End of training
Browse files- README.md +2 -2
- all_results.json +8 -0
- test_results.json +8 -0
- trainer_state.json +782 -0
README.md
CHANGED
|
@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss: 0.
|
| 22 |
-
- Accuracy: 0.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.7568
|
| 22 |
+
- Accuracy: 0.8080
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.040055803571429,
|
| 3 |
+
"eval_accuracy": 0.8079937304075235,
|
| 4 |
+
"eval_loss": 0.7567912340164185,
|
| 5 |
+
"eval_runtime": 438.2415,
|
| 6 |
+
"eval_samples_per_second": 2.912,
|
| 7 |
+
"eval_steps_per_second": 1.456
|
| 8 |
+
}
|
test_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.040055803571429,
|
| 3 |
+
"eval_accuracy": 0.8079937304075235,
|
| 4 |
+
"eval_loss": 0.7567912340164185,
|
| 5 |
+
"eval_runtime": 438.2415,
|
| 6 |
+
"eval_samples_per_second": 2.912,
|
| 7 |
+
"eval_steps_per_second": 1.456
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,782 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9941176470588236,
|
| 3 |
+
"best_model_checkpoint": "/media/cse/HDD/Shawon/shawon/MY DATA/VideoMAE_BdSLW60_FrameRateCorrected_withoutAug/checkpoint-4488",
|
| 4 |
+
"epoch": 9.040055803571429,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 8977,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.004464285714285714,
|
| 13 |
+
"grad_norm": 9.476576805114746,
|
| 14 |
+
"learning_rate": 2.2321428571428573e-06,
|
| 15 |
+
"loss": 4.1434,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.008928571428571428,
|
| 20 |
+
"grad_norm": 11.913525581359863,
|
| 21 |
+
"learning_rate": 4.464285714285715e-06,
|
| 22 |
+
"loss": 4.0979,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.013392857142857142,
|
| 27 |
+
"grad_norm": 10.909424781799316,
|
| 28 |
+
"learning_rate": 6.696428571428572e-06,
|
| 29 |
+
"loss": 4.0707,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.017857142857142856,
|
| 34 |
+
"grad_norm": 13.112800598144531,
|
| 35 |
+
"learning_rate": 8.92857142857143e-06,
|
| 36 |
+
"loss": 4.0628,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.022321428571428572,
|
| 41 |
+
"grad_norm": 12.672718048095703,
|
| 42 |
+
"learning_rate": 1.1160714285714287e-05,
|
| 43 |
+
"loss": 4.0055,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.026785714285714284,
|
| 48 |
+
"grad_norm": 13.710538864135742,
|
| 49 |
+
"learning_rate": 1.3392857142857144e-05,
|
| 50 |
+
"loss": 3.8428,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.03125,
|
| 55 |
+
"grad_norm": 15.481912612915039,
|
| 56 |
+
"learning_rate": 1.5625e-05,
|
| 57 |
+
"loss": 3.6124,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.03571428571428571,
|
| 62 |
+
"grad_norm": 15.442792892456055,
|
| 63 |
+
"learning_rate": 1.785714285714286e-05,
|
| 64 |
+
"loss": 3.1009,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.040044642857142855,
|
| 69 |
+
"eval_accuracy": 0.508235294117647,
|
| 70 |
+
"eval_loss": 2.399294137954712,
|
| 71 |
+
"eval_runtime": 297.3318,
|
| 72 |
+
"eval_samples_per_second": 2.859,
|
| 73 |
+
"eval_steps_per_second": 1.429,
|
| 74 |
+
"step": 897
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 1.0001004464285714,
|
| 78 |
+
"grad_norm": 21.445680618286133,
|
| 79 |
+
"learning_rate": 2.0089285714285717e-05,
|
| 80 |
+
"loss": 2.6823,
|
| 81 |
+
"step": 900
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 1.0045647321428572,
|
| 85 |
+
"grad_norm": 19.279504776000977,
|
| 86 |
+
"learning_rate": 2.2321428571428575e-05,
|
| 87 |
+
"loss": 2.0408,
|
| 88 |
+
"step": 1000
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 1.0090290178571428,
|
| 92 |
+
"grad_norm": 16.168643951416016,
|
| 93 |
+
"learning_rate": 2.455357142857143e-05,
|
| 94 |
+
"loss": 1.6718,
|
| 95 |
+
"step": 1100
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 1.0134933035714286,
|
| 99 |
+
"grad_norm": 10.949524879455566,
|
| 100 |
+
"learning_rate": 2.6785714285714288e-05,
|
| 101 |
+
"loss": 1.1872,
|
| 102 |
+
"step": 1200
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 1.0179575892857142,
|
| 106 |
+
"grad_norm": 14.425607681274414,
|
| 107 |
+
"learning_rate": 2.9017857142857146e-05,
|
| 108 |
+
"loss": 0.8375,
|
| 109 |
+
"step": 1300
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 1.022421875,
|
| 113 |
+
"grad_norm": 12.171178817749023,
|
| 114 |
+
"learning_rate": 3.125e-05,
|
| 115 |
+
"loss": 0.6268,
|
| 116 |
+
"step": 1400
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.0268861607142856,
|
| 120 |
+
"grad_norm": 6.816460609436035,
|
| 121 |
+
"learning_rate": 3.348214285714286e-05,
|
| 122 |
+
"loss": 0.4694,
|
| 123 |
+
"step": 1500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 1.0313504464285714,
|
| 127 |
+
"grad_norm": 4.130532741546631,
|
| 128 |
+
"learning_rate": 3.571428571428572e-05,
|
| 129 |
+
"loss": 0.3985,
|
| 130 |
+
"step": 1600
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 1.0358147321428572,
|
| 134 |
+
"grad_norm": 3.680629014968872,
|
| 135 |
+
"learning_rate": 3.794642857142857e-05,
|
| 136 |
+
"loss": 0.3152,
|
| 137 |
+
"step": 1700
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.0400558035714287,
|
| 141 |
+
"eval_accuracy": 0.9611764705882353,
|
| 142 |
+
"eval_loss": 0.21172475814819336,
|
| 143 |
+
"eval_runtime": 306.8196,
|
| 144 |
+
"eval_samples_per_second": 2.77,
|
| 145 |
+
"eval_steps_per_second": 1.385,
|
| 146 |
+
"step": 1795
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 2.000200892857143,
|
| 150 |
+
"grad_norm": 1.7801744937896729,
|
| 151 |
+
"learning_rate": 4.017857142857143e-05,
|
| 152 |
+
"loss": 0.2748,
|
| 153 |
+
"step": 1800
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 2.0046651785714285,
|
| 157 |
+
"grad_norm": 4.094645023345947,
|
| 158 |
+
"learning_rate": 4.2410714285714285e-05,
|
| 159 |
+
"loss": 0.1463,
|
| 160 |
+
"step": 1900
|
| 161 |
+
},
|
| 162 |
+
{
|
| 163 |
+
"epoch": 2.0091294642857145,
|
| 164 |
+
"grad_norm": 0.7033438682556152,
|
| 165 |
+
"learning_rate": 4.464285714285715e-05,
|
| 166 |
+
"loss": 0.1269,
|
| 167 |
+
"step": 2000
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 2.01359375,
|
| 171 |
+
"grad_norm": 0.5256723761558533,
|
| 172 |
+
"learning_rate": 4.6875e-05,
|
| 173 |
+
"loss": 0.0921,
|
| 174 |
+
"step": 2100
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 2.0180580357142857,
|
| 178 |
+
"grad_norm": 1.1959021091461182,
|
| 179 |
+
"learning_rate": 4.910714285714286e-05,
|
| 180 |
+
"loss": 0.0496,
|
| 181 |
+
"step": 2200
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 2.0225223214285712,
|
| 185 |
+
"grad_norm": 1.346044659614563,
|
| 186 |
+
"learning_rate": 4.985119047619048e-05,
|
| 187 |
+
"loss": 0.0801,
|
| 188 |
+
"step": 2300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 2.0269866071428573,
|
| 192 |
+
"grad_norm": 0.21942313015460968,
|
| 193 |
+
"learning_rate": 4.960317460317461e-05,
|
| 194 |
+
"loss": 0.1066,
|
| 195 |
+
"step": 2400
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 2.031450892857143,
|
| 199 |
+
"grad_norm": 0.10540590435266495,
|
| 200 |
+
"learning_rate": 4.9355158730158735e-05,
|
| 201 |
+
"loss": 0.0639,
|
| 202 |
+
"step": 2500
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 2.0359151785714285,
|
| 206 |
+
"grad_norm": 0.18206317722797394,
|
| 207 |
+
"learning_rate": 4.910714285714286e-05,
|
| 208 |
+
"loss": 0.0363,
|
| 209 |
+
"step": 2600
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 2.0400669642857143,
|
| 213 |
+
"eval_accuracy": 0.991764705882353,
|
| 214 |
+
"eval_loss": 0.037506770342588425,
|
| 215 |
+
"eval_runtime": 294.0638,
|
| 216 |
+
"eval_samples_per_second": 2.891,
|
| 217 |
+
"eval_steps_per_second": 1.445,
|
| 218 |
+
"step": 2693
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.000301339285714,
|
| 222 |
+
"grad_norm": 0.18602599203586578,
|
| 223 |
+
"learning_rate": 4.8859126984126984e-05,
|
| 224 |
+
"loss": 0.0633,
|
| 225 |
+
"step": 2700
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.004765625,
|
| 229 |
+
"grad_norm": 0.13055439293384552,
|
| 230 |
+
"learning_rate": 4.8611111111111115e-05,
|
| 231 |
+
"loss": 0.0147,
|
| 232 |
+
"step": 2800
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.0092299107142857,
|
| 236 |
+
"grad_norm": 0.21550215780735016,
|
| 237 |
+
"learning_rate": 4.836309523809524e-05,
|
| 238 |
+
"loss": 0.0339,
|
| 239 |
+
"step": 2900
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.0136941964285713,
|
| 243 |
+
"grad_norm": 0.17052847146987915,
|
| 244 |
+
"learning_rate": 4.811507936507937e-05,
|
| 245 |
+
"loss": 0.0361,
|
| 246 |
+
"step": 3000
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.0181584821428573,
|
| 250 |
+
"grad_norm": 0.11729184538125992,
|
| 251 |
+
"learning_rate": 4.7867063492063496e-05,
|
| 252 |
+
"loss": 0.0339,
|
| 253 |
+
"step": 3100
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.022622767857143,
|
| 257 |
+
"grad_norm": 0.02288275957107544,
|
| 258 |
+
"learning_rate": 4.761904761904762e-05,
|
| 259 |
+
"loss": 0.0384,
|
| 260 |
+
"step": 3200
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 3.0270870535714285,
|
| 264 |
+
"grad_norm": 0.026718221604824066,
|
| 265 |
+
"learning_rate": 4.7371031746031745e-05,
|
| 266 |
+
"loss": 0.0269,
|
| 267 |
+
"step": 3300
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 3.031551339285714,
|
| 271 |
+
"grad_norm": 16.089458465576172,
|
| 272 |
+
"learning_rate": 4.7123015873015876e-05,
|
| 273 |
+
"loss": 0.0447,
|
| 274 |
+
"step": 3400
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 3.036015625,
|
| 278 |
+
"grad_norm": 0.6215272545814514,
|
| 279 |
+
"learning_rate": 4.6875e-05,
|
| 280 |
+
"loss": 0.0431,
|
| 281 |
+
"step": 3500
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 3.040078125,
|
| 285 |
+
"eval_accuracy": 0.9764705882352941,
|
| 286 |
+
"eval_loss": 0.08410066366195679,
|
| 287 |
+
"eval_runtime": 296.1888,
|
| 288 |
+
"eval_samples_per_second": 2.87,
|
| 289 |
+
"eval_steps_per_second": 1.435,
|
| 290 |
+
"step": 3591
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 4.000401785714286,
|
| 294 |
+
"grad_norm": 6.967636585235596,
|
| 295 |
+
"learning_rate": 4.662698412698413e-05,
|
| 296 |
+
"loss": 0.0237,
|
| 297 |
+
"step": 3600
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 4.004866071428571,
|
| 301 |
+
"grad_norm": 0.018726671114563942,
|
| 302 |
+
"learning_rate": 4.637896825396826e-05,
|
| 303 |
+
"loss": 0.0286,
|
| 304 |
+
"step": 3700
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 4.009330357142857,
|
| 308 |
+
"grad_norm": 0.027330530807375908,
|
| 309 |
+
"learning_rate": 4.613095238095239e-05,
|
| 310 |
+
"loss": 0.0216,
|
| 311 |
+
"step": 3800
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 4.0137946428571425,
|
| 315 |
+
"grad_norm": 0.08880900591611862,
|
| 316 |
+
"learning_rate": 4.5882936507936506e-05,
|
| 317 |
+
"loss": 0.0148,
|
| 318 |
+
"step": 3900
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 4.018258928571429,
|
| 322 |
+
"grad_norm": 0.01726541668176651,
|
| 323 |
+
"learning_rate": 4.563492063492064e-05,
|
| 324 |
+
"loss": 0.0341,
|
| 325 |
+
"step": 4000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 4.0227232142857146,
|
| 329 |
+
"grad_norm": 0.5136727094650269,
|
| 330 |
+
"learning_rate": 4.538690476190476e-05,
|
| 331 |
+
"loss": 0.0286,
|
| 332 |
+
"step": 4100
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 4.0271875,
|
| 336 |
+
"grad_norm": 0.25709015130996704,
|
| 337 |
+
"learning_rate": 4.5138888888888894e-05,
|
| 338 |
+
"loss": 0.0437,
|
| 339 |
+
"step": 4200
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 4.031651785714286,
|
| 343 |
+
"grad_norm": 1.26307213306427,
|
| 344 |
+
"learning_rate": 4.489087301587302e-05,
|
| 345 |
+
"loss": 0.0636,
|
| 346 |
+
"step": 4300
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 4.036116071428571,
|
| 350 |
+
"grad_norm": 0.04352056235074997,
|
| 351 |
+
"learning_rate": 4.464285714285715e-05,
|
| 352 |
+
"loss": 0.0554,
|
| 353 |
+
"step": 4400
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 4.040044642857143,
|
| 357 |
+
"eval_accuracy": 0.9941176470588236,
|
| 358 |
+
"eval_loss": 0.03231202811002731,
|
| 359 |
+
"eval_runtime": 284.1927,
|
| 360 |
+
"eval_samples_per_second": 2.991,
|
| 361 |
+
"eval_steps_per_second": 1.495,
|
| 362 |
+
"step": 4488
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 5.000502232142857,
|
| 366 |
+
"grad_norm": 0.015468730591237545,
|
| 367 |
+
"learning_rate": 4.439484126984127e-05,
|
| 368 |
+
"loss": 0.0489,
|
| 369 |
+
"step": 4500
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 5.0049665178571425,
|
| 373 |
+
"grad_norm": 0.008214977569878101,
|
| 374 |
+
"learning_rate": 4.41468253968254e-05,
|
| 375 |
+
"loss": 0.0224,
|
| 376 |
+
"step": 4600
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 5.009430803571429,
|
| 380 |
+
"grad_norm": 0.709925651550293,
|
| 381 |
+
"learning_rate": 4.3898809523809523e-05,
|
| 382 |
+
"loss": 0.0131,
|
| 383 |
+
"step": 4700
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 5.013895089285715,
|
| 387 |
+
"grad_norm": 6.61177396774292,
|
| 388 |
+
"learning_rate": 4.3650793650793655e-05,
|
| 389 |
+
"loss": 0.0122,
|
| 390 |
+
"step": 4800
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 5.018359375,
|
| 394 |
+
"grad_norm": 0.024309329688549042,
|
| 395 |
+
"learning_rate": 4.340277777777778e-05,
|
| 396 |
+
"loss": 0.02,
|
| 397 |
+
"step": 4900
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 5.022823660714286,
|
| 401 |
+
"grad_norm": 0.0186802726238966,
|
| 402 |
+
"learning_rate": 4.315476190476191e-05,
|
| 403 |
+
"loss": 0.0262,
|
| 404 |
+
"step": 5000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 5.027287946428571,
|
| 408 |
+
"grad_norm": 0.0063638836145401,
|
| 409 |
+
"learning_rate": 4.290674603174603e-05,
|
| 410 |
+
"loss": 0.0202,
|
| 411 |
+
"step": 5100
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 5.031752232142857,
|
| 415 |
+
"grad_norm": 0.0073557039722800255,
|
| 416 |
+
"learning_rate": 4.265873015873016e-05,
|
| 417 |
+
"loss": 0.0266,
|
| 418 |
+
"step": 5200
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 5.0362165178571425,
|
| 422 |
+
"grad_norm": 13.825119972229004,
|
| 423 |
+
"learning_rate": 4.2410714285714285e-05,
|
| 424 |
+
"loss": 0.0467,
|
| 425 |
+
"step": 5300
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 5.040055803571429,
|
| 429 |
+
"eval_accuracy": 0.9835294117647059,
|
| 430 |
+
"eval_loss": 0.06136510148644447,
|
| 431 |
+
"eval_runtime": 284.8463,
|
| 432 |
+
"eval_samples_per_second": 2.984,
|
| 433 |
+
"eval_steps_per_second": 1.492,
|
| 434 |
+
"step": 5386
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 6.000602678571428,
|
| 438 |
+
"grad_norm": 65.57583618164062,
|
| 439 |
+
"learning_rate": 4.2162698412698416e-05,
|
| 440 |
+
"loss": 0.022,
|
| 441 |
+
"step": 5400
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 6.005066964285715,
|
| 445 |
+
"grad_norm": 0.026758452877402306,
|
| 446 |
+
"learning_rate": 4.191468253968254e-05,
|
| 447 |
+
"loss": 0.0287,
|
| 448 |
+
"step": 5500
|
| 449 |
+
},
|
| 450 |
+
{
|
| 451 |
+
"epoch": 6.00953125,
|
| 452 |
+
"grad_norm": 0.013503998517990112,
|
| 453 |
+
"learning_rate": 4.166666666666667e-05,
|
| 454 |
+
"loss": 0.04,
|
| 455 |
+
"step": 5600
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 6.013995535714286,
|
| 459 |
+
"grad_norm": 0.012964209541678429,
|
| 460 |
+
"learning_rate": 4.14186507936508e-05,
|
| 461 |
+
"loss": 0.0333,
|
| 462 |
+
"step": 5700
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 6.018459821428571,
|
| 466 |
+
"grad_norm": 0.009373425506055355,
|
| 467 |
+
"learning_rate": 4.117063492063492e-05,
|
| 468 |
+
"loss": 0.0163,
|
| 469 |
+
"step": 5800
|
| 470 |
+
},
|
| 471 |
+
{
|
| 472 |
+
"epoch": 6.022924107142857,
|
| 473 |
+
"grad_norm": 0.6834176778793335,
|
| 474 |
+
"learning_rate": 4.0922619047619046e-05,
|
| 475 |
+
"loss": 0.0482,
|
| 476 |
+
"step": 5900
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 6.027388392857143,
|
| 480 |
+
"grad_norm": 0.034619417041540146,
|
| 481 |
+
"learning_rate": 4.067460317460318e-05,
|
| 482 |
+
"loss": 0.0339,
|
| 483 |
+
"step": 6000
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 6.031852678571428,
|
| 487 |
+
"grad_norm": 0.031068073585629463,
|
| 488 |
+
"learning_rate": 4.04265873015873e-05,
|
| 489 |
+
"loss": 0.0399,
|
| 490 |
+
"step": 6100
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 6.036316964285715,
|
| 494 |
+
"grad_norm": 0.13247177004814148,
|
| 495 |
+
"learning_rate": 4.017857142857143e-05,
|
| 496 |
+
"loss": 0.0324,
|
| 497 |
+
"step": 6200
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 6.040066964285714,
|
| 501 |
+
"eval_accuracy": 0.9858823529411764,
|
| 502 |
+
"eval_loss": 0.09442107379436493,
|
| 503 |
+
"eval_runtime": 294.806,
|
| 504 |
+
"eval_samples_per_second": 2.883,
|
| 505 |
+
"eval_steps_per_second": 1.442,
|
| 506 |
+
"step": 6284
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 7.000703125,
|
| 510 |
+
"grad_norm": 0.008008907549083233,
|
| 511 |
+
"learning_rate": 3.993055555555556e-05,
|
| 512 |
+
"loss": 0.0342,
|
| 513 |
+
"step": 6300
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 7.005167410714286,
|
| 517 |
+
"grad_norm": 0.057924672961235046,
|
| 518 |
+
"learning_rate": 3.968253968253968e-05,
|
| 519 |
+
"loss": 0.008,
|
| 520 |
+
"step": 6400
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 7.009631696428571,
|
| 524 |
+
"grad_norm": 0.0028113469015806913,
|
| 525 |
+
"learning_rate": 3.943452380952381e-05,
|
| 526 |
+
"loss": 0.0238,
|
| 527 |
+
"step": 6500
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 7.014095982142857,
|
| 531 |
+
"grad_norm": 0.006544953212141991,
|
| 532 |
+
"learning_rate": 3.918650793650794e-05,
|
| 533 |
+
"loss": 0.0081,
|
| 534 |
+
"step": 6600
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 7.018560267857143,
|
| 538 |
+
"grad_norm": 0.007798121310770512,
|
| 539 |
+
"learning_rate": 3.893849206349206e-05,
|
| 540 |
+
"loss": 0.0178,
|
| 541 |
+
"step": 6700
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 7.023024553571428,
|
| 545 |
+
"grad_norm": 0.011922234669327736,
|
| 546 |
+
"learning_rate": 3.8690476190476195e-05,
|
| 547 |
+
"loss": 0.0146,
|
| 548 |
+
"step": 6800
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 7.027488839285715,
|
| 552 |
+
"grad_norm": 0.02987760305404663,
|
| 553 |
+
"learning_rate": 3.844246031746032e-05,
|
| 554 |
+
"loss": 0.0438,
|
| 555 |
+
"step": 6900
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 7.031953125,
|
| 559 |
+
"grad_norm": 0.010242101736366749,
|
| 560 |
+
"learning_rate": 3.8194444444444444e-05,
|
| 561 |
+
"loss": 0.0614,
|
| 562 |
+
"step": 7000
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 7.036417410714286,
|
| 566 |
+
"grad_norm": 0.030809585005044937,
|
| 567 |
+
"learning_rate": 3.794642857142857e-05,
|
| 568 |
+
"loss": 0.0381,
|
| 569 |
+
"step": 7100
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 7.040078125,
|
| 573 |
+
"eval_accuracy": 0.9870588235294118,
|
| 574 |
+
"eval_loss": 0.05825329199433327,
|
| 575 |
+
"eval_runtime": 282.1803,
|
| 576 |
+
"eval_samples_per_second": 3.012,
|
| 577 |
+
"eval_steps_per_second": 1.506,
|
| 578 |
+
"step": 7182
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 8.000803571428571,
|
| 582 |
+
"grad_norm": 1.1858717203140259,
|
| 583 |
+
"learning_rate": 3.76984126984127e-05,
|
| 584 |
+
"loss": 0.0385,
|
| 585 |
+
"step": 7200
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 8.005267857142858,
|
| 589 |
+
"grad_norm": 0.06927396357059479,
|
| 590 |
+
"learning_rate": 3.7450396825396824e-05,
|
| 591 |
+
"loss": 0.0322,
|
| 592 |
+
"step": 7300
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 8.009732142857143,
|
| 596 |
+
"grad_norm": 0.0045885867439210415,
|
| 597 |
+
"learning_rate": 3.7202380952380956e-05,
|
| 598 |
+
"loss": 0.0051,
|
| 599 |
+
"step": 7400
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 8.014196428571429,
|
| 603 |
+
"grad_norm": 0.005404896102845669,
|
| 604 |
+
"learning_rate": 3.695436507936508e-05,
|
| 605 |
+
"loss": 0.003,
|
| 606 |
+
"step": 7500
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 8.018660714285714,
|
| 610 |
+
"grad_norm": 0.003892578650265932,
|
| 611 |
+
"learning_rate": 3.6706349206349205e-05,
|
| 612 |
+
"loss": 0.0154,
|
| 613 |
+
"step": 7600
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 8.023125,
|
| 617 |
+
"grad_norm": 0.00285942736081779,
|
| 618 |
+
"learning_rate": 3.6458333333333336e-05,
|
| 619 |
+
"loss": 0.0095,
|
| 620 |
+
"step": 7700
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 8.027589285714285,
|
| 624 |
+
"grad_norm": 0.004967841785401106,
|
| 625 |
+
"learning_rate": 3.621031746031746e-05,
|
| 626 |
+
"loss": 0.0036,
|
| 627 |
+
"step": 7800
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 8.032053571428571,
|
| 631 |
+
"grad_norm": 0.0037566416431218386,
|
| 632 |
+
"learning_rate": 3.5962301587301586e-05,
|
| 633 |
+
"loss": 0.0013,
|
| 634 |
+
"step": 7900
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 8.036517857142858,
|
| 638 |
+
"grad_norm": 0.003524980042129755,
|
| 639 |
+
"learning_rate": 3.571428571428572e-05,
|
| 640 |
+
"loss": 0.0039,
|
| 641 |
+
"step": 8000
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 8.040044642857143,
|
| 645 |
+
"eval_accuracy": 0.991764705882353,
|
| 646 |
+
"eval_loss": 0.030685828998684883,
|
| 647 |
+
"eval_runtime": 277.8498,
|
| 648 |
+
"eval_samples_per_second": 3.059,
|
| 649 |
+
"eval_steps_per_second": 1.53,
|
| 650 |
+
"step": 8079
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 9.000904017857144,
|
| 654 |
+
"grad_norm": 0.009027580730617046,
|
| 655 |
+
"learning_rate": 3.546626984126984e-05,
|
| 656 |
+
"loss": 0.0276,
|
| 657 |
+
"step": 8100
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 9.005368303571428,
|
| 661 |
+
"grad_norm": 0.00775180896744132,
|
| 662 |
+
"learning_rate": 3.521825396825397e-05,
|
| 663 |
+
"loss": 0.0324,
|
| 664 |
+
"step": 8200
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 9.009832589285715,
|
| 668 |
+
"grad_norm": 0.003224916523322463,
|
| 669 |
+
"learning_rate": 3.49702380952381e-05,
|
| 670 |
+
"loss": 0.001,
|
| 671 |
+
"step": 8300
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 9.014296875,
|
| 675 |
+
"grad_norm": 0.001547032268717885,
|
| 676 |
+
"learning_rate": 3.472222222222222e-05,
|
| 677 |
+
"loss": 0.0002,
|
| 678 |
+
"step": 8400
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 9.018761160714286,
|
| 682 |
+
"grad_norm": 0.00670098839327693,
|
| 683 |
+
"learning_rate": 3.4474206349206354e-05,
|
| 684 |
+
"loss": 0.0254,
|
| 685 |
+
"step": 8500
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 9.02322544642857,
|
| 689 |
+
"grad_norm": 0.0018392472993582487,
|
| 690 |
+
"learning_rate": 3.422619047619048e-05,
|
| 691 |
+
"loss": 0.0341,
|
| 692 |
+
"step": 8600
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 9.027689732142857,
|
| 696 |
+
"grad_norm": 27.046947479248047,
|
| 697 |
+
"learning_rate": 3.397817460317461e-05,
|
| 698 |
+
"loss": 0.0559,
|
| 699 |
+
"step": 8700
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 9.032154017857144,
|
| 703 |
+
"grad_norm": 0.03754453361034393,
|
| 704 |
+
"learning_rate": 3.3730158730158734e-05,
|
| 705 |
+
"loss": 0.0343,
|
| 706 |
+
"step": 8800
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 9.036618303571428,
|
| 710 |
+
"grad_norm": 0.00891073513776064,
|
| 711 |
+
"learning_rate": 3.348214285714286e-05,
|
| 712 |
+
"loss": 0.0149,
|
| 713 |
+
"step": 8900
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 9.040055803571429,
|
| 717 |
+
"eval_accuracy": 0.9858823529411764,
|
| 718 |
+
"eval_loss": 0.06944379210472107,
|
| 719 |
+
"eval_runtime": 283.081,
|
| 720 |
+
"eval_samples_per_second": 3.003,
|
| 721 |
+
"eval_steps_per_second": 1.501,
|
| 722 |
+
"step": 8977
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 9.040055803571429,
|
| 726 |
+
"step": 8977,
|
| 727 |
+
"total_flos": 8.952618183954924e+19,
|
| 728 |
+
"train_loss": 0.48967243622379686,
|
| 729 |
+
"train_runtime": 34397.09,
|
| 730 |
+
"train_samples_per_second": 5.21,
|
| 731 |
+
"train_steps_per_second": 0.651
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 9.040055803571429,
|
| 735 |
+
"eval_accuracy": 0.9941176470588236,
|
| 736 |
+
"eval_loss": 0.03231202811002731,
|
| 737 |
+
"eval_runtime": 299.9449,
|
| 738 |
+
"eval_samples_per_second": 2.834,
|
| 739 |
+
"eval_steps_per_second": 1.417,
|
| 740 |
+
"step": 8977
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 9.040055803571429,
|
| 744 |
+
"eval_accuracy": 0.8079937304075235,
|
| 745 |
+
"eval_loss": 0.7567912340164185,
|
| 746 |
+
"eval_runtime": 438.2415,
|
| 747 |
+
"eval_samples_per_second": 2.912,
|
| 748 |
+
"eval_steps_per_second": 1.456,
|
| 749 |
+
"step": 8977
|
| 750 |
+
}
|
| 751 |
+
],
|
| 752 |
+
"logging_steps": 100,
|
| 753 |
+
"max_steps": 22400,
|
| 754 |
+
"num_input_tokens_seen": 0,
|
| 755 |
+
"num_train_epochs": 9223372036854775807,
|
| 756 |
+
"save_steps": 500,
|
| 757 |
+
"stateful_callbacks": {
|
| 758 |
+
"EarlyStoppingCallback": {
|
| 759 |
+
"args": {
|
| 760 |
+
"early_stopping_patience": 5,
|
| 761 |
+
"early_stopping_threshold": 0.0
|
| 762 |
+
},
|
| 763 |
+
"attributes": {
|
| 764 |
+
"early_stopping_patience_counter": 0
|
| 765 |
+
}
|
| 766 |
+
},
|
| 767 |
+
"TrainerControl": {
|
| 768 |
+
"args": {
|
| 769 |
+
"should_epoch_stop": false,
|
| 770 |
+
"should_evaluate": false,
|
| 771 |
+
"should_log": false,
|
| 772 |
+
"should_save": true,
|
| 773 |
+
"should_training_stop": true
|
| 774 |
+
},
|
| 775 |
+
"attributes": {}
|
| 776 |
+
}
|
| 777 |
+
},
|
| 778 |
+
"total_flos": 8.952618183954924e+19,
|
| 779 |
+
"train_batch_size": 2,
|
| 780 |
+
"trial_name": null,
|
| 781 |
+
"trial_params": null
|
| 782 |
+
}
|