End of training
Browse files- all_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +862 -0
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9259259259259259,
|
| 3 |
+
"total_flos": 1.9112276736e+17,
|
| 4 |
+
"train_loss": 1.9073539078235626,
|
| 5 |
+
"train_runtime": 260.6876,
|
| 6 |
+
"train_samples_per_second": 1.534,
|
| 7 |
+
"train_steps_per_second": 0.384
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 0.9259259259259259,
|
| 3 |
+
"total_flos": 1.9112276736e+17,
|
| 4 |
+
"train_loss": 1.9073539078235626,
|
| 5 |
+
"train_runtime": 260.6876,
|
| 6 |
+
"train_samples_per_second": 1.534,
|
| 7 |
+
"train_steps_per_second": 0.384
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,862 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9259259259259259,
|
| 5 |
+
"eval_steps": 20,
|
| 6 |
+
"global_step": 100,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.009259259259259259,
|
| 13 |
+
"grad_norm": NaN,
|
| 14 |
+
"learning_rate": 1e-05,
|
| 15 |
+
"loss": 1.9801,
|
| 16 |
+
"step": 1
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.018518518518518517,
|
| 20 |
+
"grad_norm": NaN,
|
| 21 |
+
"learning_rate": 1e-05,
|
| 22 |
+
"loss": 1.7876,
|
| 23 |
+
"step": 2
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.027777777777777776,
|
| 27 |
+
"grad_norm": Infinity,
|
| 28 |
+
"learning_rate": 1e-05,
|
| 29 |
+
"loss": 2.3855,
|
| 30 |
+
"step": 3
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.037037037037037035,
|
| 34 |
+
"grad_norm": 129.57530212402344,
|
| 35 |
+
"learning_rate": 9.9e-06,
|
| 36 |
+
"loss": 1.8948,
|
| 37 |
+
"step": 4
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.046296296296296294,
|
| 41 |
+
"grad_norm": Infinity,
|
| 42 |
+
"learning_rate": 9.9e-06,
|
| 43 |
+
"loss": 2.1217,
|
| 44 |
+
"step": 5
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05555555555555555,
|
| 48 |
+
"grad_norm": 96.91162872314453,
|
| 49 |
+
"learning_rate": 9.800000000000001e-06,
|
| 50 |
+
"loss": 2.0581,
|
| 51 |
+
"step": 6
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06481481481481481,
|
| 55 |
+
"grad_norm": 40.41334533691406,
|
| 56 |
+
"learning_rate": 9.7e-06,
|
| 57 |
+
"loss": 1.7127,
|
| 58 |
+
"step": 7
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.07407407407407407,
|
| 62 |
+
"grad_norm": 89.87907409667969,
|
| 63 |
+
"learning_rate": 9.600000000000001e-06,
|
| 64 |
+
"loss": 1.9292,
|
| 65 |
+
"step": 8
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.08333333333333333,
|
| 69 |
+
"grad_norm": 52.121402740478516,
|
| 70 |
+
"learning_rate": 9.5e-06,
|
| 71 |
+
"loss": 1.9921,
|
| 72 |
+
"step": 9
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.09259259259259259,
|
| 76 |
+
"grad_norm": 321.693115234375,
|
| 77 |
+
"learning_rate": 9.4e-06,
|
| 78 |
+
"loss": 2.1328,
|
| 79 |
+
"step": 10
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.10185185185185185,
|
| 83 |
+
"grad_norm": 125.0765609741211,
|
| 84 |
+
"learning_rate": 9.3e-06,
|
| 85 |
+
"loss": 1.986,
|
| 86 |
+
"step": 11
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1111111111111111,
|
| 90 |
+
"grad_norm": 110.41716766357422,
|
| 91 |
+
"learning_rate": 9.200000000000002e-06,
|
| 92 |
+
"loss": 1.8335,
|
| 93 |
+
"step": 12
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.12037037037037036,
|
| 97 |
+
"grad_norm": 77.33905792236328,
|
| 98 |
+
"learning_rate": 9.100000000000001e-06,
|
| 99 |
+
"loss": 1.9346,
|
| 100 |
+
"step": 13
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.12962962962962962,
|
| 104 |
+
"grad_norm": 89.59822845458984,
|
| 105 |
+
"learning_rate": 9e-06,
|
| 106 |
+
"loss": 2.0646,
|
| 107 |
+
"step": 14
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.1388888888888889,
|
| 111 |
+
"grad_norm": 23.786359786987305,
|
| 112 |
+
"learning_rate": 8.900000000000001e-06,
|
| 113 |
+
"loss": 1.9532,
|
| 114 |
+
"step": 15
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.14814814814814814,
|
| 118 |
+
"grad_norm": 49.40592956542969,
|
| 119 |
+
"learning_rate": 8.8e-06,
|
| 120 |
+
"loss": 1.9619,
|
| 121 |
+
"step": 16
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.1574074074074074,
|
| 125 |
+
"grad_norm": 146.65223693847656,
|
| 126 |
+
"learning_rate": 8.700000000000001e-06,
|
| 127 |
+
"loss": 1.9551,
|
| 128 |
+
"step": 17
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.16666666666666666,
|
| 132 |
+
"grad_norm": Infinity,
|
| 133 |
+
"learning_rate": 8.700000000000001e-06,
|
| 134 |
+
"loss": 1.6949,
|
| 135 |
+
"step": 18
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.17592592592592593,
|
| 139 |
+
"grad_norm": 236.78709411621094,
|
| 140 |
+
"learning_rate": 8.6e-06,
|
| 141 |
+
"loss": 1.8608,
|
| 142 |
+
"step": 19
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.18518518518518517,
|
| 146 |
+
"grad_norm": 184.51332092285156,
|
| 147 |
+
"learning_rate": 8.5e-06,
|
| 148 |
+
"loss": 1.9155,
|
| 149 |
+
"step": 20
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.18518518518518517,
|
| 153 |
+
"eval_loss": 1.8037505149841309,
|
| 154 |
+
"eval_map": 0.0047,
|
| 155 |
+
"eval_map_50": 0.0101,
|
| 156 |
+
"eval_map_75": 0.005,
|
| 157 |
+
"eval_map_Adult": 0.0081,
|
| 158 |
+
"eval_map_Kid": 0.0013,
|
| 159 |
+
"eval_map_large": 0.0071,
|
| 160 |
+
"eval_map_medium": 0.0035,
|
| 161 |
+
"eval_map_small": -1.0,
|
| 162 |
+
"eval_mar_1": 0.0332,
|
| 163 |
+
"eval_mar_10": 0.1231,
|
| 164 |
+
"eval_mar_100": 0.3206,
|
| 165 |
+
"eval_mar_100_Adult": 0.4727,
|
| 166 |
+
"eval_mar_100_Kid": 0.1685,
|
| 167 |
+
"eval_mar_large": 0.3067,
|
| 168 |
+
"eval_mar_medium": 0.4545,
|
| 169 |
+
"eval_mar_small": -1.0,
|
| 170 |
+
"eval_runtime": 15.3961,
|
| 171 |
+
"eval_samples_per_second": 4.936,
|
| 172 |
+
"eval_steps_per_second": 1.234,
|
| 173 |
+
"step": 20
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.19444444444444445,
|
| 177 |
+
"grad_norm": 163.67559814453125,
|
| 178 |
+
"learning_rate": 8.400000000000001e-06,
|
| 179 |
+
"loss": 2.1412,
|
| 180 |
+
"step": 21
|
| 181 |
+
},
|
| 182 |
+
{
|
| 183 |
+
"epoch": 0.2037037037037037,
|
| 184 |
+
"grad_norm": 24.548583984375,
|
| 185 |
+
"learning_rate": 8.3e-06,
|
| 186 |
+
"loss": 1.7217,
|
| 187 |
+
"step": 22
|
| 188 |
+
},
|
| 189 |
+
{
|
| 190 |
+
"epoch": 0.21296296296296297,
|
| 191 |
+
"grad_norm": 124.01173400878906,
|
| 192 |
+
"learning_rate": 8.2e-06,
|
| 193 |
+
"loss": 2.0238,
|
| 194 |
+
"step": 23
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.2222222222222222,
|
| 198 |
+
"grad_norm": 29.773996353149414,
|
| 199 |
+
"learning_rate": 8.1e-06,
|
| 200 |
+
"loss": 1.995,
|
| 201 |
+
"step": 24
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.23148148148148148,
|
| 205 |
+
"grad_norm": 294.4190368652344,
|
| 206 |
+
"learning_rate": 8.000000000000001e-06,
|
| 207 |
+
"loss": 1.7983,
|
| 208 |
+
"step": 25
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 0.24074074074074073,
|
| 212 |
+
"grad_norm": 478.20721435546875,
|
| 213 |
+
"learning_rate": 7.9e-06,
|
| 214 |
+
"loss": 2.1716,
|
| 215 |
+
"step": 26
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 0.25,
|
| 219 |
+
"grad_norm": 57.03529357910156,
|
| 220 |
+
"learning_rate": 7.800000000000002e-06,
|
| 221 |
+
"loss": 1.9222,
|
| 222 |
+
"step": 27
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 0.25925925925925924,
|
| 226 |
+
"grad_norm": 83.11935424804688,
|
| 227 |
+
"learning_rate": 7.7e-06,
|
| 228 |
+
"loss": 2.0409,
|
| 229 |
+
"step": 28
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 0.26851851851851855,
|
| 233 |
+
"grad_norm": 177.90359497070312,
|
| 234 |
+
"learning_rate": 7.600000000000001e-06,
|
| 235 |
+
"loss": 1.8105,
|
| 236 |
+
"step": 29
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.2777777777777778,
|
| 240 |
+
"grad_norm": 93.07716369628906,
|
| 241 |
+
"learning_rate": 7.500000000000001e-06,
|
| 242 |
+
"loss": 1.935,
|
| 243 |
+
"step": 30
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.28703703703703703,
|
| 247 |
+
"grad_norm": 146.4728546142578,
|
| 248 |
+
"learning_rate": 7.4e-06,
|
| 249 |
+
"loss": 1.6811,
|
| 250 |
+
"step": 31
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 0.2962962962962963,
|
| 254 |
+
"grad_norm": 223.94830322265625,
|
| 255 |
+
"learning_rate": 7.3e-06,
|
| 256 |
+
"loss": 1.922,
|
| 257 |
+
"step": 32
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.3055555555555556,
|
| 261 |
+
"grad_norm": 153.93211364746094,
|
| 262 |
+
"learning_rate": 7.2000000000000005e-06,
|
| 263 |
+
"loss": 1.9116,
|
| 264 |
+
"step": 33
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.3148148148148148,
|
| 268 |
+
"grad_norm": 193.51223754882812,
|
| 269 |
+
"learning_rate": 7.100000000000001e-06,
|
| 270 |
+
"loss": 2.1864,
|
| 271 |
+
"step": 34
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 0.32407407407407407,
|
| 275 |
+
"grad_norm": 590.421875,
|
| 276 |
+
"learning_rate": 7e-06,
|
| 277 |
+
"loss": 1.8323,
|
| 278 |
+
"step": 35
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.3333333333333333,
|
| 282 |
+
"grad_norm": 194.01913452148438,
|
| 283 |
+
"learning_rate": 6.9e-06,
|
| 284 |
+
"loss": 1.8585,
|
| 285 |
+
"step": 36
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.3425925925925926,
|
| 289 |
+
"grad_norm": 45.26372146606445,
|
| 290 |
+
"learning_rate": 6.800000000000001e-06,
|
| 291 |
+
"loss": 2.0296,
|
| 292 |
+
"step": 37
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 0.35185185185185186,
|
| 296 |
+
"grad_norm": 206.16380310058594,
|
| 297 |
+
"learning_rate": 6.700000000000001e-06,
|
| 298 |
+
"loss": 1.8509,
|
| 299 |
+
"step": 38
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 0.3611111111111111,
|
| 303 |
+
"grad_norm": 162.83180236816406,
|
| 304 |
+
"learning_rate": 6.600000000000001e-06,
|
| 305 |
+
"loss": 1.7652,
|
| 306 |
+
"step": 39
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 0.37037037037037035,
|
| 310 |
+
"grad_norm": 55.78132629394531,
|
| 311 |
+
"learning_rate": 6.5000000000000004e-06,
|
| 312 |
+
"loss": 1.8395,
|
| 313 |
+
"step": 40
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 0.37037037037037035,
|
| 317 |
+
"eval_loss": 1.7978708744049072,
|
| 318 |
+
"eval_map": 0.0045,
|
| 319 |
+
"eval_map_50": 0.0107,
|
| 320 |
+
"eval_map_75": 0.0035,
|
| 321 |
+
"eval_map_Adult": 0.0076,
|
| 322 |
+
"eval_map_Kid": 0.0014,
|
| 323 |
+
"eval_map_large": 0.0072,
|
| 324 |
+
"eval_map_medium": 0.0042,
|
| 325 |
+
"eval_map_small": -1.0,
|
| 326 |
+
"eval_mar_1": 0.0223,
|
| 327 |
+
"eval_mar_10": 0.1147,
|
| 328 |
+
"eval_mar_100": 0.3081,
|
| 329 |
+
"eval_mar_100_Adult": 0.4436,
|
| 330 |
+
"eval_mar_100_Kid": 0.1726,
|
| 331 |
+
"eval_mar_large": 0.2926,
|
| 332 |
+
"eval_mar_medium": 0.4909,
|
| 333 |
+
"eval_mar_small": -1.0,
|
| 334 |
+
"eval_runtime": 14.768,
|
| 335 |
+
"eval_samples_per_second": 5.146,
|
| 336 |
+
"eval_steps_per_second": 1.287,
|
| 337 |
+
"step": 40
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 0.37962962962962965,
|
| 341 |
+
"grad_norm": 60.19484329223633,
|
| 342 |
+
"learning_rate": 6.4000000000000006e-06,
|
| 343 |
+
"loss": 1.6552,
|
| 344 |
+
"step": 41
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.3888888888888889,
|
| 348 |
+
"grad_norm": 179.31629943847656,
|
| 349 |
+
"learning_rate": 6.300000000000001e-06,
|
| 350 |
+
"loss": 1.7017,
|
| 351 |
+
"step": 42
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.39814814814814814,
|
| 355 |
+
"grad_norm": 296.96734619140625,
|
| 356 |
+
"learning_rate": 6.200000000000001e-06,
|
| 357 |
+
"loss": 1.9784,
|
| 358 |
+
"step": 43
|
| 359 |
+
},
|
| 360 |
+
{
|
| 361 |
+
"epoch": 0.4074074074074074,
|
| 362 |
+
"grad_norm": 32.73106384277344,
|
| 363 |
+
"learning_rate": 6.1e-06,
|
| 364 |
+
"loss": 2.0133,
|
| 365 |
+
"step": 44
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 0.4166666666666667,
|
| 369 |
+
"grad_norm": 76.0937271118164,
|
| 370 |
+
"learning_rate": 6e-06,
|
| 371 |
+
"loss": 2.1171,
|
| 372 |
+
"step": 45
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 0.42592592592592593,
|
| 376 |
+
"grad_norm": 41.131385803222656,
|
| 377 |
+
"learning_rate": 5.9e-06,
|
| 378 |
+
"loss": 2.0797,
|
| 379 |
+
"step": 46
|
| 380 |
+
},
|
| 381 |
+
{
|
| 382 |
+
"epoch": 0.4351851851851852,
|
| 383 |
+
"grad_norm": 722.7262573242188,
|
| 384 |
+
"learning_rate": 5.8e-06,
|
| 385 |
+
"loss": 2.6376,
|
| 386 |
+
"step": 47
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.4444444444444444,
|
| 390 |
+
"grad_norm": 661.7726440429688,
|
| 391 |
+
"learning_rate": 5.7e-06,
|
| 392 |
+
"loss": 2.1092,
|
| 393 |
+
"step": 48
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.4537037037037037,
|
| 397 |
+
"grad_norm": 86.95410919189453,
|
| 398 |
+
"learning_rate": 5.600000000000001e-06,
|
| 399 |
+
"loss": 1.8785,
|
| 400 |
+
"step": 49
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 0.46296296296296297,
|
| 404 |
+
"grad_norm": 168.14027404785156,
|
| 405 |
+
"learning_rate": 5.500000000000001e-06,
|
| 406 |
+
"loss": 1.9761,
|
| 407 |
+
"step": 50
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 0.4722222222222222,
|
| 411 |
+
"grad_norm": Infinity,
|
| 412 |
+
"learning_rate": 5.500000000000001e-06,
|
| 413 |
+
"loss": 1.742,
|
| 414 |
+
"step": 51
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.48148148148148145,
|
| 418 |
+
"grad_norm": 25.475540161132812,
|
| 419 |
+
"learning_rate": 5.400000000000001e-06,
|
| 420 |
+
"loss": 1.922,
|
| 421 |
+
"step": 52
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 0.49074074074074076,
|
| 425 |
+
"grad_norm": 54.6159553527832,
|
| 426 |
+
"learning_rate": 5.300000000000001e-06,
|
| 427 |
+
"loss": 1.649,
|
| 428 |
+
"step": 53
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.5,
|
| 432 |
+
"grad_norm": 67.34464263916016,
|
| 433 |
+
"learning_rate": 5.2e-06,
|
| 434 |
+
"loss": 2.0628,
|
| 435 |
+
"step": 54
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.5092592592592593,
|
| 439 |
+
"grad_norm": 99.4066162109375,
|
| 440 |
+
"learning_rate": 5.1e-06,
|
| 441 |
+
"loss": 1.9407,
|
| 442 |
+
"step": 55
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 0.5185185185185185,
|
| 446 |
+
"grad_norm": 68.25653076171875,
|
| 447 |
+
"learning_rate": 5e-06,
|
| 448 |
+
"loss": 1.9328,
|
| 449 |
+
"step": 56
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 0.5277777777777778,
|
| 453 |
+
"grad_norm": 398.8371276855469,
|
| 454 |
+
"learning_rate": 4.9000000000000005e-06,
|
| 455 |
+
"loss": 1.8278,
|
| 456 |
+
"step": 57
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 0.5370370370370371,
|
| 460 |
+
"grad_norm": 49.0096435546875,
|
| 461 |
+
"learning_rate": 4.800000000000001e-06,
|
| 462 |
+
"loss": 1.9959,
|
| 463 |
+
"step": 58
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 0.5462962962962963,
|
| 467 |
+
"grad_norm": 310.508544921875,
|
| 468 |
+
"learning_rate": 4.7e-06,
|
| 469 |
+
"loss": 1.6916,
|
| 470 |
+
"step": 59
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.5555555555555556,
|
| 474 |
+
"grad_norm": 260.59576416015625,
|
| 475 |
+
"learning_rate": 4.600000000000001e-06,
|
| 476 |
+
"loss": 1.8477,
|
| 477 |
+
"step": 60
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.5555555555555556,
|
| 481 |
+
"eval_loss": 1.765649437904358,
|
| 482 |
+
"eval_map": 0.0046,
|
| 483 |
+
"eval_map_50": 0.0107,
|
| 484 |
+
"eval_map_75": 0.0041,
|
| 485 |
+
"eval_map_Adult": 0.0072,
|
| 486 |
+
"eval_map_Kid": 0.002,
|
| 487 |
+
"eval_map_large": 0.0077,
|
| 488 |
+
"eval_map_medium": 0.0044,
|
| 489 |
+
"eval_map_small": -1.0,
|
| 490 |
+
"eval_mar_1": 0.0329,
|
| 491 |
+
"eval_mar_10": 0.1148,
|
| 492 |
+
"eval_mar_100": 0.3276,
|
| 493 |
+
"eval_mar_100_Adult": 0.4309,
|
| 494 |
+
"eval_mar_100_Kid": 0.2242,
|
| 495 |
+
"eval_mar_large": 0.3137,
|
| 496 |
+
"eval_mar_medium": 0.5091,
|
| 497 |
+
"eval_mar_small": -1.0,
|
| 498 |
+
"eval_runtime": 14.8896,
|
| 499 |
+
"eval_samples_per_second": 5.104,
|
| 500 |
+
"eval_steps_per_second": 1.276,
|
| 501 |
+
"step": 60
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.5648148148148148,
|
| 505 |
+
"grad_norm": 65.56439971923828,
|
| 506 |
+
"learning_rate": 4.5e-06,
|
| 507 |
+
"loss": 1.9681,
|
| 508 |
+
"step": 61
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 0.5740740740740741,
|
| 512 |
+
"grad_norm": 99.23474884033203,
|
| 513 |
+
"learning_rate": 4.4e-06,
|
| 514 |
+
"loss": 1.7314,
|
| 515 |
+
"step": 62
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.5833333333333334,
|
| 519 |
+
"grad_norm": 118.4346694946289,
|
| 520 |
+
"learning_rate": 4.3e-06,
|
| 521 |
+
"loss": 1.6602,
|
| 522 |
+
"step": 63
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 0.5925925925925926,
|
| 526 |
+
"grad_norm": 43.443115234375,
|
| 527 |
+
"learning_rate": 4.2000000000000004e-06,
|
| 528 |
+
"loss": 2.0101,
|
| 529 |
+
"step": 64
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 0.6018518518518519,
|
| 533 |
+
"grad_norm": 112.58702850341797,
|
| 534 |
+
"learning_rate": 4.1e-06,
|
| 535 |
+
"loss": 1.656,
|
| 536 |
+
"step": 65
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.6111111111111112,
|
| 540 |
+
"grad_norm": 60.914520263671875,
|
| 541 |
+
"learning_rate": 4.000000000000001e-06,
|
| 542 |
+
"loss": 1.7922,
|
| 543 |
+
"step": 66
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.6203703703703703,
|
| 547 |
+
"grad_norm": 92.44080352783203,
|
| 548 |
+
"learning_rate": 3.900000000000001e-06,
|
| 549 |
+
"loss": 1.5629,
|
| 550 |
+
"step": 67
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 0.6296296296296297,
|
| 554 |
+
"grad_norm": 1000.2711181640625,
|
| 555 |
+
"learning_rate": 3.8000000000000005e-06,
|
| 556 |
+
"loss": 2.3308,
|
| 557 |
+
"step": 68
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.6388888888888888,
|
| 561 |
+
"grad_norm": 132.41334533691406,
|
| 562 |
+
"learning_rate": 3.7e-06,
|
| 563 |
+
"loss": 1.7623,
|
| 564 |
+
"step": 69
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.6481481481481481,
|
| 568 |
+
"grad_norm": 631.0505981445312,
|
| 569 |
+
"learning_rate": 3.6000000000000003e-06,
|
| 570 |
+
"loss": 2.043,
|
| 571 |
+
"step": 70
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 0.6574074074074074,
|
| 575 |
+
"grad_norm": 230.10476684570312,
|
| 576 |
+
"learning_rate": 3.5e-06,
|
| 577 |
+
"loss": 1.9894,
|
| 578 |
+
"step": 71
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.6666666666666666,
|
| 582 |
+
"grad_norm": 68.68940734863281,
|
| 583 |
+
"learning_rate": 3.4000000000000005e-06,
|
| 584 |
+
"loss": 1.8069,
|
| 585 |
+
"step": 72
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.6759259259259259,
|
| 589 |
+
"grad_norm": 310.83233642578125,
|
| 590 |
+
"learning_rate": 3.3000000000000006e-06,
|
| 591 |
+
"loss": 1.8285,
|
| 592 |
+
"step": 73
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 0.6851851851851852,
|
| 596 |
+
"grad_norm": 97.42311096191406,
|
| 597 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 598 |
+
"loss": 1.8649,
|
| 599 |
+
"step": 74
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.6944444444444444,
|
| 603 |
+
"grad_norm": 119.45216369628906,
|
| 604 |
+
"learning_rate": 3.1000000000000004e-06,
|
| 605 |
+
"loss": 1.6971,
|
| 606 |
+
"step": 75
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 0.7037037037037037,
|
| 610 |
+
"grad_norm": 63.03942108154297,
|
| 611 |
+
"learning_rate": 3e-06,
|
| 612 |
+
"loss": 1.8963,
|
| 613 |
+
"step": 76
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 0.7129629629629629,
|
| 617 |
+
"grad_norm": 69.41495513916016,
|
| 618 |
+
"learning_rate": 2.9e-06,
|
| 619 |
+
"loss": 1.551,
|
| 620 |
+
"step": 77
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.7222222222222222,
|
| 624 |
+
"grad_norm": 90.6502685546875,
|
| 625 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 626 |
+
"loss": 1.6615,
|
| 627 |
+
"step": 78
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.7314814814814815,
|
| 631 |
+
"grad_norm": 111.7878646850586,
|
| 632 |
+
"learning_rate": 2.7000000000000004e-06,
|
| 633 |
+
"loss": 2.3283,
|
| 634 |
+
"step": 79
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 0.7407407407407407,
|
| 638 |
+
"grad_norm": 89.25751495361328,
|
| 639 |
+
"learning_rate": 2.6e-06,
|
| 640 |
+
"loss": 1.66,
|
| 641 |
+
"step": 80
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 0.7407407407407407,
|
| 645 |
+
"eval_loss": 1.7439719438552856,
|
| 646 |
+
"eval_map": 0.0044,
|
| 647 |
+
"eval_map_50": 0.0098,
|
| 648 |
+
"eval_map_75": 0.0039,
|
| 649 |
+
"eval_map_Adult": 0.0062,
|
| 650 |
+
"eval_map_Kid": 0.0026,
|
| 651 |
+
"eval_map_large": 0.0076,
|
| 652 |
+
"eval_map_medium": 0.0049,
|
| 653 |
+
"eval_map_small": -1.0,
|
| 654 |
+
"eval_mar_1": 0.0256,
|
| 655 |
+
"eval_mar_10": 0.1029,
|
| 656 |
+
"eval_mar_100": 0.3266,
|
| 657 |
+
"eval_mar_100_Adult": 0.4,
|
| 658 |
+
"eval_mar_100_Kid": 0.2532,
|
| 659 |
+
"eval_mar_large": 0.3137,
|
| 660 |
+
"eval_mar_medium": 0.5182,
|
| 661 |
+
"eval_mar_small": -1.0,
|
| 662 |
+
"eval_runtime": 15.2125,
|
| 663 |
+
"eval_samples_per_second": 4.996,
|
| 664 |
+
"eval_steps_per_second": 1.249,
|
| 665 |
+
"step": 80
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 0.75,
|
| 669 |
+
"grad_norm": 87.07488250732422,
|
| 670 |
+
"learning_rate": 2.5e-06,
|
| 671 |
+
"loss": 1.7118,
|
| 672 |
+
"step": 81
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 0.7592592592592593,
|
| 676 |
+
"grad_norm": 59.08213424682617,
|
| 677 |
+
"learning_rate": 2.4000000000000003e-06,
|
| 678 |
+
"loss": 1.6906,
|
| 679 |
+
"step": 82
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 0.7685185185185185,
|
| 683 |
+
"grad_norm": 45.39196014404297,
|
| 684 |
+
"learning_rate": 2.3000000000000004e-06,
|
| 685 |
+
"loss": 1.7661,
|
| 686 |
+
"step": 83
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.7777777777777778,
|
| 690 |
+
"grad_norm": 42.13712692260742,
|
| 691 |
+
"learning_rate": 2.2e-06,
|
| 692 |
+
"loss": 2.0699,
|
| 693 |
+
"step": 84
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.7870370370370371,
|
| 697 |
+
"grad_norm": 355.4501647949219,
|
| 698 |
+
"learning_rate": 2.1000000000000002e-06,
|
| 699 |
+
"loss": 1.6448,
|
| 700 |
+
"step": 85
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 0.7962962962962963,
|
| 704 |
+
"grad_norm": 42.52521514892578,
|
| 705 |
+
"learning_rate": 2.0000000000000003e-06,
|
| 706 |
+
"loss": 1.8531,
|
| 707 |
+
"step": 86
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 0.8055555555555556,
|
| 711 |
+
"grad_norm": 90.40050506591797,
|
| 712 |
+
"learning_rate": 1.9000000000000002e-06,
|
| 713 |
+
"loss": 1.9209,
|
| 714 |
+
"step": 87
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.8148148148148148,
|
| 718 |
+
"grad_norm": Infinity,
|
| 719 |
+
"learning_rate": 1.9000000000000002e-06,
|
| 720 |
+
"loss": 2.0899,
|
| 721 |
+
"step": 88
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 0.8240740740740741,
|
| 725 |
+
"grad_norm": 58.782588958740234,
|
| 726 |
+
"learning_rate": 1.8000000000000001e-06,
|
| 727 |
+
"loss": 1.7296,
|
| 728 |
+
"step": 89
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.8333333333333334,
|
| 732 |
+
"grad_norm": 45.47020721435547,
|
| 733 |
+
"learning_rate": 1.7000000000000002e-06,
|
| 734 |
+
"loss": 1.6152,
|
| 735 |
+
"step": 90
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.8425925925925926,
|
| 739 |
+
"grad_norm": 25.518877029418945,
|
| 740 |
+
"learning_rate": 1.6000000000000001e-06,
|
| 741 |
+
"loss": 1.6979,
|
| 742 |
+
"step": 91
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 0.8518518518518519,
|
| 746 |
+
"grad_norm": 87.5730972290039,
|
| 747 |
+
"learning_rate": 1.5e-06,
|
| 748 |
+
"loss": 2.2574,
|
| 749 |
+
"step": 92
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 0.8611111111111112,
|
| 753 |
+
"grad_norm": 39.618709564208984,
|
| 754 |
+
"learning_rate": 1.4000000000000001e-06,
|
| 755 |
+
"loss": 2.0495,
|
| 756 |
+
"step": 93
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 0.8703703703703703,
|
| 760 |
+
"grad_norm": 206.44537353515625,
|
| 761 |
+
"learning_rate": 1.3e-06,
|
| 762 |
+
"loss": 2.2438,
|
| 763 |
+
"step": 94
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 0.8796296296296297,
|
| 767 |
+
"grad_norm": 163.28404235839844,
|
| 768 |
+
"learning_rate": 1.2000000000000002e-06,
|
| 769 |
+
"loss": 1.7708,
|
| 770 |
+
"step": 95
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.8888888888888888,
|
| 774 |
+
"grad_norm": 103.78318786621094,
|
| 775 |
+
"learning_rate": 1.1e-06,
|
| 776 |
+
"loss": 1.9995,
|
| 777 |
+
"step": 96
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.8981481481481481,
|
| 781 |
+
"grad_norm": 150.3301239013672,
|
| 782 |
+
"learning_rate": 1.0000000000000002e-06,
|
| 783 |
+
"loss": 1.8358,
|
| 784 |
+
"step": 97
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 0.9074074074074074,
|
| 788 |
+
"grad_norm": 1041.284423828125,
|
| 789 |
+
"learning_rate": 9.000000000000001e-07,
|
| 790 |
+
"loss": 1.6731,
|
| 791 |
+
"step": 98
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 0.9166666666666666,
|
| 795 |
+
"grad_norm": 290.3284912109375,
|
| 796 |
+
"learning_rate": 8.000000000000001e-07,
|
| 797 |
+
"loss": 1.9554,
|
| 798 |
+
"step": 99
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 0.9259259259259259,
|
| 802 |
+
"grad_norm": 25.19146728515625,
|
| 803 |
+
"learning_rate": 7.000000000000001e-07,
|
| 804 |
+
"loss": 2.0088,
|
| 805 |
+
"step": 100
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 0.9259259259259259,
|
| 809 |
+
"eval_loss": 1.7495548725128174,
|
| 810 |
+
"eval_map": 0.0043,
|
| 811 |
+
"eval_map_50": 0.0096,
|
| 812 |
+
"eval_map_75": 0.0039,
|
| 813 |
+
"eval_map_Adult": 0.0056,
|
| 814 |
+
"eval_map_Kid": 0.0029,
|
| 815 |
+
"eval_map_large": 0.0073,
|
| 816 |
+
"eval_map_medium": 0.0051,
|
| 817 |
+
"eval_map_small": -1.0,
|
| 818 |
+
"eval_mar_1": 0.0175,
|
| 819 |
+
"eval_mar_10": 0.0983,
|
| 820 |
+
"eval_mar_100": 0.3165,
|
| 821 |
+
"eval_mar_100_Adult": 0.3636,
|
| 822 |
+
"eval_mar_100_Kid": 0.2694,
|
| 823 |
+
"eval_mar_large": 0.3035,
|
| 824 |
+
"eval_mar_medium": 0.5364,
|
| 825 |
+
"eval_mar_small": -1.0,
|
| 826 |
+
"eval_runtime": 15.2134,
|
| 827 |
+
"eval_samples_per_second": 4.996,
|
| 828 |
+
"eval_steps_per_second": 1.249,
|
| 829 |
+
"step": 100
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.9259259259259259,
|
| 833 |
+
"step": 100,
|
| 834 |
+
"total_flos": 1.9112276736e+17,
|
| 835 |
+
"train_loss": 1.9073539078235626,
|
| 836 |
+
"train_runtime": 260.6876,
|
| 837 |
+
"train_samples_per_second": 1.534,
|
| 838 |
+
"train_steps_per_second": 0.384
|
| 839 |
+
}
|
| 840 |
+
],
|
| 841 |
+
"logging_steps": 1,
|
| 842 |
+
"max_steps": 100,
|
| 843 |
+
"num_input_tokens_seen": 0,
|
| 844 |
+
"num_train_epochs": 1,
|
| 845 |
+
"save_steps": 10,
|
| 846 |
+
"stateful_callbacks": {
|
| 847 |
+
"TrainerControl": {
|
| 848 |
+
"args": {
|
| 849 |
+
"should_epoch_stop": false,
|
| 850 |
+
"should_evaluate": false,
|
| 851 |
+
"should_log": false,
|
| 852 |
+
"should_save": true,
|
| 853 |
+
"should_training_stop": true
|
| 854 |
+
},
|
| 855 |
+
"attributes": {}
|
| 856 |
+
}
|
| 857 |
+
},
|
| 858 |
+
"total_flos": 1.9112276736e+17,
|
| 859 |
+
"train_batch_size": 4,
|
| 860 |
+
"trial_name": null,
|
| 861 |
+
"trial_params": null
|
| 862 |
+
}
|