Spatiallysaying
commited on
Commit
•
ec1f85b
1
Parent(s):
2d0bd9e
End of training
Browse files- README.md +3 -1
- all_results.json +33 -0
- eval_results.json +28 -0
- runs/Jul14_13-42-10_25399931ec8c/events.out.tfevents.1720970494.25399931ec8c.2180.1 +3 -0
- train_results.json +8 -0
- trainer_state.json +774 -0
README.md
CHANGED
@@ -2,6 +2,8 @@
|
|
2 |
license: other
|
3 |
base_model: nvidia/mit-b0
|
4 |
tags:
|
|
|
|
|
5 |
- generated_from_trainer
|
6 |
model-index:
|
7 |
- name: segformer-finetuned-rwymarkings-3k-steps
|
@@ -13,7 +15,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
13 |
|
14 |
# segformer-finetuned-rwymarkings-3k-steps
|
15 |
|
16 |
-
This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on
|
17 |
It achieves the following results on the evaluation set:
|
18 |
- Loss: 0.0182
|
19 |
- Mean Iou: 0.0441
|
|
|
2 |
license: other
|
3 |
base_model: nvidia/mit-b0
|
4 |
tags:
|
5 |
+
- image-segmentation
|
6 |
+
- vision
|
7 |
- generated_from_trainer
|
8 |
model-index:
|
9 |
- name: segformer-finetuned-rwymarkings-3k-steps
|
|
|
15 |
|
16 |
# segformer-finetuned-rwymarkings-3k-steps
|
17 |
|
18 |
+
This model is a fine-tuned version of [nvidia/mit-b0](https://huggingface.co/nvidia/mit-b0) on the Spatiallysaying/rwymarkings dataset.
|
19 |
It achieves the following results on the evaluation set:
|
20 |
- Loss: 0.0182
|
21 |
- Mean Iou: 0.0441
|
all_results.json
ADDED
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 17.341040462427745,
|
3 |
+
"eval_accuracy__backgound_": NaN,
|
4 |
+
"eval_accuracy_aim": 0.220281446734971,
|
5 |
+
"eval_accuracy_arrow": 0.0,
|
6 |
+
"eval_accuracy_chevron": 0.0,
|
7 |
+
"eval_accuracy_desig": 0.0,
|
8 |
+
"eval_accuracy_disp": 0.0,
|
9 |
+
"eval_accuracy_rwythr": 0.09711051534107834,
|
10 |
+
"eval_accuracy_tdz": 0.09081906093440521,
|
11 |
+
"eval_accuracy_thrbar": 0.0,
|
12 |
+
"eval_iou__backgound_": 0.0,
|
13 |
+
"eval_iou_aim": 0.2189167637858063,
|
14 |
+
"eval_iou_arrow": 0.0,
|
15 |
+
"eval_iou_chevron": 0.0,
|
16 |
+
"eval_iou_desig": 0.0,
|
17 |
+
"eval_iou_disp": 0.0,
|
18 |
+
"eval_iou_rwythr": 0.09582598471487361,
|
19 |
+
"eval_iou_tdz": 0.08179004249514715,
|
20 |
+
"eval_iou_thrbar": 0.0,
|
21 |
+
"eval_loss": 0.01818845421075821,
|
22 |
+
"eval_mean_accuracy": 0.05102637787630682,
|
23 |
+
"eval_mean_iou": 0.044059198999536336,
|
24 |
+
"eval_overall_accuracy": 0.08002687469672651,
|
25 |
+
"eval_runtime": 68.5586,
|
26 |
+
"eval_samples_per_second": 1.269,
|
27 |
+
"eval_steps_per_second": 0.642,
|
28 |
+
"total_flos": 1.05218665611264e+17,
|
29 |
+
"train_loss": 0.11730531366666158,
|
30 |
+
"train_runtime": 5867.3784,
|
31 |
+
"train_samples_per_second": 1.023,
|
32 |
+
"train_steps_per_second": 0.511
|
33 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 17.341040462427745,
|
3 |
+
"eval_accuracy__backgound_": NaN,
|
4 |
+
"eval_accuracy_aim": 0.220281446734971,
|
5 |
+
"eval_accuracy_arrow": 0.0,
|
6 |
+
"eval_accuracy_chevron": 0.0,
|
7 |
+
"eval_accuracy_desig": 0.0,
|
8 |
+
"eval_accuracy_disp": 0.0,
|
9 |
+
"eval_accuracy_rwythr": 0.09711051534107834,
|
10 |
+
"eval_accuracy_tdz": 0.09081906093440521,
|
11 |
+
"eval_accuracy_thrbar": 0.0,
|
12 |
+
"eval_iou__backgound_": 0.0,
|
13 |
+
"eval_iou_aim": 0.2189167637858063,
|
14 |
+
"eval_iou_arrow": 0.0,
|
15 |
+
"eval_iou_chevron": 0.0,
|
16 |
+
"eval_iou_desig": 0.0,
|
17 |
+
"eval_iou_disp": 0.0,
|
18 |
+
"eval_iou_rwythr": 0.09582598471487361,
|
19 |
+
"eval_iou_tdz": 0.08179004249514715,
|
20 |
+
"eval_iou_thrbar": 0.0,
|
21 |
+
"eval_loss": 0.01818845421075821,
|
22 |
+
"eval_mean_accuracy": 0.05102637787630682,
|
23 |
+
"eval_mean_iou": 0.044059198999536336,
|
24 |
+
"eval_overall_accuracy": 0.08002687469672651,
|
25 |
+
"eval_runtime": 68.5586,
|
26 |
+
"eval_samples_per_second": 1.269,
|
27 |
+
"eval_steps_per_second": 0.642
|
28 |
+
}
|
runs/Jul14_13-42-10_25399931ec8c/events.out.tfevents.1720970494.25399931ec8c.2180.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:648ecacbc3d0d58611d4876e18a6787452c2466ee3ab9af328d17ed418d30f5d
|
3 |
+
size 1537
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 17.341040462427745,
|
3 |
+
"total_flos": 1.05218665611264e+17,
|
4 |
+
"train_loss": 0.11730531366666158,
|
5 |
+
"train_runtime": 5867.3784,
|
6 |
+
"train_samples_per_second": 1.023,
|
7 |
+
"train_steps_per_second": 0.511
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,774 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 17.341040462427745,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 3000,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.5780346820809249,
|
13 |
+
"grad_norm": 5.030792713165283,
|
14 |
+
"learning_rate": 5.8003333333333334e-05,
|
15 |
+
"loss": 1.6294,
|
16 |
+
"step": 100
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 1.0,
|
20 |
+
"eval_accuracy__backgound_": NaN,
|
21 |
+
"eval_accuracy_aim": 0.0,
|
22 |
+
"eval_accuracy_arrow": 0.0,
|
23 |
+
"eval_accuracy_chevron": 0.0,
|
24 |
+
"eval_accuracy_desig": 0.0,
|
25 |
+
"eval_accuracy_disp": 0.0,
|
26 |
+
"eval_accuracy_rwythr": 0.0,
|
27 |
+
"eval_accuracy_tdz": 0.0,
|
28 |
+
"eval_accuracy_thrbar": 0.0,
|
29 |
+
"eval_iou__backgound_": 0.0,
|
30 |
+
"eval_iou_aim": 0.0,
|
31 |
+
"eval_iou_arrow": 0.0,
|
32 |
+
"eval_iou_chevron": 0.0,
|
33 |
+
"eval_iou_desig": 0.0,
|
34 |
+
"eval_iou_disp": 0.0,
|
35 |
+
"eval_iou_rwythr": 0.0,
|
36 |
+
"eval_iou_tdz": 0.0,
|
37 |
+
"eval_iou_thrbar": 0.0,
|
38 |
+
"eval_loss": 0.5447909235954285,
|
39 |
+
"eval_mean_accuracy": 0.0,
|
40 |
+
"eval_mean_iou": 0.0,
|
41 |
+
"eval_overall_accuracy": 0.0,
|
42 |
+
"eval_runtime": 69.2081,
|
43 |
+
"eval_samples_per_second": 1.257,
|
44 |
+
"eval_steps_per_second": 0.636,
|
45 |
+
"step": 173
|
46 |
+
},
|
47 |
+
{
|
48 |
+
"epoch": 1.1560693641618498,
|
49 |
+
"grad_norm": 3.1198508739471436,
|
50 |
+
"learning_rate": 5.600666666666667e-05,
|
51 |
+
"loss": 0.7662,
|
52 |
+
"step": 200
|
53 |
+
},
|
54 |
+
{
|
55 |
+
"epoch": 1.7341040462427746,
|
56 |
+
"grad_norm": 2.538226366043091,
|
57 |
+
"learning_rate": 5.401e-05,
|
58 |
+
"loss": 0.3371,
|
59 |
+
"step": 300
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 2.0,
|
63 |
+
"eval_accuracy__backgound_": NaN,
|
64 |
+
"eval_accuracy_aim": 0.0,
|
65 |
+
"eval_accuracy_arrow": 0.0,
|
66 |
+
"eval_accuracy_chevron": 0.0,
|
67 |
+
"eval_accuracy_desig": 0.0,
|
68 |
+
"eval_accuracy_disp": 0.0,
|
69 |
+
"eval_accuracy_rwythr": 0.0,
|
70 |
+
"eval_accuracy_tdz": 0.0,
|
71 |
+
"eval_accuracy_thrbar": 0.0,
|
72 |
+
"eval_iou__backgound_": 0.0,
|
73 |
+
"eval_iou_aim": 0.0,
|
74 |
+
"eval_iou_arrow": 0.0,
|
75 |
+
"eval_iou_chevron": 0.0,
|
76 |
+
"eval_iou_desig": 0.0,
|
77 |
+
"eval_iou_disp": 0.0,
|
78 |
+
"eval_iou_rwythr": 0.0,
|
79 |
+
"eval_iou_tdz": 0.0,
|
80 |
+
"eval_iou_thrbar": 0.0,
|
81 |
+
"eval_loss": 0.11072597652673721,
|
82 |
+
"eval_mean_accuracy": 0.0,
|
83 |
+
"eval_mean_iou": 0.0,
|
84 |
+
"eval_overall_accuracy": 0.0,
|
85 |
+
"eval_runtime": 69.363,
|
86 |
+
"eval_samples_per_second": 1.254,
|
87 |
+
"eval_steps_per_second": 0.634,
|
88 |
+
"step": 346
|
89 |
+
},
|
90 |
+
{
|
91 |
+
"epoch": 2.3121387283236996,
|
92 |
+
"grad_norm": 0.97201007604599,
|
93 |
+
"learning_rate": 5.201333333333334e-05,
|
94 |
+
"loss": 0.1452,
|
95 |
+
"step": 400
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 2.8901734104046244,
|
99 |
+
"grad_norm": 0.2807599604129791,
|
100 |
+
"learning_rate": 5.001666666666667e-05,
|
101 |
+
"loss": 0.0724,
|
102 |
+
"step": 500
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 3.0,
|
106 |
+
"eval_accuracy__backgound_": NaN,
|
107 |
+
"eval_accuracy_aim": 0.0,
|
108 |
+
"eval_accuracy_arrow": 0.0,
|
109 |
+
"eval_accuracy_chevron": 0.0,
|
110 |
+
"eval_accuracy_desig": 0.0,
|
111 |
+
"eval_accuracy_disp": 0.0,
|
112 |
+
"eval_accuracy_rwythr": 0.0,
|
113 |
+
"eval_accuracy_tdz": 0.0,
|
114 |
+
"eval_accuracy_thrbar": 0.0,
|
115 |
+
"eval_iou__backgound_": 0.0,
|
116 |
+
"eval_iou_aim": 0.0,
|
117 |
+
"eval_iou_arrow": 0.0,
|
118 |
+
"eval_iou_chevron": 0.0,
|
119 |
+
"eval_iou_desig": 0.0,
|
120 |
+
"eval_iou_disp": 0.0,
|
121 |
+
"eval_iou_rwythr": 0.0,
|
122 |
+
"eval_iou_tdz": 0.0,
|
123 |
+
"eval_iou_thrbar": 0.0,
|
124 |
+
"eval_loss": 0.04833525791764259,
|
125 |
+
"eval_mean_accuracy": 0.0,
|
126 |
+
"eval_mean_iou": 0.0,
|
127 |
+
"eval_overall_accuracy": 0.0,
|
128 |
+
"eval_runtime": 70.0111,
|
129 |
+
"eval_samples_per_second": 1.243,
|
130 |
+
"eval_steps_per_second": 0.628,
|
131 |
+
"step": 519
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 3.468208092485549,
|
135 |
+
"grad_norm": 0.4320279359817505,
|
136 |
+
"learning_rate": 4.8020000000000004e-05,
|
137 |
+
"loss": 0.0508,
|
138 |
+
"step": 600
|
139 |
+
},
|
140 |
+
{
|
141 |
+
"epoch": 4.0,
|
142 |
+
"eval_accuracy__backgound_": NaN,
|
143 |
+
"eval_accuracy_aim": 0.0,
|
144 |
+
"eval_accuracy_arrow": 0.0,
|
145 |
+
"eval_accuracy_chevron": 0.0,
|
146 |
+
"eval_accuracy_desig": 0.0,
|
147 |
+
"eval_accuracy_disp": 0.0,
|
148 |
+
"eval_accuracy_rwythr": 0.0,
|
149 |
+
"eval_accuracy_tdz": 0.0,
|
150 |
+
"eval_accuracy_thrbar": 0.0,
|
151 |
+
"eval_iou__backgound_": 0.0,
|
152 |
+
"eval_iou_aim": 0.0,
|
153 |
+
"eval_iou_arrow": 0.0,
|
154 |
+
"eval_iou_chevron": 0.0,
|
155 |
+
"eval_iou_desig": 0.0,
|
156 |
+
"eval_iou_disp": 0.0,
|
157 |
+
"eval_iou_rwythr": 0.0,
|
158 |
+
"eval_iou_tdz": 0.0,
|
159 |
+
"eval_iou_thrbar": 0.0,
|
160 |
+
"eval_loss": 0.0330786257982254,
|
161 |
+
"eval_mean_accuracy": 0.0,
|
162 |
+
"eval_mean_iou": 0.0,
|
163 |
+
"eval_overall_accuracy": 0.0,
|
164 |
+
"eval_runtime": 68.3376,
|
165 |
+
"eval_samples_per_second": 1.273,
|
166 |
+
"eval_steps_per_second": 0.644,
|
167 |
+
"step": 692
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 4.046242774566474,
|
171 |
+
"grad_norm": 0.4302203953266144,
|
172 |
+
"learning_rate": 4.602333333333333e-05,
|
173 |
+
"loss": 0.0417,
|
174 |
+
"step": 700
|
175 |
+
},
|
176 |
+
{
|
177 |
+
"epoch": 4.624277456647399,
|
178 |
+
"grad_norm": 0.31269383430480957,
|
179 |
+
"learning_rate": 4.402666666666667e-05,
|
180 |
+
"loss": 0.0369,
|
181 |
+
"step": 800
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 5.0,
|
185 |
+
"eval_accuracy__backgound_": NaN,
|
186 |
+
"eval_accuracy_aim": 0.0018516232563881001,
|
187 |
+
"eval_accuracy_arrow": 0.0,
|
188 |
+
"eval_accuracy_chevron": 0.0,
|
189 |
+
"eval_accuracy_desig": 0.0,
|
190 |
+
"eval_accuracy_disp": 0.0,
|
191 |
+
"eval_accuracy_rwythr": 0.0,
|
192 |
+
"eval_accuracy_tdz": 0.0,
|
193 |
+
"eval_accuracy_thrbar": 0.0,
|
194 |
+
"eval_iou__backgound_": 0.0,
|
195 |
+
"eval_iou_aim": 0.0018516232563881001,
|
196 |
+
"eval_iou_arrow": 0.0,
|
197 |
+
"eval_iou_chevron": 0.0,
|
198 |
+
"eval_iou_desig": 0.0,
|
199 |
+
"eval_iou_disp": 0.0,
|
200 |
+
"eval_iou_rwythr": 0.0,
|
201 |
+
"eval_iou_tdz": 0.0,
|
202 |
+
"eval_iou_thrbar": 0.0,
|
203 |
+
"eval_loss": 0.028909126296639442,
|
204 |
+
"eval_mean_accuracy": 0.00023145290704851252,
|
205 |
+
"eval_mean_iou": 0.00020573591737645558,
|
206 |
+
"eval_overall_accuracy": 0.00037325967675711993,
|
207 |
+
"eval_runtime": 71.0072,
|
208 |
+
"eval_samples_per_second": 1.225,
|
209 |
+
"eval_steps_per_second": 0.62,
|
210 |
+
"step": 865
|
211 |
+
},
|
212 |
+
{
|
213 |
+
"epoch": 5.202312138728324,
|
214 |
+
"grad_norm": 0.2979387044906616,
|
215 |
+
"learning_rate": 4.203e-05,
|
216 |
+
"loss": 0.0306,
|
217 |
+
"step": 900
|
218 |
+
},
|
219 |
+
{
|
220 |
+
"epoch": 5.780346820809249,
|
221 |
+
"grad_norm": 0.17238232493400574,
|
222 |
+
"learning_rate": 4.003333333333334e-05,
|
223 |
+
"loss": 0.0272,
|
224 |
+
"step": 1000
|
225 |
+
},
|
226 |
+
{
|
227 |
+
"epoch": 6.0,
|
228 |
+
"eval_accuracy__backgound_": NaN,
|
229 |
+
"eval_accuracy_aim": 0.08529811134427849,
|
230 |
+
"eval_accuracy_arrow": 0.0,
|
231 |
+
"eval_accuracy_chevron": 0.0,
|
232 |
+
"eval_accuracy_desig": 0.0,
|
233 |
+
"eval_accuracy_disp": 0.0,
|
234 |
+
"eval_accuracy_rwythr": 0.0,
|
235 |
+
"eval_accuracy_tdz": 0.010660608178958406,
|
236 |
+
"eval_accuracy_thrbar": 0.0,
|
237 |
+
"eval_iou__backgound_": 0.0,
|
238 |
+
"eval_iou_aim": 0.08447949141145547,
|
239 |
+
"eval_iou_arrow": 0.0,
|
240 |
+
"eval_iou_chevron": 0.0,
|
241 |
+
"eval_iou_desig": 0.0,
|
242 |
+
"eval_iou_disp": 0.0,
|
243 |
+
"eval_iou_rwythr": 0.0,
|
244 |
+
"eval_iou_tdz": 0.01050034427358274,
|
245 |
+
"eval_iou_thrbar": 0.0,
|
246 |
+
"eval_loss": 0.027622828260064125,
|
247 |
+
"eval_mean_accuracy": 0.011994839940404612,
|
248 |
+
"eval_mean_iou": 0.010553315076115356,
|
249 |
+
"eval_overall_accuracy": 0.019471713137496422,
|
250 |
+
"eval_runtime": 70.32,
|
251 |
+
"eval_samples_per_second": 1.237,
|
252 |
+
"eval_steps_per_second": 0.626,
|
253 |
+
"step": 1038
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 6.358381502890174,
|
257 |
+
"grad_norm": 0.1439923644065857,
|
258 |
+
"learning_rate": 3.803666666666667e-05,
|
259 |
+
"loss": 0.0265,
|
260 |
+
"step": 1100
|
261 |
+
},
|
262 |
+
{
|
263 |
+
"epoch": 6.936416184971098,
|
264 |
+
"grad_norm": 0.13196904957294464,
|
265 |
+
"learning_rate": 3.604e-05,
|
266 |
+
"loss": 0.0258,
|
267 |
+
"step": 1200
|
268 |
+
},
|
269 |
+
{
|
270 |
+
"epoch": 7.0,
|
271 |
+
"eval_accuracy__backgound_": NaN,
|
272 |
+
"eval_accuracy_aim": 0.048018763115664734,
|
273 |
+
"eval_accuracy_arrow": 0.0,
|
274 |
+
"eval_accuracy_chevron": 0.0,
|
275 |
+
"eval_accuracy_desig": 0.0,
|
276 |
+
"eval_accuracy_disp": 0.0,
|
277 |
+
"eval_accuracy_rwythr": 0.0,
|
278 |
+
"eval_accuracy_tdz": 0.011767447279506,
|
279 |
+
"eval_accuracy_thrbar": 0.0,
|
280 |
+
"eval_iou__backgound_": 0.0,
|
281 |
+
"eval_iou_aim": 0.04800987349583462,
|
282 |
+
"eval_iou_arrow": 0.0,
|
283 |
+
"eval_iou_chevron": 0.0,
|
284 |
+
"eval_iou_desig": 0.0,
|
285 |
+
"eval_iou_disp": 0.0,
|
286 |
+
"eval_iou_rwythr": 0.0,
|
287 |
+
"eval_iou_tdz": 0.011650709424385742,
|
288 |
+
"eval_iou_thrbar": 0.0,
|
289 |
+
"eval_loss": 0.023318186402320862,
|
290 |
+
"eval_mean_accuracy": 0.007473276299396342,
|
291 |
+
"eval_mean_iou": 0.006628953657802262,
|
292 |
+
"eval_overall_accuracy": 0.012193149440732584,
|
293 |
+
"eval_runtime": 71.6771,
|
294 |
+
"eval_samples_per_second": 1.214,
|
295 |
+
"eval_steps_per_second": 0.614,
|
296 |
+
"step": 1211
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 7.514450867052023,
|
300 |
+
"grad_norm": 0.3215765058994293,
|
301 |
+
"learning_rate": 3.404333333333333e-05,
|
302 |
+
"loss": 0.0235,
|
303 |
+
"step": 1300
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 8.0,
|
307 |
+
"eval_accuracy__backgound_": NaN,
|
308 |
+
"eval_accuracy_aim": 0.11078879150722133,
|
309 |
+
"eval_accuracy_arrow": 0.0,
|
310 |
+
"eval_accuracy_chevron": 0.0,
|
311 |
+
"eval_accuracy_desig": 0.0,
|
312 |
+
"eval_accuracy_disp": 0.0,
|
313 |
+
"eval_accuracy_rwythr": 0.002383080131069407,
|
314 |
+
"eval_accuracy_tdz": 0.02330187580100198,
|
315 |
+
"eval_accuracy_thrbar": 0.0,
|
316 |
+
"eval_iou__backgound_": 0.0,
|
317 |
+
"eval_iou_aim": 0.11068631682801998,
|
318 |
+
"eval_iou_arrow": 0.0,
|
319 |
+
"eval_iou_chevron": 0.0,
|
320 |
+
"eval_iou_desig": 0.0,
|
321 |
+
"eval_iou_disp": 0.0,
|
322 |
+
"eval_iou_rwythr": 0.002383080131069407,
|
323 |
+
"eval_iou_tdz": 0.022372615918116227,
|
324 |
+
"eval_iou_thrbar": 0.0,
|
325 |
+
"eval_loss": 0.02208574488759041,
|
326 |
+
"eval_mean_accuracy": 0.01705921842991159,
|
327 |
+
"eval_mean_iou": 0.015049112541911734,
|
328 |
+
"eval_overall_accuracy": 0.027708310004603536,
|
329 |
+
"eval_runtime": 74.1523,
|
330 |
+
"eval_samples_per_second": 1.173,
|
331 |
+
"eval_steps_per_second": 0.593,
|
332 |
+
"step": 1384
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 8.092485549132949,
|
336 |
+
"grad_norm": 0.08234093338251114,
|
337 |
+
"learning_rate": 3.2046666666666666e-05,
|
338 |
+
"loss": 0.0227,
|
339 |
+
"step": 1400
|
340 |
+
},
|
341 |
+
{
|
342 |
+
"epoch": 8.670520231213873,
|
343 |
+
"grad_norm": 0.11765847355127335,
|
344 |
+
"learning_rate": 3.0050000000000002e-05,
|
345 |
+
"loss": 0.0213,
|
346 |
+
"step": 1500
|
347 |
+
},
|
348 |
+
{
|
349 |
+
"epoch": 9.0,
|
350 |
+
"eval_accuracy__backgound_": NaN,
|
351 |
+
"eval_accuracy_aim": 0.1351067769411184,
|
352 |
+
"eval_accuracy_arrow": 0.0,
|
353 |
+
"eval_accuracy_chevron": 0.0,
|
354 |
+
"eval_accuracy_desig": 0.0,
|
355 |
+
"eval_accuracy_disp": 0.0,
|
356 |
+
"eval_accuracy_rwythr": 0.0016383675901102174,
|
357 |
+
"eval_accuracy_tdz": 0.02365140393801701,
|
358 |
+
"eval_accuracy_thrbar": 0.0,
|
359 |
+
"eval_iou__backgound_": 0.0,
|
360 |
+
"eval_iou_aim": 0.13460828926331325,
|
361 |
+
"eval_iou_arrow": 0.0,
|
362 |
+
"eval_iou_chevron": 0.0,
|
363 |
+
"eval_iou_desig": 0.0,
|
364 |
+
"eval_iou_disp": 0.0,
|
365 |
+
"eval_iou_rwythr": 0.0016383675901102174,
|
366 |
+
"eval_iou_tdz": 0.0230931118821455,
|
367 |
+
"eval_iou_thrbar": 0.0,
|
368 |
+
"eval_loss": 0.020945318043231964,
|
369 |
+
"eval_mean_accuracy": 0.020049568558655702,
|
370 |
+
"eval_mean_iou": 0.017704418748396553,
|
371 |
+
"eval_overall_accuracy": 0.032560685802446095,
|
372 |
+
"eval_runtime": 68.7392,
|
373 |
+
"eval_samples_per_second": 1.266,
|
374 |
+
"eval_steps_per_second": 0.64,
|
375 |
+
"step": 1557
|
376 |
+
},
|
377 |
+
{
|
378 |
+
"epoch": 9.248554913294798,
|
379 |
+
"grad_norm": 0.3116740882396698,
|
380 |
+
"learning_rate": 2.805333333333333e-05,
|
381 |
+
"loss": 0.0218,
|
382 |
+
"step": 1600
|
383 |
+
},
|
384 |
+
{
|
385 |
+
"epoch": 9.826589595375722,
|
386 |
+
"grad_norm": 0.3269786238670349,
|
387 |
+
"learning_rate": 2.6056666666666667e-05,
|
388 |
+
"loss": 0.0201,
|
389 |
+
"step": 1700
|
390 |
+
},
|
391 |
+
{
|
392 |
+
"epoch": 10.0,
|
393 |
+
"eval_accuracy__backgound_": NaN,
|
394 |
+
"eval_accuracy_aim": 0.17343537834835204,
|
395 |
+
"eval_accuracy_arrow": 0.0,
|
396 |
+
"eval_accuracy_chevron": 0.0,
|
397 |
+
"eval_accuracy_desig": 0.0,
|
398 |
+
"eval_accuracy_disp": 0.0,
|
399 |
+
"eval_accuracy_rwythr": 0.02114983616324099,
|
400 |
+
"eval_accuracy_tdz": 0.05947803798205756,
|
401 |
+
"eval_accuracy_thrbar": 0.0,
|
402 |
+
"eval_iou__backgound_": 0.0,
|
403 |
+
"eval_iou_aim": 0.1726150254929664,
|
404 |
+
"eval_iou_arrow": 0.0,
|
405 |
+
"eval_iou_chevron": 0.0,
|
406 |
+
"eval_iou_desig": 0.0,
|
407 |
+
"eval_iou_disp": 0.0,
|
408 |
+
"eval_iou_rwythr": 0.02114983616324099,
|
409 |
+
"eval_iou_tdz": 0.05587173032724089,
|
410 |
+
"eval_iou_thrbar": 0.0,
|
411 |
+
"eval_loss": 0.020602751523256302,
|
412 |
+
"eval_mean_accuracy": 0.03175790656170632,
|
413 |
+
"eval_mean_iou": 0.02773739910927203,
|
414 |
+
"eval_overall_accuracy": 0.05119878566185162,
|
415 |
+
"eval_runtime": 67.8225,
|
416 |
+
"eval_samples_per_second": 1.283,
|
417 |
+
"eval_steps_per_second": 0.649,
|
418 |
+
"step": 1730
|
419 |
+
},
|
420 |
+
{
|
421 |
+
"epoch": 10.404624277456648,
|
422 |
+
"grad_norm": 0.9702034592628479,
|
423 |
+
"learning_rate": 2.406e-05,
|
424 |
+
"loss": 0.0178,
|
425 |
+
"step": 1800
|
426 |
+
},
|
427 |
+
{
|
428 |
+
"epoch": 10.982658959537572,
|
429 |
+
"grad_norm": 0.1186259537935257,
|
430 |
+
"learning_rate": 2.2063333333333333e-05,
|
431 |
+
"loss": 0.0203,
|
432 |
+
"step": 1900
|
433 |
+
},
|
434 |
+
{
|
435 |
+
"epoch": 11.0,
|
436 |
+
"eval_accuracy__backgound_": NaN,
|
437 |
+
"eval_accuracy_aim": 0.1511541784964819,
|
438 |
+
"eval_accuracy_arrow": 0.0,
|
439 |
+
"eval_accuracy_chevron": 0.0,
|
440 |
+
"eval_accuracy_desig": 0.0,
|
441 |
+
"eval_accuracy_disp": 0.0,
|
442 |
+
"eval_accuracy_rwythr": 0.02770330652368186,
|
443 |
+
"eval_accuracy_tdz": 0.04631247815449144,
|
444 |
+
"eval_accuracy_thrbar": 0.0,
|
445 |
+
"eval_iou__backgound_": 0.0,
|
446 |
+
"eval_iou_aim": 0.1505131829635548,
|
447 |
+
"eval_iou_arrow": 0.0,
|
448 |
+
"eval_iou_chevron": 0.0,
|
449 |
+
"eval_iou_desig": 0.0,
|
450 |
+
"eval_iou_disp": 0.0,
|
451 |
+
"eval_iou_rwythr": 0.02770330652368186,
|
452 |
+
"eval_iou_tdz": 0.04319713105846555,
|
453 |
+
"eval_iou_thrbar": 0.0,
|
454 |
+
"eval_loss": 0.01977250538766384,
|
455 |
+
"eval_mean_accuracy": 0.0281462453968319,
|
456 |
+
"eval_mean_iou": 0.02460151339396691,
|
457 |
+
"eval_overall_accuracy": 0.044990233038458186,
|
458 |
+
"eval_runtime": 68.4515,
|
459 |
+
"eval_samples_per_second": 1.271,
|
460 |
+
"eval_steps_per_second": 0.643,
|
461 |
+
"step": 1903
|
462 |
+
},
|
463 |
+
{
|
464 |
+
"epoch": 11.560693641618498,
|
465 |
+
"grad_norm": 0.22837448120117188,
|
466 |
+
"learning_rate": 2.006666666666667e-05,
|
467 |
+
"loss": 0.0172,
|
468 |
+
"step": 2000
|
469 |
+
},
|
470 |
+
{
|
471 |
+
"epoch": 12.0,
|
472 |
+
"eval_accuracy__backgound_": NaN,
|
473 |
+
"eval_accuracy_aim": 0.21447969386495494,
|
474 |
+
"eval_accuracy_arrow": 0.0,
|
475 |
+
"eval_accuracy_chevron": 0.0,
|
476 |
+
"eval_accuracy_desig": 0.0,
|
477 |
+
"eval_accuracy_disp": 0.0,
|
478 |
+
"eval_accuracy_rwythr": 0.059204647006255585,
|
479 |
+
"eval_accuracy_tdz": 0.07439123849469882,
|
480 |
+
"eval_accuracy_thrbar": 0.0,
|
481 |
+
"eval_iou__backgound_": 0.0,
|
482 |
+
"eval_iou_aim": 0.21191608732772288,
|
483 |
+
"eval_iou_arrow": 0.0,
|
484 |
+
"eval_iou_chevron": 0.0,
|
485 |
+
"eval_iou_desig": 0.0,
|
486 |
+
"eval_iou_disp": 0.0,
|
487 |
+
"eval_iou_rwythr": 0.058936911557565426,
|
488 |
+
"eval_iou_tdz": 0.06803409696323921,
|
489 |
+
"eval_iou_thrbar": 0.0,
|
490 |
+
"eval_loss": 0.019217312335968018,
|
491 |
+
"eval_mean_accuracy": 0.04350944742073867,
|
492 |
+
"eval_mean_iou": 0.0376541217609475,
|
493 |
+
"eval_overall_accuracy": 0.06901571423239147,
|
494 |
+
"eval_runtime": 68.0189,
|
495 |
+
"eval_samples_per_second": 1.279,
|
496 |
+
"eval_steps_per_second": 0.647,
|
497 |
+
"step": 2076
|
498 |
+
},
|
499 |
+
{
|
500 |
+
"epoch": 12.138728323699421,
|
501 |
+
"grad_norm": 0.6837244629859924,
|
502 |
+
"learning_rate": 1.807e-05,
|
503 |
+
"loss": 0.0171,
|
504 |
+
"step": 2100
|
505 |
+
},
|
506 |
+
{
|
507 |
+
"epoch": 12.716763005780347,
|
508 |
+
"grad_norm": 0.17025157809257507,
|
509 |
+
"learning_rate": 1.6073333333333334e-05,
|
510 |
+
"loss": 0.0168,
|
511 |
+
"step": 2200
|
512 |
+
},
|
513 |
+
{
|
514 |
+
"epoch": 13.0,
|
515 |
+
"eval_accuracy__backgound_": NaN,
|
516 |
+
"eval_accuracy_aim": 0.18843352672509567,
|
517 |
+
"eval_accuracy_arrow": 0.0,
|
518 |
+
"eval_accuracy_chevron": 0.0,
|
519 |
+
"eval_accuracy_desig": 0.0,
|
520 |
+
"eval_accuracy_disp": 0.0,
|
521 |
+
"eval_accuracy_rwythr": 0.04617217753946976,
|
522 |
+
"eval_accuracy_tdz": 0.07037166491902598,
|
523 |
+
"eval_accuracy_thrbar": 0.0,
|
524 |
+
"eval_iou__backgound_": 0.0,
|
525 |
+
"eval_iou_aim": 0.187553753532375,
|
526 |
+
"eval_iou_arrow": 0.0,
|
527 |
+
"eval_iou_chevron": 0.0,
|
528 |
+
"eval_iou_desig": 0.0,
|
529 |
+
"eval_iou_disp": 0.0,
|
530 |
+
"eval_iou_rwythr": 0.04606240713224369,
|
531 |
+
"eval_iou_tdz": 0.06446104589114195,
|
532 |
+
"eval_iou_thrbar": 0.0,
|
533 |
+
"eval_loss": 0.018872098997235298,
|
534 |
+
"eval_mean_accuracy": 0.03812217114794893,
|
535 |
+
"eval_mean_iou": 0.03311968961730674,
|
536 |
+
"eval_overall_accuracy": 0.06072934940838341,
|
537 |
+
"eval_runtime": 67.7405,
|
538 |
+
"eval_samples_per_second": 1.284,
|
539 |
+
"eval_steps_per_second": 0.65,
|
540 |
+
"step": 2249
|
541 |
+
},
|
542 |
+
{
|
543 |
+
"epoch": 13.294797687861271,
|
544 |
+
"grad_norm": 0.11161621659994125,
|
545 |
+
"learning_rate": 1.4076666666666663e-05,
|
546 |
+
"loss": 0.0171,
|
547 |
+
"step": 2300
|
548 |
+
},
|
549 |
+
{
|
550 |
+
"epoch": 13.872832369942197,
|
551 |
+
"grad_norm": 0.22759991884231567,
|
552 |
+
"learning_rate": 1.2079999999999998e-05,
|
553 |
+
"loss": 0.0169,
|
554 |
+
"step": 2400
|
555 |
+
},
|
556 |
+
{
|
557 |
+
"epoch": 14.0,
|
558 |
+
"eval_accuracy__backgound_": NaN,
|
559 |
+
"eval_accuracy_aim": 0.21244290828292803,
|
560 |
+
"eval_accuracy_arrow": 0.0,
|
561 |
+
"eval_accuracy_chevron": 0.0,
|
562 |
+
"eval_accuracy_desig": 0.0,
|
563 |
+
"eval_accuracy_disp": 0.0,
|
564 |
+
"eval_accuracy_rwythr": 0.0627792672028597,
|
565 |
+
"eval_accuracy_tdz": 0.07858557613887918,
|
566 |
+
"eval_accuracy_thrbar": 0.0,
|
567 |
+
"eval_iou__backgound_": 0.0,
|
568 |
+
"eval_iou_aim": 0.21116564417177913,
|
569 |
+
"eval_iou_arrow": 0.0,
|
570 |
+
"eval_iou_chevron": 0.0,
|
571 |
+
"eval_iou_desig": 0.0,
|
572 |
+
"eval_iou_disp": 0.0,
|
573 |
+
"eval_iou_rwythr": 0.06229677800768549,
|
574 |
+
"eval_iou_tdz": 0.07162198035572073,
|
575 |
+
"eval_iou_thrbar": 0.0,
|
576 |
+
"eval_loss": 0.01845720410346985,
|
577 |
+
"eval_mean_accuracy": 0.04422596895308336,
|
578 |
+
"eval_mean_iou": 0.03834271139279837,
|
579 |
+
"eval_overall_accuracy": 0.07009816729498712,
|
580 |
+
"eval_runtime": 68.7802,
|
581 |
+
"eval_samples_per_second": 1.265,
|
582 |
+
"eval_steps_per_second": 0.64,
|
583 |
+
"step": 2422
|
584 |
+
},
|
585 |
+
{
|
586 |
+
"epoch": 14.45086705202312,
|
587 |
+
"grad_norm": 0.13816995918750763,
|
588 |
+
"learning_rate": 1.0083333333333332e-05,
|
589 |
+
"loss": 0.0172,
|
590 |
+
"step": 2500
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 15.0,
|
594 |
+
"eval_accuracy__backgound_": NaN,
|
595 |
+
"eval_accuracy_aim": 0.24626589309961733,
|
596 |
+
"eval_accuracy_arrow": 0.0,
|
597 |
+
"eval_accuracy_chevron": 0.0,
|
598 |
+
"eval_accuracy_desig": 0.0,
|
599 |
+
"eval_accuracy_disp": 0.0,
|
600 |
+
"eval_accuracy_rwythr": 0.10284480190646411,
|
601 |
+
"eval_accuracy_tdz": 0.09169288127694279,
|
602 |
+
"eval_accuracy_thrbar": 0.0,
|
603 |
+
"eval_iou__backgound_": 0.0,
|
604 |
+
"eval_iou_aim": 0.24432061723103302,
|
605 |
+
"eval_iou_arrow": 0.0,
|
606 |
+
"eval_iou_chevron": 0.0,
|
607 |
+
"eval_iou_desig": 0.0,
|
608 |
+
"eval_iou_disp": 0.0,
|
609 |
+
"eval_iou_rwythr": 0.10126118199149435,
|
610 |
+
"eval_iou_tdz": 0.08299936722210505,
|
611 |
+
"eval_iou_thrbar": 0.0,
|
612 |
+
"eval_loss": 0.018359314650297165,
|
613 |
+
"eval_mean_accuracy": 0.05510044703537803,
|
614 |
+
"eval_mean_iou": 0.047620129604959156,
|
615 |
+
"eval_overall_accuracy": 0.08640961516927326,
|
616 |
+
"eval_runtime": 68.3856,
|
617 |
+
"eval_samples_per_second": 1.272,
|
618 |
+
"eval_steps_per_second": 0.643,
|
619 |
+
"step": 2595
|
620 |
+
},
|
621 |
+
{
|
622 |
+
"epoch": 15.028901734104046,
|
623 |
+
"grad_norm": 0.2364118993282318,
|
624 |
+
"learning_rate": 8.086666666666665e-06,
|
625 |
+
"loss": 0.0169,
|
626 |
+
"step": 2600
|
627 |
+
},
|
628 |
+
{
|
629 |
+
"epoch": 15.606936416184972,
|
630 |
+
"grad_norm": 0.16393928229808807,
|
631 |
+
"learning_rate": 6.0899999999999984e-06,
|
632 |
+
"loss": 0.0159,
|
633 |
+
"step": 2700
|
634 |
+
},
|
635 |
+
{
|
636 |
+
"epoch": 16.0,
|
637 |
+
"eval_accuracy__backgound_": NaN,
|
638 |
+
"eval_accuracy_aim": 0.24929021108505123,
|
639 |
+
"eval_accuracy_arrow": 0.0,
|
640 |
+
"eval_accuracy_chevron": 0.0,
|
641 |
+
"eval_accuracy_desig": 0.0,
|
642 |
+
"eval_accuracy_disp": 0.0,
|
643 |
+
"eval_accuracy_rwythr": 0.12250521298778672,
|
644 |
+
"eval_accuracy_tdz": 0.12023767913317022,
|
645 |
+
"eval_accuracy_thrbar": 0.0,
|
646 |
+
"eval_iou__backgound_": 0.0,
|
647 |
+
"eval_iou_aim": 0.24682229283793694,
|
648 |
+
"eval_iou_arrow": 0.0,
|
649 |
+
"eval_iou_chevron": 0.0,
|
650 |
+
"eval_iou_desig": 0.0,
|
651 |
+
"eval_iou_disp": 0.0,
|
652 |
+
"eval_iou_rwythr": 0.11991543956845022,
|
653 |
+
"eval_iou_tdz": 0.1043900465304471,
|
654 |
+
"eval_iou_thrbar": 0.0,
|
655 |
+
"eval_loss": 0.018234960734844208,
|
656 |
+
"eval_mean_accuracy": 0.06150413790075102,
|
657 |
+
"eval_mean_iou": 0.05234753099298158,
|
658 |
+
"eval_overall_accuracy": 0.09640053251713884,
|
659 |
+
"eval_runtime": 67.7047,
|
660 |
+
"eval_samples_per_second": 1.285,
|
661 |
+
"eval_steps_per_second": 0.65,
|
662 |
+
"step": 2768
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 16.184971098265898,
|
666 |
+
"grad_norm": 0.11604613810777664,
|
667 |
+
"learning_rate": 4.093333333333332e-06,
|
668 |
+
"loss": 0.0153,
|
669 |
+
"step": 2800
|
670 |
+
},
|
671 |
+
{
|
672 |
+
"epoch": 16.76300578034682,
|
673 |
+
"grad_norm": 0.23215174674987793,
|
674 |
+
"learning_rate": 2.096666666666666e-06,
|
675 |
+
"loss": 0.0163,
|
676 |
+
"step": 2900
|
677 |
+
},
|
678 |
+
{
|
679 |
+
"epoch": 17.0,
|
680 |
+
"eval_accuracy__backgound_": NaN,
|
681 |
+
"eval_accuracy_aim": 0.24138995185779533,
|
682 |
+
"eval_accuracy_arrow": 0.0,
|
683 |
+
"eval_accuracy_chevron": 0.0,
|
684 |
+
"eval_accuracy_desig": 0.0,
|
685 |
+
"eval_accuracy_disp": 0.0,
|
686 |
+
"eval_accuracy_rwythr": 0.11669645516830504,
|
687 |
+
"eval_accuracy_tdz": 0.09868344401724338,
|
688 |
+
"eval_accuracy_thrbar": 0.0,
|
689 |
+
"eval_iou__backgound_": 0.0,
|
690 |
+
"eval_iou_aim": 0.2396739796543694,
|
691 |
+
"eval_iou_arrow": 0.0,
|
692 |
+
"eval_iou_chevron": 0.0,
|
693 |
+
"eval_iou_desig": 0.0,
|
694 |
+
"eval_iou_disp": 0.0,
|
695 |
+
"eval_iou_rwythr": 0.11457190904438108,
|
696 |
+
"eval_iou_tdz": 0.08854275559272423,
|
697 |
+
"eval_iou_thrbar": 0.0,
|
698 |
+
"eval_loss": 0.01814867928624153,
|
699 |
+
"eval_mean_accuracy": 0.05709623138041797,
|
700 |
+
"eval_mean_iou": 0.0491987382546083,
|
701 |
+
"eval_overall_accuracy": 0.08923394672340214,
|
702 |
+
"eval_runtime": 68.5645,
|
703 |
+
"eval_samples_per_second": 1.269,
|
704 |
+
"eval_steps_per_second": 0.642,
|
705 |
+
"step": 2941
|
706 |
+
},
|
707 |
+
{
|
708 |
+
"epoch": 17.341040462427745,
|
709 |
+
"grad_norm": 0.09813344478607178,
|
710 |
+
"learning_rate": 1e-07,
|
711 |
+
"loss": 0.0152,
|
712 |
+
"step": 3000
|
713 |
+
},
|
714 |
+
{
|
715 |
+
"epoch": 17.341040462427745,
|
716 |
+
"eval_accuracy__backgound_": NaN,
|
717 |
+
"eval_accuracy_aim": 0.220281446734971,
|
718 |
+
"eval_accuracy_arrow": 0.0,
|
719 |
+
"eval_accuracy_chevron": 0.0,
|
720 |
+
"eval_accuracy_desig": 0.0,
|
721 |
+
"eval_accuracy_disp": 0.0,
|
722 |
+
"eval_accuracy_rwythr": 0.09711051534107834,
|
723 |
+
"eval_accuracy_tdz": 0.09081906093440521,
|
724 |
+
"eval_accuracy_thrbar": 0.0,
|
725 |
+
"eval_iou__backgound_": 0.0,
|
726 |
+
"eval_iou_aim": 0.2189167637858063,
|
727 |
+
"eval_iou_arrow": 0.0,
|
728 |
+
"eval_iou_chevron": 0.0,
|
729 |
+
"eval_iou_desig": 0.0,
|
730 |
+
"eval_iou_disp": 0.0,
|
731 |
+
"eval_iou_rwythr": 0.09582598471487361,
|
732 |
+
"eval_iou_tdz": 0.08179004249514715,
|
733 |
+
"eval_iou_thrbar": 0.0,
|
734 |
+
"eval_loss": 0.01818845421075821,
|
735 |
+
"eval_mean_accuracy": 0.05102637787630682,
|
736 |
+
"eval_mean_iou": 0.044059198999536336,
|
737 |
+
"eval_overall_accuracy": 0.08002687469672651,
|
738 |
+
"eval_runtime": 69.51,
|
739 |
+
"eval_samples_per_second": 1.252,
|
740 |
+
"eval_steps_per_second": 0.633,
|
741 |
+
"step": 3000
|
742 |
+
},
|
743 |
+
{
|
744 |
+
"epoch": 17.341040462427745,
|
745 |
+
"step": 3000,
|
746 |
+
"total_flos": 1.05218665611264e+17,
|
747 |
+
"train_loss": 0.11730531366666158,
|
748 |
+
"train_runtime": 5867.3784,
|
749 |
+
"train_samples_per_second": 1.023,
|
750 |
+
"train_steps_per_second": 0.511
|
751 |
+
}
|
752 |
+
],
|
753 |
+
"logging_steps": 100,
|
754 |
+
"max_steps": 3000,
|
755 |
+
"num_input_tokens_seen": 0,
|
756 |
+
"num_train_epochs": 18,
|
757 |
+
"save_steps": 500,
|
758 |
+
"stateful_callbacks": {
|
759 |
+
"TrainerControl": {
|
760 |
+
"args": {
|
761 |
+
"should_epoch_stop": false,
|
762 |
+
"should_evaluate": false,
|
763 |
+
"should_log": false,
|
764 |
+
"should_save": true,
|
765 |
+
"should_training_stop": true
|
766 |
+
},
|
767 |
+
"attributes": {}
|
768 |
+
}
|
769 |
+
},
|
770 |
+
"total_flos": 1.05218665611264e+17,
|
771 |
+
"train_batch_size": 2,
|
772 |
+
"trial_name": null,
|
773 |
+
"trial_params": null
|
774 |
+
}
|