End of training
Browse files- README.md +1 -1
- all_results.json +10 -10
- eval_results.json +5 -5
- runs/Apr19_11-49-01_c9d5019042b9/events.out.tfevents.1713532637.c9d5019042b9.34.1 +3 -0
- train_results.json +6 -6
- trainer_state.json +443 -100
README.md
CHANGED
|
@@ -17,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [microsoft/swin-base-patch4-window7-224-in22k](https://huggingface.co/microsoft/swin-base-patch4-window7-224-in22k) on the imagefolder dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 0.
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [microsoft/swin-base-patch4-window7-224-in22k](https://huggingface.co/microsoft/swin-base-patch4-window7-224-in22k) on the imagefolder dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.0008
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
all_results.json
CHANGED
|
@@ -1,12 +1,12 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 9.
|
| 3 |
-
"eval_loss": 0.
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples_per_second":
|
| 6 |
-
"eval_steps_per_second":
|
| 7 |
-
"total_flos":
|
| 8 |
-
"train_loss": 0.
|
| 9 |
-
"train_runtime":
|
| 10 |
-
"train_samples_per_second":
|
| 11 |
-
"train_steps_per_second": 0.
|
| 12 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 9.87,
|
| 3 |
+
"eval_loss": 0.000785744923632592,
|
| 4 |
+
"eval_runtime": 26.9588,
|
| 5 |
+
"eval_samples_per_second": 58.682,
|
| 6 |
+
"eval_steps_per_second": 0.927,
|
| 7 |
+
"total_flos": 1.100143961239951e+19,
|
| 8 |
+
"train_loss": 0.03171633874193173,
|
| 9 |
+
"train_runtime": 4576.6164,
|
| 10 |
+
"train_samples_per_second": 31.093,
|
| 11 |
+
"train_steps_per_second": 0.12
|
| 12 |
}
|
eval_results.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 9.
|
| 3 |
-
"eval_loss": 0.
|
| 4 |
-
"eval_runtime":
|
| 5 |
-
"eval_samples_per_second":
|
| 6 |
-
"eval_steps_per_second":
|
| 7 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 9.87,
|
| 3 |
+
"eval_loss": 0.000785744923632592,
|
| 4 |
+
"eval_runtime": 26.9588,
|
| 5 |
+
"eval_samples_per_second": 58.682,
|
| 6 |
+
"eval_steps_per_second": 0.927
|
| 7 |
}
|
runs/Apr19_11-49-01_c9d5019042b9/events.out.tfevents.1713532637.c9d5019042b9.34.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1cb67ff6354834d15459e10a754777753b7ebb154ac46348fc4174ad25c1eae9
|
| 3 |
+
size 359
|
train_results.json
CHANGED
|
@@ -1,8 +1,8 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch": 9.
|
| 3 |
-
"total_flos":
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
-
"train_samples_per_second":
|
| 7 |
-
"train_steps_per_second": 0.
|
| 8 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 9.87,
|
| 3 |
+
"total_flos": 1.100143961239951e+19,
|
| 4 |
+
"train_loss": 0.03171633874193173,
|
| 5 |
+
"train_runtime": 4576.6164,
|
| 6 |
+
"train_samples_per_second": 31.093,
|
| 7 |
+
"train_steps_per_second": 0.12
|
| 8 |
}
|
trainer_state.json
CHANGED
|
@@ -1,152 +1,495 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "swin-base-patch4-window7-224-in22k-Kontur-competition-1.3K/checkpoint-
|
| 4 |
-
"epoch": 9.
|
| 5 |
"eval_steps": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
-
"epoch": 0.
|
| 13 |
-
"
|
| 14 |
-
"
|
| 15 |
-
"
|
| 16 |
-
"
|
| 17 |
-
"step": 6
|
| 18 |
},
|
| 19 |
{
|
| 20 |
-
"epoch":
|
| 21 |
-
"grad_norm": 3.
|
| 22 |
-
"learning_rate":
|
| 23 |
-
"loss": 0.
|
| 24 |
-
"step":
|
| 25 |
},
|
| 26 |
{
|
| 27 |
-
"epoch":
|
| 28 |
-
"
|
| 29 |
-
"
|
| 30 |
-
"
|
| 31 |
-
"
|
| 32 |
-
"step": 12
|
| 33 |
},
|
| 34 |
{
|
| 35 |
-
"epoch":
|
| 36 |
-
"
|
| 37 |
-
"
|
| 38 |
-
"
|
| 39 |
-
"
|
| 40 |
-
"step": 18
|
| 41 |
},
|
| 42 |
{
|
| 43 |
-
"epoch":
|
| 44 |
-
"grad_norm":
|
| 45 |
-
"learning_rate":
|
| 46 |
-
"loss": 0.
|
| 47 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
},
|
| 49 |
{
|
| 50 |
"epoch": 4.0,
|
| 51 |
-
"eval_loss": 0.
|
| 52 |
-
"eval_runtime":
|
| 53 |
-
"eval_samples_per_second":
|
| 54 |
-
"eval_steps_per_second":
|
| 55 |
-
"step":
|
| 56 |
},
|
| 57 |
{
|
| 58 |
-
"epoch": 4.
|
| 59 |
-
"grad_norm":
|
| 60 |
-
"learning_rate":
|
| 61 |
-
"loss": 0.
|
| 62 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
},
|
| 64 |
{
|
| 65 |
-
"epoch": 4.
|
| 66 |
-
"eval_loss": 0.
|
| 67 |
-
"eval_runtime":
|
| 68 |
-
"eval_samples_per_second":
|
| 69 |
-
"eval_steps_per_second":
|
| 70 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 71 |
},
|
| 72 |
{
|
| 73 |
"epoch": 5.92,
|
| 74 |
-
"
|
| 75 |
-
"
|
| 76 |
-
"
|
| 77 |
-
"
|
| 78 |
-
"step": 37
|
| 79 |
},
|
| 80 |
{
|
| 81 |
-
"epoch":
|
| 82 |
-
"
|
| 83 |
-
"
|
| 84 |
-
"
|
| 85 |
-
"
|
|
|
|
| 86 |
},
|
| 87 |
{
|
| 88 |
-
"epoch": 6.
|
| 89 |
-
"
|
| 90 |
-
"
|
| 91 |
-
"
|
| 92 |
-
"
|
| 93 |
-
"step": 43
|
| 94 |
},
|
| 95 |
{
|
| 96 |
-
"epoch":
|
| 97 |
-
"grad_norm":
|
| 98 |
-
"learning_rate":
|
| 99 |
-
"loss": 0.
|
| 100 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 101 |
},
|
| 102 |
{
|
| 103 |
"epoch": 8.0,
|
| 104 |
-
"eval_loss": 0.
|
| 105 |
-
"eval_runtime":
|
| 106 |
-
"eval_samples_per_second":
|
| 107 |
-
"eval_steps_per_second":
|
| 108 |
-
"step":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 109 |
},
|
| 110 |
{
|
| 111 |
-
"epoch": 8.
|
| 112 |
-
"
|
| 113 |
-
"
|
| 114 |
-
"
|
| 115 |
-
"
|
| 116 |
-
"step": 56
|
| 117 |
},
|
| 118 |
{
|
| 119 |
-
"epoch":
|
| 120 |
-
"grad_norm": 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 121 |
"learning_rate": 0.0,
|
| 122 |
-
"loss": 0.
|
| 123 |
-
"step":
|
| 124 |
},
|
| 125 |
{
|
| 126 |
-
"epoch": 9.
|
| 127 |
-
"eval_loss": 0.
|
| 128 |
-
"eval_runtime":
|
| 129 |
-
"eval_samples_per_second":
|
| 130 |
-
"eval_steps_per_second":
|
| 131 |
-
"step":
|
| 132 |
},
|
| 133 |
{
|
| 134 |
-
"epoch": 9.
|
| 135 |
-
"step":
|
| 136 |
-
"total_flos":
|
| 137 |
-
"train_loss": 0.
|
| 138 |
-
"train_runtime":
|
| 139 |
-
"train_samples_per_second":
|
| 140 |
-
"train_steps_per_second": 0.
|
| 141 |
}
|
| 142 |
],
|
| 143 |
"logging_steps": 10,
|
| 144 |
-
"max_steps":
|
| 145 |
"num_input_tokens_seen": 0,
|
| 146 |
"num_train_epochs": 10,
|
| 147 |
"save_steps": 500,
|
| 148 |
-
"total_flos":
|
| 149 |
-
"train_batch_size":
|
| 150 |
"trial_name": null,
|
| 151 |
"trial_params": null
|
| 152 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.000785744923632592,
|
| 3 |
+
"best_model_checkpoint": "swin-base-patch4-window7-224-in22k-Kontur-competition-1.3K/checkpoint-334",
|
| 4 |
+
"epoch": 9.865470852017937,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 550,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
{
|
| 12 |
+
"epoch": 0.18,
|
| 13 |
+
"grad_norm": 3.5358009338378906,
|
| 14 |
+
"learning_rate": 9.090909090909091e-06,
|
| 15 |
+
"loss": 0.674,
|
| 16 |
+
"step": 10
|
|
|
|
| 17 |
},
|
| 18 |
{
|
| 19 |
+
"epoch": 0.36,
|
| 20 |
+
"grad_norm": 3.1279354095458984,
|
| 21 |
+
"learning_rate": 1.8181818181818182e-05,
|
| 22 |
+
"loss": 0.4594,
|
| 23 |
+
"step": 20
|
| 24 |
},
|
| 25 |
{
|
| 26 |
+
"epoch": 0.54,
|
| 27 |
+
"grad_norm": 2.273749351501465,
|
| 28 |
+
"learning_rate": 2.7272727272727273e-05,
|
| 29 |
+
"loss": 0.1591,
|
| 30 |
+
"step": 30
|
|
|
|
| 31 |
},
|
| 32 |
{
|
| 33 |
+
"epoch": 0.72,
|
| 34 |
+
"grad_norm": 16.241113662719727,
|
| 35 |
+
"learning_rate": 3.6363636363636364e-05,
|
| 36 |
+
"loss": 0.0193,
|
| 37 |
+
"step": 40
|
|
|
|
| 38 |
},
|
| 39 |
{
|
| 40 |
+
"epoch": 0.9,
|
| 41 |
+
"grad_norm": 5.422853469848633,
|
| 42 |
+
"learning_rate": 4.545454545454546e-05,
|
| 43 |
+
"loss": 0.0593,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.99,
|
| 48 |
+
"eval_loss": 0.029350830242037773,
|
| 49 |
+
"eval_runtime": 35.1661,
|
| 50 |
+
"eval_samples_per_second": 44.986,
|
| 51 |
+
"eval_steps_per_second": 0.711,
|
| 52 |
+
"step": 55
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.08,
|
| 56 |
+
"grad_norm": 6.060277938842773,
|
| 57 |
+
"learning_rate": 4.94949494949495e-05,
|
| 58 |
+
"loss": 0.0433,
|
| 59 |
+
"step": 60
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.26,
|
| 63 |
+
"grad_norm": 0.930535614490509,
|
| 64 |
+
"learning_rate": 4.848484848484849e-05,
|
| 65 |
+
"loss": 0.018,
|
| 66 |
+
"step": 70
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 1.43,
|
| 70 |
+
"grad_norm": 4.823692321777344,
|
| 71 |
+
"learning_rate": 4.7474747474747476e-05,
|
| 72 |
+
"loss": 0.0436,
|
| 73 |
+
"step": 80
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 1.61,
|
| 77 |
+
"grad_norm": 1.3981133699417114,
|
| 78 |
+
"learning_rate": 4.6464646464646464e-05,
|
| 79 |
+
"loss": 0.0184,
|
| 80 |
+
"step": 90
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 1.79,
|
| 84 |
+
"grad_norm": 0.6671331524848938,
|
| 85 |
+
"learning_rate": 4.545454545454546e-05,
|
| 86 |
+
"loss": 0.0201,
|
| 87 |
+
"step": 100
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.97,
|
| 91 |
+
"grad_norm": 1.8355345726013184,
|
| 92 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 93 |
+
"loss": 0.0098,
|
| 94 |
+
"step": 110
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 1.99,
|
| 98 |
+
"eval_loss": 0.03146681934595108,
|
| 99 |
+
"eval_runtime": 28.3147,
|
| 100 |
+
"eval_samples_per_second": 55.872,
|
| 101 |
+
"eval_steps_per_second": 0.883,
|
| 102 |
+
"step": 111
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 2.15,
|
| 106 |
+
"grad_norm": 0.02300291508436203,
|
| 107 |
+
"learning_rate": 4.343434343434344e-05,
|
| 108 |
+
"loss": 0.0131,
|
| 109 |
+
"step": 120
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 2.33,
|
| 113 |
+
"grad_norm": 7.2036871910095215,
|
| 114 |
+
"learning_rate": 4.242424242424243e-05,
|
| 115 |
+
"loss": 0.0215,
|
| 116 |
+
"step": 130
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.51,
|
| 120 |
+
"grad_norm": 1.0436112880706787,
|
| 121 |
+
"learning_rate": 4.141414141414142e-05,
|
| 122 |
+
"loss": 0.0224,
|
| 123 |
+
"step": 140
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.69,
|
| 127 |
+
"grad_norm": 2.077420949935913,
|
| 128 |
+
"learning_rate": 4.0404040404040405e-05,
|
| 129 |
+
"loss": 0.0133,
|
| 130 |
+
"step": 150
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.87,
|
| 134 |
+
"grad_norm": 0.17279785871505737,
|
| 135 |
+
"learning_rate": 3.939393939393939e-05,
|
| 136 |
+
"loss": 0.0066,
|
| 137 |
+
"step": 160
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 3.0,
|
| 141 |
+
"eval_loss": 0.03221270069479942,
|
| 142 |
+
"eval_runtime": 27.4288,
|
| 143 |
+
"eval_samples_per_second": 57.677,
|
| 144 |
+
"eval_steps_per_second": 0.911,
|
| 145 |
+
"step": 167
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 3.05,
|
| 149 |
+
"grad_norm": 0.07281704246997833,
|
| 150 |
+
"learning_rate": 3.838383838383838e-05,
|
| 151 |
+
"loss": 0.0034,
|
| 152 |
+
"step": 170
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.23,
|
| 156 |
+
"grad_norm": 0.06530273705720901,
|
| 157 |
+
"learning_rate": 3.7373737373737376e-05,
|
| 158 |
+
"loss": 0.0038,
|
| 159 |
+
"step": 180
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 3.41,
|
| 163 |
+
"grad_norm": 0.2904812693595886,
|
| 164 |
+
"learning_rate": 3.6363636363636364e-05,
|
| 165 |
+
"loss": 0.0045,
|
| 166 |
+
"step": 190
|
| 167 |
+
},
|
| 168 |
+
{
|
| 169 |
+
"epoch": 3.59,
|
| 170 |
+
"grad_norm": 0.01948358118534088,
|
| 171 |
+
"learning_rate": 3.535353535353535e-05,
|
| 172 |
+
"loss": 0.0052,
|
| 173 |
+
"step": 200
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 3.77,
|
| 177 |
+
"grad_norm": 0.24817952513694763,
|
| 178 |
+
"learning_rate": 3.434343434343435e-05,
|
| 179 |
+
"loss": 0.004,
|
| 180 |
+
"step": 210
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 3.95,
|
| 184 |
+
"grad_norm": 1.0436078310012817,
|
| 185 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 186 |
+
"loss": 0.0179,
|
| 187 |
+
"step": 220
|
| 188 |
},
|
| 189 |
{
|
| 190 |
"epoch": 4.0,
|
| 191 |
+
"eval_loss": 0.006767116021364927,
|
| 192 |
+
"eval_runtime": 27.0156,
|
| 193 |
+
"eval_samples_per_second": 58.559,
|
| 194 |
+
"eval_steps_per_second": 0.925,
|
| 195 |
+
"step": 223
|
| 196 |
},
|
| 197 |
{
|
| 198 |
+
"epoch": 4.13,
|
| 199 |
+
"grad_norm": 0.4218502342700958,
|
| 200 |
+
"learning_rate": 3.232323232323233e-05,
|
| 201 |
+
"loss": 0.011,
|
| 202 |
+
"step": 230
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 4.3,
|
| 206 |
+
"grad_norm": 1.2248862981796265,
|
| 207 |
+
"learning_rate": 3.131313131313132e-05,
|
| 208 |
+
"loss": 0.0043,
|
| 209 |
+
"step": 240
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 4.48,
|
| 213 |
+
"grad_norm": 3.450435161590576,
|
| 214 |
+
"learning_rate": 3.0303030303030306e-05,
|
| 215 |
+
"loss": 0.0115,
|
| 216 |
+
"step": 250
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 4.66,
|
| 220 |
+
"grad_norm": 0.10841131955385208,
|
| 221 |
+
"learning_rate": 2.9292929292929294e-05,
|
| 222 |
+
"loss": 0.0189,
|
| 223 |
+
"step": 260
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 4.84,
|
| 227 |
+
"grad_norm": 2.503843307495117,
|
| 228 |
+
"learning_rate": 2.8282828282828282e-05,
|
| 229 |
+
"loss": 0.0078,
|
| 230 |
+
"step": 270
|
| 231 |
},
|
| 232 |
{
|
| 233 |
+
"epoch": 4.99,
|
| 234 |
+
"eval_loss": 0.0032722819596529007,
|
| 235 |
+
"eval_runtime": 27.2058,
|
| 236 |
+
"eval_samples_per_second": 58.149,
|
| 237 |
+
"eval_steps_per_second": 0.919,
|
| 238 |
+
"step": 278
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 5.02,
|
| 242 |
+
"grad_norm": 1.6022576093673706,
|
| 243 |
+
"learning_rate": 2.7272727272727273e-05,
|
| 244 |
+
"loss": 0.0069,
|
| 245 |
+
"step": 280
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 5.2,
|
| 249 |
+
"grad_norm": 0.23544766008853912,
|
| 250 |
+
"learning_rate": 2.6262626262626268e-05,
|
| 251 |
+
"loss": 0.0077,
|
| 252 |
+
"step": 290
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 5.38,
|
| 256 |
+
"grad_norm": 1.4383817911148071,
|
| 257 |
+
"learning_rate": 2.5252525252525256e-05,
|
| 258 |
+
"loss": 0.002,
|
| 259 |
+
"step": 300
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 5.56,
|
| 263 |
+
"grad_norm": 0.105356365442276,
|
| 264 |
+
"learning_rate": 2.4242424242424244e-05,
|
| 265 |
+
"loss": 0.0017,
|
| 266 |
+
"step": 310
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 5.74,
|
| 270 |
+
"grad_norm": 0.7489631175994873,
|
| 271 |
+
"learning_rate": 2.3232323232323232e-05,
|
| 272 |
+
"loss": 0.0012,
|
| 273 |
+
"step": 320
|
| 274 |
},
|
| 275 |
{
|
| 276 |
"epoch": 5.92,
|
| 277 |
+
"grad_norm": 1.5391371250152588,
|
| 278 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 279 |
+
"loss": 0.0015,
|
| 280 |
+
"step": 330
|
|
|
|
| 281 |
},
|
| 282 |
{
|
| 283 |
+
"epoch": 5.99,
|
| 284 |
+
"eval_loss": 0.000785744923632592,
|
| 285 |
+
"eval_runtime": 26.43,
|
| 286 |
+
"eval_samples_per_second": 59.856,
|
| 287 |
+
"eval_steps_per_second": 0.946,
|
| 288 |
+
"step": 334
|
| 289 |
},
|
| 290 |
{
|
| 291 |
+
"epoch": 6.1,
|
| 292 |
+
"grad_norm": 0.05875357240438461,
|
| 293 |
+
"learning_rate": 2.1212121212121215e-05,
|
| 294 |
+
"loss": 0.002,
|
| 295 |
+
"step": 340
|
|
|
|
| 296 |
},
|
| 297 |
{
|
| 298 |
+
"epoch": 6.28,
|
| 299 |
+
"grad_norm": 2.8288419246673584,
|
| 300 |
+
"learning_rate": 2.0202020202020203e-05,
|
| 301 |
+
"loss": 0.0036,
|
| 302 |
+
"step": 350
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 6.46,
|
| 306 |
+
"grad_norm": 0.018499189987778664,
|
| 307 |
+
"learning_rate": 1.919191919191919e-05,
|
| 308 |
+
"loss": 0.0007,
|
| 309 |
+
"step": 360
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 6.64,
|
| 313 |
+
"grad_norm": 0.00614353409036994,
|
| 314 |
+
"learning_rate": 1.8181818181818182e-05,
|
| 315 |
+
"loss": 0.0036,
|
| 316 |
+
"step": 370
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 6.82,
|
| 320 |
+
"grad_norm": 0.01476567517966032,
|
| 321 |
+
"learning_rate": 1.7171717171717173e-05,
|
| 322 |
+
"loss": 0.0041,
|
| 323 |
+
"step": 380
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 7.0,
|
| 327 |
+
"grad_norm": 0.16370485723018646,
|
| 328 |
+
"learning_rate": 1.6161616161616165e-05,
|
| 329 |
+
"loss": 0.0017,
|
| 330 |
+
"step": 390
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 7.0,
|
| 334 |
+
"eval_loss": 0.007814141921699047,
|
| 335 |
+
"eval_runtime": 26.7049,
|
| 336 |
+
"eval_samples_per_second": 59.24,
|
| 337 |
+
"eval_steps_per_second": 0.936,
|
| 338 |
+
"step": 390
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 7.17,
|
| 342 |
+
"grad_norm": 0.09164857864379883,
|
| 343 |
+
"learning_rate": 1.5151515151515153e-05,
|
| 344 |
+
"loss": 0.0004,
|
| 345 |
+
"step": 400
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 7.35,
|
| 349 |
+
"grad_norm": 0.19655057787895203,
|
| 350 |
+
"learning_rate": 1.4141414141414141e-05,
|
| 351 |
+
"loss": 0.0005,
|
| 352 |
+
"step": 410
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 7.53,
|
| 356 |
+
"grad_norm": 0.0026813277509063482,
|
| 357 |
+
"learning_rate": 1.3131313131313134e-05,
|
| 358 |
+
"loss": 0.0003,
|
| 359 |
+
"step": 420
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 7.71,
|
| 363 |
+
"grad_norm": 0.7142350673675537,
|
| 364 |
+
"learning_rate": 1.2121212121212122e-05,
|
| 365 |
+
"loss": 0.0005,
|
| 366 |
+
"step": 430
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 7.89,
|
| 370 |
+
"grad_norm": 0.004237661603838205,
|
| 371 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 372 |
+
"loss": 0.0008,
|
| 373 |
+
"step": 440
|
| 374 |
},
|
| 375 |
{
|
| 376 |
"epoch": 8.0,
|
| 377 |
+
"eval_loss": 0.002696491777896881,
|
| 378 |
+
"eval_runtime": 26.1158,
|
| 379 |
+
"eval_samples_per_second": 60.576,
|
| 380 |
+
"eval_steps_per_second": 0.957,
|
| 381 |
+
"step": 446
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 8.07,
|
| 385 |
+
"grad_norm": 0.010493074543774128,
|
| 386 |
+
"learning_rate": 1.0101010101010101e-05,
|
| 387 |
+
"loss": 0.0001,
|
| 388 |
+
"step": 450
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 8.25,
|
| 392 |
+
"grad_norm": 0.015726672485470772,
|
| 393 |
+
"learning_rate": 9.090909090909091e-06,
|
| 394 |
+
"loss": 0.0003,
|
| 395 |
+
"step": 460
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 8.43,
|
| 399 |
+
"grad_norm": 2.051889657974243,
|
| 400 |
+
"learning_rate": 8.080808080808082e-06,
|
| 401 |
+
"loss": 0.0023,
|
| 402 |
+
"step": 470
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 8.61,
|
| 406 |
+
"grad_norm": 0.001357908477075398,
|
| 407 |
+
"learning_rate": 7.0707070707070704e-06,
|
| 408 |
+
"loss": 0.0018,
|
| 409 |
+
"step": 480
|
| 410 |
},
|
| 411 |
{
|
| 412 |
+
"epoch": 8.79,
|
| 413 |
+
"grad_norm": 0.0037935995496809483,
|
| 414 |
+
"learning_rate": 6.060606060606061e-06,
|
| 415 |
+
"loss": 0.0001,
|
| 416 |
+
"step": 490
|
|
|
|
| 417 |
},
|
| 418 |
{
|
| 419 |
+
"epoch": 8.97,
|
| 420 |
+
"grad_norm": 0.014398468658328056,
|
| 421 |
+
"learning_rate": 5.050505050505051e-06,
|
| 422 |
+
"loss": 0.0019,
|
| 423 |
+
"step": 500
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 8.99,
|
| 427 |
+
"eval_loss": 0.0011395993642508984,
|
| 428 |
+
"eval_runtime": 26.5634,
|
| 429 |
+
"eval_samples_per_second": 59.556,
|
| 430 |
+
"eval_steps_per_second": 0.941,
|
| 431 |
+
"step": 501
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 9.15,
|
| 435 |
+
"grad_norm": 0.04990648478269577,
|
| 436 |
+
"learning_rate": 4.040404040404041e-06,
|
| 437 |
+
"loss": 0.0008,
|
| 438 |
+
"step": 510
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 9.33,
|
| 442 |
+
"grad_norm": 0.009485547430813313,
|
| 443 |
+
"learning_rate": 3.0303030303030305e-06,
|
| 444 |
+
"loss": 0.0024,
|
| 445 |
+
"step": 520
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 9.51,
|
| 449 |
+
"grad_norm": 0.007252832874655724,
|
| 450 |
+
"learning_rate": 2.0202020202020206e-06,
|
| 451 |
+
"loss": 0.0007,
|
| 452 |
+
"step": 530
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 9.69,
|
| 456 |
+
"grad_norm": 0.007266469299793243,
|
| 457 |
+
"learning_rate": 1.0101010101010103e-06,
|
| 458 |
+
"loss": 0.0003,
|
| 459 |
+
"step": 540
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 9.87,
|
| 463 |
+
"grad_norm": 0.01849411241710186,
|
| 464 |
"learning_rate": 0.0,
|
| 465 |
+
"loss": 0.0014,
|
| 466 |
+
"step": 550
|
| 467 |
},
|
| 468 |
{
|
| 469 |
+
"epoch": 9.87,
|
| 470 |
+
"eval_loss": 0.0035754255950450897,
|
| 471 |
+
"eval_runtime": 37.9782,
|
| 472 |
+
"eval_samples_per_second": 41.655,
|
| 473 |
+
"eval_steps_per_second": 0.658,
|
| 474 |
+
"step": 550
|
| 475 |
},
|
| 476 |
{
|
| 477 |
+
"epoch": 9.87,
|
| 478 |
+
"step": 550,
|
| 479 |
+
"total_flos": 1.100143961239951e+19,
|
| 480 |
+
"train_loss": 0.03171633874193173,
|
| 481 |
+
"train_runtime": 4576.6164,
|
| 482 |
+
"train_samples_per_second": 31.093,
|
| 483 |
+
"train_steps_per_second": 0.12
|
| 484 |
}
|
| 485 |
],
|
| 486 |
"logging_steps": 10,
|
| 487 |
+
"max_steps": 550,
|
| 488 |
"num_input_tokens_seen": 0,
|
| 489 |
"num_train_epochs": 10,
|
| 490 |
"save_steps": 500,
|
| 491 |
+
"total_flos": 1.100143961239951e+19,
|
| 492 |
+
"train_batch_size": 64,
|
| 493 |
"trial_name": null,
|
| 494 |
"trial_params": null
|
| 495 |
}
|