Commit
•
688f085
1
Parent(s):
3e431b7
End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"eval_accuracy": 0.8994708994708994,
|
| 4 |
+
"eval_loss": 0.3382543921470642,
|
| 5 |
+
"eval_runtime": 71.1959,
|
| 6 |
+
"eval_samples_per_second": 26.546,
|
| 7 |
+
"eval_steps_per_second": 1.671,
|
| 8 |
+
"total_flos": 1.6742089052991848e+19,
|
| 9 |
+
"train_loss": 0.5761546168379679,
|
| 10 |
+
"train_runtime": 6609.6541,
|
| 11 |
+
"train_samples_per_second": 9.722,
|
| 12 |
+
"train_steps_per_second": 0.152
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"eval_accuracy": 0.8994708994708994,
|
| 4 |
+
"eval_loss": 0.3382543921470642,
|
| 5 |
+
"eval_runtime": 71.1959,
|
| 6 |
+
"eval_samples_per_second": 26.546,
|
| 7 |
+
"eval_steps_per_second": 1.671
|
| 8 |
+
}
|
runs/Apr10_03-38-39_275a0a6eb96c/events.out.tfevents.1681105637.275a0a6eb96c.23.5
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a02947e5d4b952bd5414b5ecb9cd8ec5ad50c31359afad25234faa12491b5d38
|
| 3 |
+
size 363
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"total_flos": 1.6742089052991848e+19,
|
| 4 |
+
"train_loss": 0.5761546168379679,
|
| 5 |
+
"train_runtime": 6609.6541,
|
| 6 |
+
"train_samples_per_second": 9.722,
|
| 7 |
+
"train_steps_per_second": 0.152
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,679 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8994708994708994,
|
| 3 |
+
"best_model_checkpoint": "Har_Finetuned-ViT-Hybrid/checkpoint-1002",
|
| 4 |
+
"epoch": 5.997014925373135,
|
| 5 |
+
"global_step": 1002,
|
| 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": 4.950495049504951e-06,
|
| 13 |
+
"loss": 2.8031,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.12,
|
| 18 |
+
"learning_rate": 9.900990099009901e-06,
|
| 19 |
+
"loss": 2.5937,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.18,
|
| 24 |
+
"learning_rate": 1.4851485148514851e-05,
|
| 25 |
+
"loss": 2.2631,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.24,
|
| 30 |
+
"learning_rate": 1.9801980198019803e-05,
|
| 31 |
+
"loss": 1.8858,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.3,
|
| 36 |
+
"learning_rate": 2.4752475247524754e-05,
|
| 37 |
+
"loss": 1.5579,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.36,
|
| 42 |
+
"learning_rate": 2.9702970297029702e-05,
|
| 43 |
+
"loss": 1.1878,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.42,
|
| 48 |
+
"learning_rate": 3.465346534653465e-05,
|
| 49 |
+
"loss": 0.9966,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.48,
|
| 54 |
+
"learning_rate": 3.9603960396039605e-05,
|
| 55 |
+
"loss": 0.9343,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.54,
|
| 60 |
+
"learning_rate": 4.455445544554456e-05,
|
| 61 |
+
"loss": 0.897,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.6,
|
| 66 |
+
"learning_rate": 4.950495049504951e-05,
|
| 67 |
+
"loss": 0.8911,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.66,
|
| 72 |
+
"learning_rate": 4.9500554938956716e-05,
|
| 73 |
+
"loss": 0.8231,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.72,
|
| 78 |
+
"learning_rate": 4.894561598224195e-05,
|
| 79 |
+
"loss": 0.8594,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.78,
|
| 84 |
+
"learning_rate": 4.8390677025527196e-05,
|
| 85 |
+
"loss": 0.7903,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.84,
|
| 90 |
+
"learning_rate": 4.783573806881244e-05,
|
| 91 |
+
"loss": 0.7901,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.9,
|
| 96 |
+
"learning_rate": 4.7280799112097675e-05,
|
| 97 |
+
"loss": 0.7695,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.96,
|
| 102 |
+
"learning_rate": 4.672586015538291e-05,
|
| 103 |
+
"loss": 0.7923,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.0,
|
| 108 |
+
"eval_accuracy": 0.8698412698412699,
|
| 109 |
+
"eval_loss": 0.44196292757987976,
|
| 110 |
+
"eval_runtime": 70.4064,
|
| 111 |
+
"eval_samples_per_second": 26.844,
|
| 112 |
+
"eval_steps_per_second": 1.69,
|
| 113 |
+
"step": 167
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 1.02,
|
| 117 |
+
"learning_rate": 4.617092119866815e-05,
|
| 118 |
+
"loss": 0.6929,
|
| 119 |
+
"step": 170
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.08,
|
| 123 |
+
"learning_rate": 4.5615982241953385e-05,
|
| 124 |
+
"loss": 0.6409,
|
| 125 |
+
"step": 180
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.14,
|
| 129 |
+
"learning_rate": 4.506104328523862e-05,
|
| 130 |
+
"loss": 0.6611,
|
| 131 |
+
"step": 190
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 1.2,
|
| 135 |
+
"learning_rate": 4.4506104328523865e-05,
|
| 136 |
+
"loss": 0.6694,
|
| 137 |
+
"step": 200
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.26,
|
| 141 |
+
"learning_rate": 4.395116537180911e-05,
|
| 142 |
+
"loss": 0.622,
|
| 143 |
+
"step": 210
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.32,
|
| 147 |
+
"learning_rate": 4.3396226415094345e-05,
|
| 148 |
+
"loss": 0.6312,
|
| 149 |
+
"step": 220
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.38,
|
| 153 |
+
"learning_rate": 4.284128745837958e-05,
|
| 154 |
+
"loss": 0.6552,
|
| 155 |
+
"step": 230
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 1.44,
|
| 159 |
+
"learning_rate": 4.228634850166482e-05,
|
| 160 |
+
"loss": 0.6335,
|
| 161 |
+
"step": 240
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.5,
|
| 165 |
+
"learning_rate": 4.1731409544950054e-05,
|
| 166 |
+
"loss": 0.6952,
|
| 167 |
+
"step": 250
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 1.56,
|
| 171 |
+
"learning_rate": 4.11764705882353e-05,
|
| 172 |
+
"loss": 0.5943,
|
| 173 |
+
"step": 260
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.61,
|
| 177 |
+
"learning_rate": 4.062153163152054e-05,
|
| 178 |
+
"loss": 0.5843,
|
| 179 |
+
"step": 270
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.67,
|
| 183 |
+
"learning_rate": 4.006659267480578e-05,
|
| 184 |
+
"loss": 0.6073,
|
| 185 |
+
"step": 280
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.73,
|
| 189 |
+
"learning_rate": 3.9511653718091014e-05,
|
| 190 |
+
"loss": 0.6046,
|
| 191 |
+
"step": 290
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.79,
|
| 195 |
+
"learning_rate": 3.895671476137625e-05,
|
| 196 |
+
"loss": 0.5643,
|
| 197 |
+
"step": 300
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.85,
|
| 201 |
+
"learning_rate": 3.8401775804661487e-05,
|
| 202 |
+
"loss": 0.6454,
|
| 203 |
+
"step": 310
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.91,
|
| 207 |
+
"learning_rate": 3.784683684794672e-05,
|
| 208 |
+
"loss": 0.5922,
|
| 209 |
+
"step": 320
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 1.97,
|
| 213 |
+
"learning_rate": 3.7291897891231966e-05,
|
| 214 |
+
"loss": 0.5555,
|
| 215 |
+
"step": 330
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 2.0,
|
| 219 |
+
"eval_accuracy": 0.882010582010582,
|
| 220 |
+
"eval_loss": 0.3810536563396454,
|
| 221 |
+
"eval_runtime": 70.4153,
|
| 222 |
+
"eval_samples_per_second": 26.841,
|
| 223 |
+
"eval_steps_per_second": 1.69,
|
| 224 |
+
"step": 334
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.04,
|
| 228 |
+
"learning_rate": 3.673695893451721e-05,
|
| 229 |
+
"loss": 0.6331,
|
| 230 |
+
"step": 340
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.1,
|
| 234 |
+
"learning_rate": 3.6182019977802446e-05,
|
| 235 |
+
"loss": 0.4524,
|
| 236 |
+
"step": 350
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 2.16,
|
| 240 |
+
"learning_rate": 3.562708102108768e-05,
|
| 241 |
+
"loss": 0.4958,
|
| 242 |
+
"step": 360
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 2.21,
|
| 246 |
+
"learning_rate": 3.507214206437292e-05,
|
| 247 |
+
"loss": 0.5111,
|
| 248 |
+
"step": 370
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 2.27,
|
| 252 |
+
"learning_rate": 3.4517203107658156e-05,
|
| 253 |
+
"loss": 0.5885,
|
| 254 |
+
"step": 380
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 2.33,
|
| 258 |
+
"learning_rate": 3.39622641509434e-05,
|
| 259 |
+
"loss": 0.5497,
|
| 260 |
+
"step": 390
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 2.39,
|
| 264 |
+
"learning_rate": 3.340732519422864e-05,
|
| 265 |
+
"loss": 0.4644,
|
| 266 |
+
"step": 400
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.45,
|
| 270 |
+
"learning_rate": 3.285238623751388e-05,
|
| 271 |
+
"loss": 0.5118,
|
| 272 |
+
"step": 410
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.51,
|
| 276 |
+
"learning_rate": 3.2297447280799115e-05,
|
| 277 |
+
"loss": 0.5281,
|
| 278 |
+
"step": 420
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 2.57,
|
| 282 |
+
"learning_rate": 3.174250832408435e-05,
|
| 283 |
+
"loss": 0.5025,
|
| 284 |
+
"step": 430
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 2.63,
|
| 288 |
+
"learning_rate": 3.118756936736959e-05,
|
| 289 |
+
"loss": 0.4469,
|
| 290 |
+
"step": 440
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 2.69,
|
| 294 |
+
"learning_rate": 3.0632630410654825e-05,
|
| 295 |
+
"loss": 0.4594,
|
| 296 |
+
"step": 450
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 2.75,
|
| 300 |
+
"learning_rate": 3.0077691453940065e-05,
|
| 301 |
+
"loss": 0.5019,
|
| 302 |
+
"step": 460
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 2.81,
|
| 306 |
+
"learning_rate": 2.9522752497225308e-05,
|
| 307 |
+
"loss": 0.5002,
|
| 308 |
+
"step": 470
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 2.87,
|
| 312 |
+
"learning_rate": 2.8967813540510548e-05,
|
| 313 |
+
"loss": 0.4876,
|
| 314 |
+
"step": 480
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 2.93,
|
| 318 |
+
"learning_rate": 2.8412874583795784e-05,
|
| 319 |
+
"loss": 0.5479,
|
| 320 |
+
"step": 490
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 2.99,
|
| 324 |
+
"learning_rate": 2.785793562708102e-05,
|
| 325 |
+
"loss": 0.4734,
|
| 326 |
+
"step": 500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 3.0,
|
| 330 |
+
"eval_accuracy": 0.8957671957671958,
|
| 331 |
+
"eval_loss": 0.34475043416023254,
|
| 332 |
+
"eval_runtime": 70.2044,
|
| 333 |
+
"eval_samples_per_second": 26.921,
|
| 334 |
+
"eval_steps_per_second": 1.695,
|
| 335 |
+
"step": 501
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 3.05,
|
| 339 |
+
"learning_rate": 2.730299667036626e-05,
|
| 340 |
+
"loss": 0.4618,
|
| 341 |
+
"step": 510
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 3.11,
|
| 345 |
+
"learning_rate": 2.6748057713651497e-05,
|
| 346 |
+
"loss": 0.4161,
|
| 347 |
+
"step": 520
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 3.17,
|
| 351 |
+
"learning_rate": 2.619311875693674e-05,
|
| 352 |
+
"loss": 0.3976,
|
| 353 |
+
"step": 530
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 3.23,
|
| 357 |
+
"learning_rate": 2.563817980022198e-05,
|
| 358 |
+
"loss": 0.3763,
|
| 359 |
+
"step": 540
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 3.29,
|
| 363 |
+
"learning_rate": 2.5083240843507217e-05,
|
| 364 |
+
"loss": 0.4015,
|
| 365 |
+
"step": 550
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 3.35,
|
| 369 |
+
"learning_rate": 2.4528301886792453e-05,
|
| 370 |
+
"loss": 0.3991,
|
| 371 |
+
"step": 560
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 3.41,
|
| 375 |
+
"learning_rate": 2.3973362930077693e-05,
|
| 376 |
+
"loss": 0.4204,
|
| 377 |
+
"step": 570
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 3.47,
|
| 381 |
+
"learning_rate": 2.3418423973362933e-05,
|
| 382 |
+
"loss": 0.3972,
|
| 383 |
+
"step": 580
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 3.53,
|
| 387 |
+
"learning_rate": 2.286348501664817e-05,
|
| 388 |
+
"loss": 0.3807,
|
| 389 |
+
"step": 590
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 3.59,
|
| 393 |
+
"learning_rate": 2.230854605993341e-05,
|
| 394 |
+
"loss": 0.4301,
|
| 395 |
+
"step": 600
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 3.65,
|
| 399 |
+
"learning_rate": 2.175360710321865e-05,
|
| 400 |
+
"loss": 0.4949,
|
| 401 |
+
"step": 610
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 3.71,
|
| 405 |
+
"learning_rate": 2.1198668146503886e-05,
|
| 406 |
+
"loss": 0.3774,
|
| 407 |
+
"step": 620
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 3.77,
|
| 411 |
+
"learning_rate": 2.0643729189789122e-05,
|
| 412 |
+
"loss": 0.3323,
|
| 413 |
+
"step": 630
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 3.83,
|
| 417 |
+
"learning_rate": 2.0088790233074362e-05,
|
| 418 |
+
"loss": 0.4668,
|
| 419 |
+
"step": 640
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 3.89,
|
| 423 |
+
"learning_rate": 1.9533851276359602e-05,
|
| 424 |
+
"loss": 0.4075,
|
| 425 |
+
"step": 650
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 3.95,
|
| 429 |
+
"learning_rate": 1.897891231964484e-05,
|
| 430 |
+
"loss": 0.4019,
|
| 431 |
+
"step": 660
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 4.0,
|
| 435 |
+
"eval_accuracy": 0.8925925925925926,
|
| 436 |
+
"eval_loss": 0.35210713744163513,
|
| 437 |
+
"eval_runtime": 70.2455,
|
| 438 |
+
"eval_samples_per_second": 26.906,
|
| 439 |
+
"eval_steps_per_second": 1.694,
|
| 440 |
+
"step": 668
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 4.01,
|
| 444 |
+
"learning_rate": 1.842397336293008e-05,
|
| 445 |
+
"loss": 0.4296,
|
| 446 |
+
"step": 670
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 4.07,
|
| 450 |
+
"learning_rate": 1.786903440621532e-05,
|
| 451 |
+
"loss": 0.4051,
|
| 452 |
+
"step": 680
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 4.13,
|
| 456 |
+
"learning_rate": 1.7314095449500555e-05,
|
| 457 |
+
"loss": 0.3749,
|
| 458 |
+
"step": 690
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 4.19,
|
| 462 |
+
"learning_rate": 1.6759156492785795e-05,
|
| 463 |
+
"loss": 0.3502,
|
| 464 |
+
"step": 700
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 4.25,
|
| 468 |
+
"learning_rate": 1.6204217536071035e-05,
|
| 469 |
+
"loss": 0.308,
|
| 470 |
+
"step": 710
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 4.31,
|
| 474 |
+
"learning_rate": 1.564927857935627e-05,
|
| 475 |
+
"loss": 0.3514,
|
| 476 |
+
"step": 720
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 4.37,
|
| 480 |
+
"learning_rate": 1.509433962264151e-05,
|
| 481 |
+
"loss": 0.3379,
|
| 482 |
+
"step": 730
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 4.43,
|
| 486 |
+
"learning_rate": 1.4539400665926751e-05,
|
| 487 |
+
"loss": 0.3101,
|
| 488 |
+
"step": 740
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 4.49,
|
| 492 |
+
"learning_rate": 1.3984461709211987e-05,
|
| 493 |
+
"loss": 0.3375,
|
| 494 |
+
"step": 750
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 4.55,
|
| 498 |
+
"learning_rate": 1.3429522752497226e-05,
|
| 499 |
+
"loss": 0.3195,
|
| 500 |
+
"step": 760
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 4.61,
|
| 504 |
+
"learning_rate": 1.2874583795782464e-05,
|
| 505 |
+
"loss": 0.3878,
|
| 506 |
+
"step": 770
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 4.67,
|
| 510 |
+
"learning_rate": 1.2319644839067704e-05,
|
| 511 |
+
"loss": 0.3555,
|
| 512 |
+
"step": 780
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 4.73,
|
| 516 |
+
"learning_rate": 1.1764705882352942e-05,
|
| 517 |
+
"loss": 0.3177,
|
| 518 |
+
"step": 790
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 4.79,
|
| 522 |
+
"learning_rate": 1.120976692563818e-05,
|
| 523 |
+
"loss": 0.294,
|
| 524 |
+
"step": 800
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 4.85,
|
| 528 |
+
"learning_rate": 1.0654827968923418e-05,
|
| 529 |
+
"loss": 0.3676,
|
| 530 |
+
"step": 810
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 4.91,
|
| 534 |
+
"learning_rate": 1.0099889012208658e-05,
|
| 535 |
+
"loss": 0.3074,
|
| 536 |
+
"step": 820
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 4.97,
|
| 540 |
+
"learning_rate": 9.544950055493896e-06,
|
| 541 |
+
"loss": 0.3622,
|
| 542 |
+
"step": 830
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 5.0,
|
| 546 |
+
"eval_accuracy": 0.8925925925925926,
|
| 547 |
+
"eval_loss": 0.350492000579834,
|
| 548 |
+
"eval_runtime": 70.3191,
|
| 549 |
+
"eval_samples_per_second": 26.877,
|
| 550 |
+
"eval_steps_per_second": 1.692,
|
| 551 |
+
"step": 835
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 5.03,
|
| 555 |
+
"learning_rate": 8.990011098779134e-06,
|
| 556 |
+
"loss": 0.3231,
|
| 557 |
+
"step": 840
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 5.09,
|
| 561 |
+
"learning_rate": 8.435072142064374e-06,
|
| 562 |
+
"loss": 0.3339,
|
| 563 |
+
"step": 850
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 5.15,
|
| 567 |
+
"learning_rate": 7.880133185349611e-06,
|
| 568 |
+
"loss": 0.2668,
|
| 569 |
+
"step": 860
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 5.21,
|
| 573 |
+
"learning_rate": 7.325194228634851e-06,
|
| 574 |
+
"loss": 0.2918,
|
| 575 |
+
"step": 870
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 5.27,
|
| 579 |
+
"learning_rate": 6.77025527192009e-06,
|
| 580 |
+
"loss": 0.2817,
|
| 581 |
+
"step": 880
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 5.33,
|
| 585 |
+
"learning_rate": 6.215316315205328e-06,
|
| 586 |
+
"loss": 0.269,
|
| 587 |
+
"step": 890
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 5.39,
|
| 591 |
+
"learning_rate": 5.660377358490566e-06,
|
| 592 |
+
"loss": 0.3038,
|
| 593 |
+
"step": 900
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 5.45,
|
| 597 |
+
"learning_rate": 5.105438401775805e-06,
|
| 598 |
+
"loss": 0.2663,
|
| 599 |
+
"step": 910
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 5.51,
|
| 603 |
+
"learning_rate": 4.5504994450610434e-06,
|
| 604 |
+
"loss": 0.2752,
|
| 605 |
+
"step": 920
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 5.57,
|
| 609 |
+
"learning_rate": 3.9955604883462825e-06,
|
| 610 |
+
"loss": 0.3432,
|
| 611 |
+
"step": 930
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 5.63,
|
| 615 |
+
"learning_rate": 3.4406215316315207e-06,
|
| 616 |
+
"loss": 0.267,
|
| 617 |
+
"step": 940
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 5.69,
|
| 621 |
+
"learning_rate": 2.8856825749167593e-06,
|
| 622 |
+
"loss": 0.3169,
|
| 623 |
+
"step": 950
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 5.75,
|
| 627 |
+
"learning_rate": 2.330743618201998e-06,
|
| 628 |
+
"loss": 0.2609,
|
| 629 |
+
"step": 960
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 5.81,
|
| 633 |
+
"learning_rate": 1.7758046614872365e-06,
|
| 634 |
+
"loss": 0.3151,
|
| 635 |
+
"step": 970
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 5.87,
|
| 639 |
+
"learning_rate": 1.2208657047724751e-06,
|
| 640 |
+
"loss": 0.2772,
|
| 641 |
+
"step": 980
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 5.93,
|
| 645 |
+
"learning_rate": 6.659267480577136e-07,
|
| 646 |
+
"loss": 0.358,
|
| 647 |
+
"step": 990
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 5.99,
|
| 651 |
+
"learning_rate": 1.1098779134295228e-07,
|
| 652 |
+
"loss": 0.2921,
|
| 653 |
+
"step": 1000
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 6.0,
|
| 657 |
+
"eval_accuracy": 0.8994708994708994,
|
| 658 |
+
"eval_loss": 0.3382543921470642,
|
| 659 |
+
"eval_runtime": 70.2009,
|
| 660 |
+
"eval_samples_per_second": 26.923,
|
| 661 |
+
"eval_steps_per_second": 1.695,
|
| 662 |
+
"step": 1002
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 6.0,
|
| 666 |
+
"step": 1002,
|
| 667 |
+
"total_flos": 1.6742089052991848e+19,
|
| 668 |
+
"train_loss": 0.5761546168379679,
|
| 669 |
+
"train_runtime": 6609.6541,
|
| 670 |
+
"train_samples_per_second": 9.722,
|
| 671 |
+
"train_steps_per_second": 0.152
|
| 672 |
+
}
|
| 673 |
+
],
|
| 674 |
+
"max_steps": 1002,
|
| 675 |
+
"num_train_epochs": 6,
|
| 676 |
+
"total_flos": 1.6742089052991848e+19,
|
| 677 |
+
"trial_name": null,
|
| 678 |
+
"trial_params": null
|
| 679 |
+
}
|