Training in progress, epoch 0
Browse files- all_results.json +13 -0
- eval_results.json +8 -0
- pytorch_model.bin +1 -1
- runs/Apr17_11-49-22_5910a653ef0c/events.out.tfevents.1681734268.5910a653ef0c.229.2 +3 -0
- runs/Apr17_12-28-29_5910a653ef0c/1681734512.5820327/events.out.tfevents.1681734512.5910a653ef0c.229.4 +3 -0
- runs/Apr17_12-28-29_5910a653ef0c/events.out.tfevents.1681734512.5910a653ef0c.229.3 +3 -0
- train_results.json +8 -0
- trainer_state.json +592 -0
- training_args.bin +1 -1
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.99,
|
| 3 |
+
"eval_accuracy": 0.9629629629629629,
|
| 4 |
+
"eval_loss": 0.10010068118572235,
|
| 5 |
+
"eval_runtime": 7.7425,
|
| 6 |
+
"eval_samples_per_second": 80.207,
|
| 7 |
+
"eval_steps_per_second": 10.074,
|
| 8 |
+
"total_flos": 6.917311166047027e+17,
|
| 9 |
+
"train_loss": 0.36484567570960386,
|
| 10 |
+
"train_runtime": 2062.5819,
|
| 11 |
+
"train_samples_per_second": 13.529,
|
| 12 |
+
"train_steps_per_second": 0.422
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.99,
|
| 3 |
+
"eval_accuracy": 0.9629629629629629,
|
| 4 |
+
"eval_loss": 0.10010068118572235,
|
| 5 |
+
"eval_runtime": 7.7425,
|
| 6 |
+
"eval_samples_per_second": 80.207,
|
| 7 |
+
"eval_steps_per_second": 10.074
|
| 8 |
+
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 110401009
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ce5cf7207cb6d9e7122febc95f95b6efdc36a9865e0e37de3e7e12e097393b8
|
| 3 |
size 110401009
|
runs/Apr17_11-49-22_5910a653ef0c/events.out.tfevents.1681734268.5910a653ef0c.229.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8036eb840d359d4f7c63c5a1ae37f9d49878a1d5d4d93de625c6f14c3f903e52
|
| 3 |
+
size 411
|
runs/Apr17_12-28-29_5910a653ef0c/1681734512.5820327/events.out.tfevents.1681734512.5910a653ef0c.229.4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f534247907c703633817dc347589f46ecd08d40efc664a49057b6e4c6082bb17
|
| 3 |
+
size 5970
|
runs/Apr17_12-28-29_5910a653ef0c/events.out.tfevents.1681734512.5910a653ef0c.229.3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:49b882046160ccf5b53d9d005fe63081f9450da35518b5f94b957912fbe6515b
|
| 3 |
+
size 7496
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.99,
|
| 3 |
+
"total_flos": 6.917311166047027e+17,
|
| 4 |
+
"train_loss": 0.36484567570960386,
|
| 5 |
+
"train_runtime": 2062.5819,
|
| 6 |
+
"train_samples_per_second": 13.529,
|
| 7 |
+
"train_steps_per_second": 0.422
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,592 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9629629629629629,
|
| 3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-bottom_cleaned_data/checkpoint-698",
|
| 4 |
+
"epoch": 4.98567335243553,
|
| 5 |
+
"global_step": 870,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.06,
|
| 12 |
+
"learning_rate": 5.747126436781609e-06,
|
| 13 |
+
"loss": 1.4399,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.11,
|
| 18 |
+
"learning_rate": 1.1494252873563218e-05,
|
| 19 |
+
"loss": 1.3855,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.17,
|
| 24 |
+
"learning_rate": 1.7241379310344828e-05,
|
| 25 |
+
"loss": 1.2914,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.23,
|
| 30 |
+
"learning_rate": 2.2988505747126437e-05,
|
| 31 |
+
"loss": 1.1756,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.29,
|
| 36 |
+
"learning_rate": 2.8735632183908045e-05,
|
| 37 |
+
"loss": 0.993,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.34,
|
| 42 |
+
"learning_rate": 3.4482758620689657e-05,
|
| 43 |
+
"loss": 0.8509,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.4,
|
| 48 |
+
"learning_rate": 4.0229885057471265e-05,
|
| 49 |
+
"loss": 0.7269,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.46,
|
| 54 |
+
"learning_rate": 4.597701149425287e-05,
|
| 55 |
+
"loss": 0.6418,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.52,
|
| 60 |
+
"learning_rate": 4.980842911877395e-05,
|
| 61 |
+
"loss": 0.6033,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.57,
|
| 66 |
+
"learning_rate": 4.916985951468711e-05,
|
| 67 |
+
"loss": 0.6462,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.63,
|
| 72 |
+
"learning_rate": 4.853128991060026e-05,
|
| 73 |
+
"loss": 0.6602,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.69,
|
| 78 |
+
"learning_rate": 4.789272030651341e-05,
|
| 79 |
+
"loss": 0.4678,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.74,
|
| 84 |
+
"learning_rate": 4.725415070242657e-05,
|
| 85 |
+
"loss": 0.4999,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.8,
|
| 90 |
+
"learning_rate": 4.661558109833972e-05,
|
| 91 |
+
"loss": 0.5158,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.86,
|
| 96 |
+
"learning_rate": 4.597701149425287e-05,
|
| 97 |
+
"loss": 0.4182,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.92,
|
| 102 |
+
"learning_rate": 4.5338441890166025e-05,
|
| 103 |
+
"loss": 0.4787,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.97,
|
| 108 |
+
"learning_rate": 4.469987228607918e-05,
|
| 109 |
+
"loss": 0.4343,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 1.0,
|
| 114 |
+
"eval_accuracy": 0.927536231884058,
|
| 115 |
+
"eval_loss": 0.2457878440618515,
|
| 116 |
+
"eval_runtime": 141.3791,
|
| 117 |
+
"eval_samples_per_second": 4.392,
|
| 118 |
+
"eval_steps_per_second": 0.552,
|
| 119 |
+
"step": 174
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.03,
|
| 123 |
+
"learning_rate": 4.406130268199234e-05,
|
| 124 |
+
"loss": 0.502,
|
| 125 |
+
"step": 180
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.09,
|
| 129 |
+
"learning_rate": 4.342273307790549e-05,
|
| 130 |
+
"loss": 0.5438,
|
| 131 |
+
"step": 190
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 1.15,
|
| 135 |
+
"learning_rate": 4.278416347381865e-05,
|
| 136 |
+
"loss": 0.4235,
|
| 137 |
+
"step": 200
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.2,
|
| 141 |
+
"learning_rate": 4.21455938697318e-05,
|
| 142 |
+
"loss": 0.3872,
|
| 143 |
+
"step": 210
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.26,
|
| 147 |
+
"learning_rate": 4.1507024265644955e-05,
|
| 148 |
+
"loss": 0.3741,
|
| 149 |
+
"step": 220
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.32,
|
| 153 |
+
"learning_rate": 4.086845466155811e-05,
|
| 154 |
+
"loss": 0.3162,
|
| 155 |
+
"step": 230
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 1.38,
|
| 159 |
+
"learning_rate": 4.0229885057471265e-05,
|
| 160 |
+
"loss": 0.306,
|
| 161 |
+
"step": 240
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.43,
|
| 165 |
+
"learning_rate": 3.959131545338442e-05,
|
| 166 |
+
"loss": 0.3393,
|
| 167 |
+
"step": 250
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 1.49,
|
| 171 |
+
"learning_rate": 3.8952745849297575e-05,
|
| 172 |
+
"loss": 0.4019,
|
| 173 |
+
"step": 260
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.55,
|
| 177 |
+
"learning_rate": 3.831417624521073e-05,
|
| 178 |
+
"loss": 0.3279,
|
| 179 |
+
"step": 270
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.6,
|
| 183 |
+
"learning_rate": 3.7675606641123885e-05,
|
| 184 |
+
"loss": 0.2847,
|
| 185 |
+
"step": 280
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.66,
|
| 189 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 190 |
+
"loss": 0.3104,
|
| 191 |
+
"step": 290
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.72,
|
| 195 |
+
"learning_rate": 3.6398467432950195e-05,
|
| 196 |
+
"loss": 0.3329,
|
| 197 |
+
"step": 300
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.78,
|
| 201 |
+
"learning_rate": 3.5759897828863347e-05,
|
| 202 |
+
"loss": 0.3435,
|
| 203 |
+
"step": 310
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.83,
|
| 207 |
+
"learning_rate": 3.51213282247765e-05,
|
| 208 |
+
"loss": 0.3614,
|
| 209 |
+
"step": 320
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 1.89,
|
| 213 |
+
"learning_rate": 3.4482758620689657e-05,
|
| 214 |
+
"loss": 0.3221,
|
| 215 |
+
"step": 330
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 1.95,
|
| 219 |
+
"learning_rate": 3.3844189016602815e-05,
|
| 220 |
+
"loss": 0.3086,
|
| 221 |
+
"step": 340
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 2.0,
|
| 225 |
+
"eval_accuracy": 0.9516908212560387,
|
| 226 |
+
"eval_loss": 0.13658131659030914,
|
| 227 |
+
"eval_runtime": 8.9796,
|
| 228 |
+
"eval_samples_per_second": 69.157,
|
| 229 |
+
"eval_steps_per_second": 8.686,
|
| 230 |
+
"step": 349
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.01,
|
| 234 |
+
"learning_rate": 3.3205619412515967e-05,
|
| 235 |
+
"loss": 0.3944,
|
| 236 |
+
"step": 350
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 2.06,
|
| 240 |
+
"learning_rate": 3.256704980842912e-05,
|
| 241 |
+
"loss": 0.3012,
|
| 242 |
+
"step": 360
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 2.12,
|
| 246 |
+
"learning_rate": 3.192848020434228e-05,
|
| 247 |
+
"loss": 0.3217,
|
| 248 |
+
"step": 370
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 2.18,
|
| 252 |
+
"learning_rate": 3.128991060025543e-05,
|
| 253 |
+
"loss": 0.3616,
|
| 254 |
+
"step": 380
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 2.23,
|
| 258 |
+
"learning_rate": 3.065134099616858e-05,
|
| 259 |
+
"loss": 0.2652,
|
| 260 |
+
"step": 390
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 2.29,
|
| 264 |
+
"learning_rate": 3.0012771392081738e-05,
|
| 265 |
+
"loss": 0.3129,
|
| 266 |
+
"step": 400
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.35,
|
| 270 |
+
"learning_rate": 2.9374201787994893e-05,
|
| 271 |
+
"loss": 0.3538,
|
| 272 |
+
"step": 410
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.41,
|
| 276 |
+
"learning_rate": 2.8735632183908045e-05,
|
| 277 |
+
"loss": 0.2775,
|
| 278 |
+
"step": 420
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 2.46,
|
| 282 |
+
"learning_rate": 2.8097062579821203e-05,
|
| 283 |
+
"loss": 0.3021,
|
| 284 |
+
"step": 430
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 2.52,
|
| 288 |
+
"learning_rate": 2.745849297573436e-05,
|
| 289 |
+
"loss": 0.2512,
|
| 290 |
+
"step": 440
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 2.58,
|
| 294 |
+
"learning_rate": 2.681992337164751e-05,
|
| 295 |
+
"loss": 0.2721,
|
| 296 |
+
"step": 450
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 2.64,
|
| 300 |
+
"learning_rate": 2.618135376756067e-05,
|
| 301 |
+
"loss": 0.2555,
|
| 302 |
+
"step": 460
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 2.69,
|
| 306 |
+
"learning_rate": 2.554278416347382e-05,
|
| 307 |
+
"loss": 0.2576,
|
| 308 |
+
"step": 470
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 2.75,
|
| 312 |
+
"learning_rate": 2.4904214559386975e-05,
|
| 313 |
+
"loss": 0.3178,
|
| 314 |
+
"step": 480
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 2.81,
|
| 318 |
+
"learning_rate": 2.426564495530013e-05,
|
| 319 |
+
"loss": 0.2992,
|
| 320 |
+
"step": 490
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 2.87,
|
| 324 |
+
"learning_rate": 2.3627075351213285e-05,
|
| 325 |
+
"loss": 0.2333,
|
| 326 |
+
"step": 500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 2.92,
|
| 330 |
+
"learning_rate": 2.2988505747126437e-05,
|
| 331 |
+
"loss": 0.2409,
|
| 332 |
+
"step": 510
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 2.98,
|
| 336 |
+
"learning_rate": 2.234993614303959e-05,
|
| 337 |
+
"loss": 0.2377,
|
| 338 |
+
"step": 520
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 3.0,
|
| 342 |
+
"eval_accuracy": 0.9468599033816425,
|
| 343 |
+
"eval_loss": 0.1384800225496292,
|
| 344 |
+
"eval_runtime": 7.5854,
|
| 345 |
+
"eval_samples_per_second": 81.868,
|
| 346 |
+
"eval_steps_per_second": 10.283,
|
| 347 |
+
"step": 523
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 3.04,
|
| 351 |
+
"learning_rate": 2.1711366538952747e-05,
|
| 352 |
+
"loss": 0.2089,
|
| 353 |
+
"step": 530
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 3.09,
|
| 357 |
+
"learning_rate": 2.10727969348659e-05,
|
| 358 |
+
"loss": 0.2406,
|
| 359 |
+
"step": 540
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 3.15,
|
| 363 |
+
"learning_rate": 2.0434227330779057e-05,
|
| 364 |
+
"loss": 0.242,
|
| 365 |
+
"step": 550
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 3.21,
|
| 369 |
+
"learning_rate": 1.979565772669221e-05,
|
| 370 |
+
"loss": 0.2607,
|
| 371 |
+
"step": 560
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 3.27,
|
| 375 |
+
"learning_rate": 1.9157088122605367e-05,
|
| 376 |
+
"loss": 0.2147,
|
| 377 |
+
"step": 570
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 3.32,
|
| 381 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 382 |
+
"loss": 0.2211,
|
| 383 |
+
"step": 580
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 3.38,
|
| 387 |
+
"learning_rate": 1.7879948914431673e-05,
|
| 388 |
+
"loss": 0.2235,
|
| 389 |
+
"step": 590
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 3.44,
|
| 393 |
+
"learning_rate": 1.7241379310344828e-05,
|
| 394 |
+
"loss": 0.2362,
|
| 395 |
+
"step": 600
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 3.5,
|
| 399 |
+
"learning_rate": 1.6602809706257983e-05,
|
| 400 |
+
"loss": 0.2407,
|
| 401 |
+
"step": 610
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 3.55,
|
| 405 |
+
"learning_rate": 1.596424010217114e-05,
|
| 406 |
+
"loss": 0.2809,
|
| 407 |
+
"step": 620
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 3.61,
|
| 411 |
+
"learning_rate": 1.532567049808429e-05,
|
| 412 |
+
"loss": 0.1666,
|
| 413 |
+
"step": 630
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 3.67,
|
| 417 |
+
"learning_rate": 1.4687100893997447e-05,
|
| 418 |
+
"loss": 0.2207,
|
| 419 |
+
"step": 640
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 3.72,
|
| 423 |
+
"learning_rate": 1.4048531289910602e-05,
|
| 424 |
+
"loss": 0.2079,
|
| 425 |
+
"step": 650
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 3.78,
|
| 429 |
+
"learning_rate": 1.3409961685823755e-05,
|
| 430 |
+
"loss": 0.2262,
|
| 431 |
+
"step": 660
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 3.84,
|
| 435 |
+
"learning_rate": 1.277139208173691e-05,
|
| 436 |
+
"loss": 0.1975,
|
| 437 |
+
"step": 670
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 3.9,
|
| 441 |
+
"learning_rate": 1.2132822477650065e-05,
|
| 442 |
+
"loss": 0.1849,
|
| 443 |
+
"step": 680
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 3.95,
|
| 447 |
+
"learning_rate": 1.1494252873563218e-05,
|
| 448 |
+
"loss": 0.2451,
|
| 449 |
+
"step": 690
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 4.0,
|
| 453 |
+
"eval_accuracy": 0.9629629629629629,
|
| 454 |
+
"eval_loss": 0.10010068118572235,
|
| 455 |
+
"eval_runtime": 9.1061,
|
| 456 |
+
"eval_samples_per_second": 68.196,
|
| 457 |
+
"eval_steps_per_second": 8.566,
|
| 458 |
+
"step": 698
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 4.01,
|
| 462 |
+
"learning_rate": 1.0855683269476373e-05,
|
| 463 |
+
"loss": 0.22,
|
| 464 |
+
"step": 700
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 4.07,
|
| 468 |
+
"learning_rate": 1.0217113665389528e-05,
|
| 469 |
+
"loss": 0.2175,
|
| 470 |
+
"step": 710
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 4.13,
|
| 474 |
+
"learning_rate": 9.578544061302683e-06,
|
| 475 |
+
"loss": 0.1948,
|
| 476 |
+
"step": 720
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 4.18,
|
| 480 |
+
"learning_rate": 8.939974457215837e-06,
|
| 481 |
+
"loss": 0.168,
|
| 482 |
+
"step": 730
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 4.24,
|
| 486 |
+
"learning_rate": 8.301404853128992e-06,
|
| 487 |
+
"loss": 0.1665,
|
| 488 |
+
"step": 740
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 4.3,
|
| 492 |
+
"learning_rate": 7.662835249042145e-06,
|
| 493 |
+
"loss": 0.183,
|
| 494 |
+
"step": 750
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 4.36,
|
| 498 |
+
"learning_rate": 7.024265644955301e-06,
|
| 499 |
+
"loss": 0.1918,
|
| 500 |
+
"step": 760
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 4.41,
|
| 504 |
+
"learning_rate": 6.385696040868455e-06,
|
| 505 |
+
"loss": 0.1743,
|
| 506 |
+
"step": 770
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 4.47,
|
| 510 |
+
"learning_rate": 5.747126436781609e-06,
|
| 511 |
+
"loss": 0.1328,
|
| 512 |
+
"step": 780
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 4.53,
|
| 516 |
+
"learning_rate": 5.108556832694764e-06,
|
| 517 |
+
"loss": 0.1959,
|
| 518 |
+
"step": 790
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 4.58,
|
| 522 |
+
"learning_rate": 4.469987228607918e-06,
|
| 523 |
+
"loss": 0.172,
|
| 524 |
+
"step": 800
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 4.64,
|
| 528 |
+
"learning_rate": 3.8314176245210725e-06,
|
| 529 |
+
"loss": 0.1872,
|
| 530 |
+
"step": 810
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 4.7,
|
| 534 |
+
"learning_rate": 3.1928480204342275e-06,
|
| 535 |
+
"loss": 0.2218,
|
| 536 |
+
"step": 820
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 4.76,
|
| 540 |
+
"learning_rate": 2.554278416347382e-06,
|
| 541 |
+
"loss": 0.1797,
|
| 542 |
+
"step": 830
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 4.81,
|
| 546 |
+
"learning_rate": 1.9157088122605362e-06,
|
| 547 |
+
"loss": 0.1487,
|
| 548 |
+
"step": 840
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 4.87,
|
| 552 |
+
"learning_rate": 1.277139208173691e-06,
|
| 553 |
+
"loss": 0.2479,
|
| 554 |
+
"step": 850
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 4.93,
|
| 558 |
+
"learning_rate": 6.385696040868455e-07,
|
| 559 |
+
"loss": 0.1359,
|
| 560 |
+
"step": 860
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 4.99,
|
| 564 |
+
"learning_rate": 0.0,
|
| 565 |
+
"loss": 0.1148,
|
| 566 |
+
"step": 870
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 4.99,
|
| 570 |
+
"eval_accuracy": 0.961352657004831,
|
| 571 |
+
"eval_loss": 0.09633708000183105,
|
| 572 |
+
"eval_runtime": 7.7205,
|
| 573 |
+
"eval_samples_per_second": 80.435,
|
| 574 |
+
"eval_steps_per_second": 10.103,
|
| 575 |
+
"step": 870
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 4.99,
|
| 579 |
+
"step": 870,
|
| 580 |
+
"total_flos": 6.917311166047027e+17,
|
| 581 |
+
"train_loss": 0.36484567570960386,
|
| 582 |
+
"train_runtime": 2062.5819,
|
| 583 |
+
"train_samples_per_second": 13.529,
|
| 584 |
+
"train_steps_per_second": 0.422
|
| 585 |
+
}
|
| 586 |
+
],
|
| 587 |
+
"max_steps": 870,
|
| 588 |
+
"num_train_epochs": 5,
|
| 589 |
+
"total_flos": 6.917311166047027e+17,
|
| 590 |
+
"trial_name": null,
|
| 591 |
+
"trial_params": null
|
| 592 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3643
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3269c8bf9b3171151c206c3d4adc88506d8dd8c6ae4508d917fe180083ccbf30
|
| 3 |
size 3643
|