366 |
+
"learning_rate": 8.333333333333334e-06,
|
| 367 |
+
"loss": 0.6943,
|
| 368 |
+
"step": 100
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 6.38,
|
| 372 |
+
"learning_rate": 8.500000000000002e-06,
|
| 373 |
+
"loss": 0.6689,
|
| 374 |
+
"step": 102
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 6.5,
|
| 378 |
+
"learning_rate": 8.666666666666668e-06,
|
| 379 |
+
"loss": 0.6919,
|
| 380 |
+
"step": 104
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.62,
|
| 384 |
+
"learning_rate": 8.833333333333334e-06,
|
| 385 |
+
"loss": 0.674,
|
| 386 |
+
"step": 106
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 6.75,
|
| 390 |
+
"learning_rate": 9e-06,
|
| 391 |
+
"loss": 0.6754,
|
| 392 |
+
"step": 108
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 6.88,
|
| 396 |
+
"learning_rate": 9.166666666666666e-06,
|
| 397 |
+
"loss": 0.6999,
|
| 398 |
+
"step": 110
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 7.0,
|
| 402 |
+
"learning_rate": 9.333333333333334e-06,
|
| 403 |
+
"loss": 0.6714,
|
| 404 |
+
"step": 112
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 7.0,
|
| 408 |
+
"eval_accuracy": 0.57,
|
| 409 |
+
"eval_loss": 0.6851080060005188,
|
| 410 |
+
"eval_macro_f1": 0.3735431235431235,
|
| 411 |
+
"eval_runtime": 1.4333,
|
| 412 |
+
"eval_samples_per_second": 279.083,
|
| 413 |
+
"eval_steps_per_second": 2.791,
|
| 414 |
+
"step": 112
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.12,
|
| 418 |
+
"learning_rate": 9.5e-06,
|
| 419 |
+
"loss": 0.694,
|
| 420 |
+
"step": 114
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.25,
|
| 424 |
+
"learning_rate": 9.666666666666667e-06,
|
| 425 |
+
"loss": 0.6837,
|
| 426 |
+
"step": 116
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.38,
|
| 430 |
+
"learning_rate": 9.833333333333333e-06,
|
| 431 |
+
"loss": 0.6692,
|
| 432 |
+
"step": 118
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.5,
|
| 436 |
+
"learning_rate": 1e-05,
|
| 437 |
+
"loss": 0.7002,
|
| 438 |
+
"step": 120
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 7.62,
|
| 442 |
+
"learning_rate": 1.0166666666666667e-05,
|
| 443 |
+
"loss": 0.6706,
|
| 444 |
+
"step": 122
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.75,
|
| 448 |
+
"learning_rate": 1.0333333333333333e-05,
|
| 449 |
+
"loss": 0.6592,
|
| 450 |
+
"step": 124
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 7.88,
|
| 454 |
+
"learning_rate": 1.05e-05,
|
| 455 |
+
"loss": 0.7135,
|
| 456 |
+
"step": 126
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 8.0,
|
| 460 |
+
"learning_rate": 1.0666666666666667e-05,
|
| 461 |
+
"loss": 0.6858,
|
| 462 |
+
"step": 128
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.0,
|
| 466 |
+
"eval_accuracy": 0.5725,
|
| 467 |
+
"eval_loss": 0.6847579479217529,
|
| 468 |
+
"eval_macro_f1": 0.3746285714285715,
|
| 469 |
+
"eval_runtime": 1.2984,
|
| 470 |
+
"eval_samples_per_second": 308.064,
|
| 471 |
+
"eval_steps_per_second": 3.081,
|
| 472 |
+
"step": 128
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.12,
|
| 476 |
+
"learning_rate": 1.0833333333333334e-05,
|
| 477 |
+
"loss": 0.6926,
|
| 478 |
+
"step": 130
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 8.25,
|
| 482 |
+
"learning_rate": 1.1000000000000001e-05,
|
| 483 |
+
"loss": 0.6814,
|
| 484 |
+
"step": 132
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 8.38,
|
| 488 |
+
"learning_rate": 1.1166666666666668e-05,
|
| 489 |
+
"loss": 0.669,
|
| 490 |
+
"step": 134
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.5,
|
| 494 |
+
"learning_rate": 1.1333333333333334e-05,
|
| 495 |
+
"loss": 0.6969,
|
| 496 |
+
"step": 136
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 8.62,
|
| 500 |
+
"learning_rate": 1.1500000000000002e-05,
|
| 501 |
+
"loss": 0.6732,
|
| 502 |
+
"step": 138
|
| 503 |
+
},
|
| 504 |
+
{
|
| 505 |
+
"epoch": 8.75,
|
| 506 |
+
"learning_rate": 1.1666666666666668e-05,
|
| 507 |
+
"loss": 0.6704,
|
| 508 |
+
"step": 140
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 8.88,
|
| 512 |
+
"learning_rate": 1.1833333333333334e-05,
|
| 513 |
+
"loss": 0.6989,
|
| 514 |
+
"step": 142
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 9.0,
|
| 518 |
+
"learning_rate": 1.2e-05,
|
| 519 |
+
"loss": 0.6831,
|
| 520 |
+
"step": 144
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 9.0,
|
| 524 |
+
"eval_accuracy": 0.575,
|
| 525 |
+
"eval_loss": 0.6859992742538452,
|
| 526 |
+
"eval_macro_f1": 0.38581596155930487,
|
| 527 |
+
"eval_runtime": 1.4972,
|
| 528 |
+
"eval_samples_per_second": 267.159,
|
| 529 |
+
"eval_steps_per_second": 2.672,
|
| 530 |
+
"step": 144
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 9.12,
|
| 534 |
+
"learning_rate": 1.2166666666666668e-05,
|
| 535 |
+
"loss": 0.6949,
|
| 536 |
+
"step": 146
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.25,
|
| 540 |
+
"learning_rate": 1.2333333333333334e-05,
|
| 541 |
+
"loss": 0.6692,
|
| 542 |
+
"step": 148
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 9.38,
|
| 546 |
+
"learning_rate": 1.25e-05,
|
| 547 |
+
"loss": 0.6792,
|
| 548 |
+
"step": 150
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 9.5,
|
| 552 |
+
"learning_rate": 1.2666666666666668e-05,
|
| 553 |
+
"loss": 0.6972,
|
| 554 |
+
"step": 152
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 9.62,
|
| 558 |
+
"learning_rate": 1.2833333333333333e-05,
|
| 559 |
+
"loss": 0.6766,
|
| 560 |
+
"step": 154
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 9.75,
|
| 564 |
+
"learning_rate": 1.3000000000000001e-05,
|
| 565 |
+
"loss": 0.678,
|
| 566 |
+
"step": 156
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 9.88,
|
| 570 |
+
"learning_rate": 1.3166666666666665e-05,
|
| 571 |
+
"loss": 0.6887,
|
| 572 |
+
"step": 158
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 10.0,
|
| 576 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 577 |
+
"loss": 0.6676,
|
| 578 |
+
"step": 160
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.0,
|
| 582 |
+
"eval_accuracy": 0.575,
|
| 583 |
+
"eval_loss": 0.6874198913574219,
|
| 584 |
+
"eval_macro_f1": 0.42159164370045255,
|
| 585 |
+
"eval_runtime": 1.4475,
|
| 586 |
+
"eval_samples_per_second": 276.347,
|
| 587 |
+
"eval_steps_per_second": 2.763,
|
| 588 |
+
"step": 160
|
| 589 |
+
}
|
| 590 |
+
],
|
| 591 |
+
"max_steps": 320,
|
| 592 |
+
"num_train_epochs": 20,
|
| 593 |
+
"total_flos": 634934132736000.0,
|
| 594 |
+
"trial_name": null,
|
| 595 |
+
"trial_params": null
|
| 596 |
+
}
|
scaling_performance/2000/fine-tuned/.DS_Store
CHANGED
|
Binary files a/scaling_performance/2000/fine-tuned/.DS_Store and b/scaling_performance/2000/fine-tuned/.DS_Store differ
|
|
|