Commit
•
e1a50a2
1
Parent(s):
be6e6ef
End of training
Browse files- README.md +3 -3
- all_results.json +13 -0
- eval_results.json +8 -0
- runs/May21_21-09-57_804716d6a90a/events.out.tfevents.1716327469.804716d6a90a.958.7 +3 -0
- train_results.json +8 -0
- trainer_state.json +1035 -0
README.md
CHANGED
@@ -22,7 +22,7 @@ model-index:
|
|
22 |
metrics:
|
23 |
- name: Accuracy
|
24 |
type: accuracy
|
25 |
-
value: 0.
|
26 |
---
|
27 |
|
28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -32,8 +32,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
32 |
|
33 |
This model is a fine-tuned version of [facebook/deit-base-distilled-patch16-224](https://huggingface.co/facebook/deit-base-distilled-patch16-224) on the imagefolder dataset.
|
34 |
It achieves the following results on the evaluation set:
|
35 |
-
- Loss: 0.
|
36 |
-
- Accuracy: 0.
|
37 |
|
38 |
## Model description
|
39 |
|
|
|
22 |
metrics:
|
23 |
- name: Accuracy
|
24 |
type: accuracy
|
25 |
+
value: 0.8354430379746836
|
26 |
---
|
27 |
|
28 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
32 |
|
33 |
This model is a fine-tuned version of [facebook/deit-base-distilled-patch16-224](https://huggingface.co/facebook/deit-base-distilled-patch16-224) on the imagefolder dataset.
|
34 |
It achieves the following results on the evaluation set:
|
35 |
+
- Loss: 0.4908
|
36 |
+
- Accuracy: 0.8354
|
37 |
|
38 |
## Model description
|
39 |
|
all_results.json
ADDED
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 85.71428571428571,
|
3 |
+
"eval_accuracy": 0.8354430379746836,
|
4 |
+
"eval_loss": 0.49078744649887085,
|
5 |
+
"eval_runtime": 1.1234,
|
6 |
+
"eval_samples_per_second": 70.325,
|
7 |
+
"eval_steps_per_second": 2.671,
|
8 |
+
"total_flos": 2.9362240500074496e+18,
|
9 |
+
"train_loss": 0.24529865741729737,
|
10 |
+
"train_runtime": 1655.9157,
|
11 |
+
"train_samples_per_second": 26.692,
|
12 |
+
"train_steps_per_second": 0.181
|
13 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 85.71428571428571,
|
3 |
+
"eval_accuracy": 0.8354430379746836,
|
4 |
+
"eval_loss": 0.49078744649887085,
|
5 |
+
"eval_runtime": 1.1234,
|
6 |
+
"eval_samples_per_second": 70.325,
|
7 |
+
"eval_steps_per_second": 2.671
|
8 |
+
}
|
runs/May21_21-09-57_804716d6a90a/events.out.tfevents.1716327469.804716d6a90a.958.7
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:bbd7294aeace86228443da1b3fa72698850b6901db629d73db9cc4d71b32bb3f
|
3 |
+
size 411
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 85.71428571428571,
|
3 |
+
"total_flos": 2.9362240500074496e+18,
|
4 |
+
"train_loss": 0.24529865741729737,
|
5 |
+
"train_runtime": 1655.9157,
|
6 |
+
"train_samples_per_second": 26.692,
|
7 |
+
"train_steps_per_second": 0.181
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,1035 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.8354430379746836,
|
3 |
+
"best_model_checkpoint": "deit-base-distilled-patch16-224-55-fold4/checkpoint-154",
|
4 |
+
"epoch": 85.71428571428571,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 300,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.8571428571428571,
|
13 |
+
"eval_accuracy": 0.5063291139240507,
|
14 |
+
"eval_loss": 0.829655110836029,
|
15 |
+
"eval_runtime": 1.0986,
|
16 |
+
"eval_samples_per_second": 71.91,
|
17 |
+
"eval_steps_per_second": 2.731,
|
18 |
+
"step": 3
|
19 |
+
},
|
20 |
+
{
|
21 |
+
"epoch": 2.0,
|
22 |
+
"eval_accuracy": 0.620253164556962,
|
23 |
+
"eval_loss": 0.6548452973365784,
|
24 |
+
"eval_runtime": 1.1343,
|
25 |
+
"eval_samples_per_second": 69.644,
|
26 |
+
"eval_steps_per_second": 2.645,
|
27 |
+
"step": 7
|
28 |
+
},
|
29 |
+
{
|
30 |
+
"epoch": 2.857142857142857,
|
31 |
+
"grad_norm": 2.468780040740967,
|
32 |
+
"learning_rate": 1.6666666666666667e-05,
|
33 |
+
"loss": 0.6826,
|
34 |
+
"step": 10
|
35 |
+
},
|
36 |
+
{
|
37 |
+
"epoch": 2.857142857142857,
|
38 |
+
"eval_accuracy": 0.5189873417721519,
|
39 |
+
"eval_loss": 0.6513058543205261,
|
40 |
+
"eval_runtime": 1.0973,
|
41 |
+
"eval_samples_per_second": 71.995,
|
42 |
+
"eval_steps_per_second": 2.734,
|
43 |
+
"step": 10
|
44 |
+
},
|
45 |
+
{
|
46 |
+
"epoch": 4.0,
|
47 |
+
"eval_accuracy": 0.7341772151898734,
|
48 |
+
"eval_loss": 0.6023642420768738,
|
49 |
+
"eval_runtime": 1.1154,
|
50 |
+
"eval_samples_per_second": 70.826,
|
51 |
+
"eval_steps_per_second": 2.69,
|
52 |
+
"step": 14
|
53 |
+
},
|
54 |
+
{
|
55 |
+
"epoch": 4.857142857142857,
|
56 |
+
"eval_accuracy": 0.5822784810126582,
|
57 |
+
"eval_loss": 0.6472656726837158,
|
58 |
+
"eval_runtime": 1.1141,
|
59 |
+
"eval_samples_per_second": 70.912,
|
60 |
+
"eval_steps_per_second": 2.693,
|
61 |
+
"step": 17
|
62 |
+
},
|
63 |
+
{
|
64 |
+
"epoch": 5.714285714285714,
|
65 |
+
"grad_norm": 2.0615956783294678,
|
66 |
+
"learning_rate": 3.3333333333333335e-05,
|
67 |
+
"loss": 0.6198,
|
68 |
+
"step": 20
|
69 |
+
},
|
70 |
+
{
|
71 |
+
"epoch": 6.0,
|
72 |
+
"eval_accuracy": 0.6835443037974683,
|
73 |
+
"eval_loss": 0.5716904997825623,
|
74 |
+
"eval_runtime": 1.1213,
|
75 |
+
"eval_samples_per_second": 70.456,
|
76 |
+
"eval_steps_per_second": 2.676,
|
77 |
+
"step": 21
|
78 |
+
},
|
79 |
+
{
|
80 |
+
"epoch": 6.857142857142857,
|
81 |
+
"eval_accuracy": 0.5569620253164557,
|
82 |
+
"eval_loss": 0.7837380766868591,
|
83 |
+
"eval_runtime": 1.1326,
|
84 |
+
"eval_samples_per_second": 69.751,
|
85 |
+
"eval_steps_per_second": 2.649,
|
86 |
+
"step": 24
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 8.0,
|
90 |
+
"eval_accuracy": 0.759493670886076,
|
91 |
+
"eval_loss": 0.5271920561790466,
|
92 |
+
"eval_runtime": 1.1388,
|
93 |
+
"eval_samples_per_second": 69.374,
|
94 |
+
"eval_steps_per_second": 2.634,
|
95 |
+
"step": 28
|
96 |
+
},
|
97 |
+
{
|
98 |
+
"epoch": 8.571428571428571,
|
99 |
+
"grad_norm": 5.306596279144287,
|
100 |
+
"learning_rate": 5e-05,
|
101 |
+
"loss": 0.5409,
|
102 |
+
"step": 30
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"epoch": 8.857142857142858,
|
106 |
+
"eval_accuracy": 0.7341772151898734,
|
107 |
+
"eval_loss": 0.5162191390991211,
|
108 |
+
"eval_runtime": 1.123,
|
109 |
+
"eval_samples_per_second": 70.344,
|
110 |
+
"eval_steps_per_second": 2.671,
|
111 |
+
"step": 31
|
112 |
+
},
|
113 |
+
{
|
114 |
+
"epoch": 10.0,
|
115 |
+
"eval_accuracy": 0.6708860759493671,
|
116 |
+
"eval_loss": 0.6155683994293213,
|
117 |
+
"eval_runtime": 1.1442,
|
118 |
+
"eval_samples_per_second": 69.046,
|
119 |
+
"eval_steps_per_second": 2.622,
|
120 |
+
"step": 35
|
121 |
+
},
|
122 |
+
{
|
123 |
+
"epoch": 10.857142857142858,
|
124 |
+
"eval_accuracy": 0.6582278481012658,
|
125 |
+
"eval_loss": 0.5843802094459534,
|
126 |
+
"eval_runtime": 1.1353,
|
127 |
+
"eval_samples_per_second": 69.586,
|
128 |
+
"eval_steps_per_second": 2.642,
|
129 |
+
"step": 38
|
130 |
+
},
|
131 |
+
{
|
132 |
+
"epoch": 11.428571428571429,
|
133 |
+
"grad_norm": 9.625292778015137,
|
134 |
+
"learning_rate": 4.814814814814815e-05,
|
135 |
+
"loss": 0.5177,
|
136 |
+
"step": 40
|
137 |
+
},
|
138 |
+
{
|
139 |
+
"epoch": 12.0,
|
140 |
+
"eval_accuracy": 0.6708860759493671,
|
141 |
+
"eval_loss": 0.7012839913368225,
|
142 |
+
"eval_runtime": 1.1292,
|
143 |
+
"eval_samples_per_second": 69.962,
|
144 |
+
"eval_steps_per_second": 2.657,
|
145 |
+
"step": 42
|
146 |
+
},
|
147 |
+
{
|
148 |
+
"epoch": 12.857142857142858,
|
149 |
+
"eval_accuracy": 0.7721518987341772,
|
150 |
+
"eval_loss": 0.5016757249832153,
|
151 |
+
"eval_runtime": 1.1469,
|
152 |
+
"eval_samples_per_second": 68.881,
|
153 |
+
"eval_steps_per_second": 2.616,
|
154 |
+
"step": 45
|
155 |
+
},
|
156 |
+
{
|
157 |
+
"epoch": 14.0,
|
158 |
+
"eval_accuracy": 0.7974683544303798,
|
159 |
+
"eval_loss": 0.4722585082054138,
|
160 |
+
"eval_runtime": 1.1187,
|
161 |
+
"eval_samples_per_second": 70.62,
|
162 |
+
"eval_steps_per_second": 2.682,
|
163 |
+
"step": 49
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 14.285714285714286,
|
167 |
+
"grad_norm": 2.9092471599578857,
|
168 |
+
"learning_rate": 4.62962962962963e-05,
|
169 |
+
"loss": 0.4995,
|
170 |
+
"step": 50
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 14.857142857142858,
|
174 |
+
"eval_accuracy": 0.6835443037974683,
|
175 |
+
"eval_loss": 0.7056770324707031,
|
176 |
+
"eval_runtime": 1.1276,
|
177 |
+
"eval_samples_per_second": 70.062,
|
178 |
+
"eval_steps_per_second": 2.661,
|
179 |
+
"step": 52
|
180 |
+
},
|
181 |
+
{
|
182 |
+
"epoch": 16.0,
|
183 |
+
"eval_accuracy": 0.8227848101265823,
|
184 |
+
"eval_loss": 0.46686723828315735,
|
185 |
+
"eval_runtime": 1.1347,
|
186 |
+
"eval_samples_per_second": 69.622,
|
187 |
+
"eval_steps_per_second": 2.644,
|
188 |
+
"step": 56
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 16.857142857142858,
|
192 |
+
"eval_accuracy": 0.7468354430379747,
|
193 |
+
"eval_loss": 0.5737015008926392,
|
194 |
+
"eval_runtime": 1.1389,
|
195 |
+
"eval_samples_per_second": 69.365,
|
196 |
+
"eval_steps_per_second": 2.634,
|
197 |
+
"step": 59
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 17.142857142857142,
|
201 |
+
"grad_norm": 9.162846565246582,
|
202 |
+
"learning_rate": 4.4444444444444447e-05,
|
203 |
+
"loss": 0.3767,
|
204 |
+
"step": 60
|
205 |
+
},
|
206 |
+
{
|
207 |
+
"epoch": 18.0,
|
208 |
+
"eval_accuracy": 0.810126582278481,
|
209 |
+
"eval_loss": 0.5034801959991455,
|
210 |
+
"eval_runtime": 1.1365,
|
211 |
+
"eval_samples_per_second": 69.513,
|
212 |
+
"eval_steps_per_second": 2.64,
|
213 |
+
"step": 63
|
214 |
+
},
|
215 |
+
{
|
216 |
+
"epoch": 18.857142857142858,
|
217 |
+
"eval_accuracy": 0.620253164556962,
|
218 |
+
"eval_loss": 0.899321436882019,
|
219 |
+
"eval_runtime": 1.1354,
|
220 |
+
"eval_samples_per_second": 69.582,
|
221 |
+
"eval_steps_per_second": 2.642,
|
222 |
+
"step": 66
|
223 |
+
},
|
224 |
+
{
|
225 |
+
"epoch": 20.0,
|
226 |
+
"grad_norm": 3.9183833599090576,
|
227 |
+
"learning_rate": 4.259259259259259e-05,
|
228 |
+
"loss": 0.3602,
|
229 |
+
"step": 70
|
230 |
+
},
|
231 |
+
{
|
232 |
+
"epoch": 20.0,
|
233 |
+
"eval_accuracy": 0.7974683544303798,
|
234 |
+
"eval_loss": 0.5424658060073853,
|
235 |
+
"eval_runtime": 1.1335,
|
236 |
+
"eval_samples_per_second": 69.693,
|
237 |
+
"eval_steps_per_second": 2.647,
|
238 |
+
"step": 70
|
239 |
+
},
|
240 |
+
{
|
241 |
+
"epoch": 20.857142857142858,
|
242 |
+
"eval_accuracy": 0.7721518987341772,
|
243 |
+
"eval_loss": 0.560451090335846,
|
244 |
+
"eval_runtime": 1.105,
|
245 |
+
"eval_samples_per_second": 71.49,
|
246 |
+
"eval_steps_per_second": 2.715,
|
247 |
+
"step": 73
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 22.0,
|
251 |
+
"eval_accuracy": 0.7974683544303798,
|
252 |
+
"eval_loss": 0.42595717310905457,
|
253 |
+
"eval_runtime": 1.1303,
|
254 |
+
"eval_samples_per_second": 69.894,
|
255 |
+
"eval_steps_per_second": 2.654,
|
256 |
+
"step": 77
|
257 |
+
},
|
258 |
+
{
|
259 |
+
"epoch": 22.857142857142858,
|
260 |
+
"grad_norm": 4.727555274963379,
|
261 |
+
"learning_rate": 4.074074074074074e-05,
|
262 |
+
"loss": 0.3479,
|
263 |
+
"step": 80
|
264 |
+
},
|
265 |
+
{
|
266 |
+
"epoch": 22.857142857142858,
|
267 |
+
"eval_accuracy": 0.8227848101265823,
|
268 |
+
"eval_loss": 0.4116832911968231,
|
269 |
+
"eval_runtime": 1.13,
|
270 |
+
"eval_samples_per_second": 69.912,
|
271 |
+
"eval_steps_per_second": 2.655,
|
272 |
+
"step": 80
|
273 |
+
},
|
274 |
+
{
|
275 |
+
"epoch": 24.0,
|
276 |
+
"eval_accuracy": 0.7974683544303798,
|
277 |
+
"eval_loss": 0.4017145037651062,
|
278 |
+
"eval_runtime": 1.1227,
|
279 |
+
"eval_samples_per_second": 70.364,
|
280 |
+
"eval_steps_per_second": 2.672,
|
281 |
+
"step": 84
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 24.857142857142858,
|
285 |
+
"eval_accuracy": 0.8227848101265823,
|
286 |
+
"eval_loss": 0.4401136338710785,
|
287 |
+
"eval_runtime": 1.152,
|
288 |
+
"eval_samples_per_second": 68.577,
|
289 |
+
"eval_steps_per_second": 2.604,
|
290 |
+
"step": 87
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 25.714285714285715,
|
294 |
+
"grad_norm": 3.6636221408843994,
|
295 |
+
"learning_rate": 3.888888888888889e-05,
|
296 |
+
"loss": 0.2845,
|
297 |
+
"step": 90
|
298 |
+
},
|
299 |
+
{
|
300 |
+
"epoch": 26.0,
|
301 |
+
"eval_accuracy": 0.7974683544303798,
|
302 |
+
"eval_loss": 0.44896697998046875,
|
303 |
+
"eval_runtime": 1.1191,
|
304 |
+
"eval_samples_per_second": 70.591,
|
305 |
+
"eval_steps_per_second": 2.681,
|
306 |
+
"step": 91
|
307 |
+
},
|
308 |
+
{
|
309 |
+
"epoch": 26.857142857142858,
|
310 |
+
"eval_accuracy": 0.810126582278481,
|
311 |
+
"eval_loss": 0.5079873204231262,
|
312 |
+
"eval_runtime": 1.1016,
|
313 |
+
"eval_samples_per_second": 71.715,
|
314 |
+
"eval_steps_per_second": 2.723,
|
315 |
+
"step": 94
|
316 |
+
},
|
317 |
+
{
|
318 |
+
"epoch": 28.0,
|
319 |
+
"eval_accuracy": 0.7974683544303798,
|
320 |
+
"eval_loss": 0.5014074444770813,
|
321 |
+
"eval_runtime": 1.1345,
|
322 |
+
"eval_samples_per_second": 69.632,
|
323 |
+
"eval_steps_per_second": 2.644,
|
324 |
+
"step": 98
|
325 |
+
},
|
326 |
+
{
|
327 |
+
"epoch": 28.571428571428573,
|
328 |
+
"grad_norm": 5.449157238006592,
|
329 |
+
"learning_rate": 3.7037037037037037e-05,
|
330 |
+
"loss": 0.2316,
|
331 |
+
"step": 100
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 28.857142857142858,
|
335 |
+
"eval_accuracy": 0.7721518987341772,
|
336 |
+
"eval_loss": 0.5384952425956726,
|
337 |
+
"eval_runtime": 1.1315,
|
338 |
+
"eval_samples_per_second": 69.82,
|
339 |
+
"eval_steps_per_second": 2.651,
|
340 |
+
"step": 101
|
341 |
+
},
|
342 |
+
{
|
343 |
+
"epoch": 30.0,
|
344 |
+
"eval_accuracy": 0.759493670886076,
|
345 |
+
"eval_loss": 0.5642827153205872,
|
346 |
+
"eval_runtime": 1.1314,
|
347 |
+
"eval_samples_per_second": 69.826,
|
348 |
+
"eval_steps_per_second": 2.652,
|
349 |
+
"step": 105
|
350 |
+
},
|
351 |
+
{
|
352 |
+
"epoch": 30.857142857142858,
|
353 |
+
"eval_accuracy": 0.7848101265822784,
|
354 |
+
"eval_loss": 0.48871567845344543,
|
355 |
+
"eval_runtime": 1.1405,
|
356 |
+
"eval_samples_per_second": 69.271,
|
357 |
+
"eval_steps_per_second": 2.631,
|
358 |
+
"step": 108
|
359 |
+
},
|
360 |
+
{
|
361 |
+
"epoch": 31.428571428571427,
|
362 |
+
"grad_norm": 4.334333896636963,
|
363 |
+
"learning_rate": 3.518518518518519e-05,
|
364 |
+
"loss": 0.2212,
|
365 |
+
"step": 110
|
366 |
+
},
|
367 |
+
{
|
368 |
+
"epoch": 32.0,
|
369 |
+
"eval_accuracy": 0.7974683544303798,
|
370 |
+
"eval_loss": 0.5253000855445862,
|
371 |
+
"eval_runtime": 1.1319,
|
372 |
+
"eval_samples_per_second": 69.795,
|
373 |
+
"eval_steps_per_second": 2.65,
|
374 |
+
"step": 112
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 32.857142857142854,
|
378 |
+
"eval_accuracy": 0.7974683544303798,
|
379 |
+
"eval_loss": 0.4632531702518463,
|
380 |
+
"eval_runtime": 1.1344,
|
381 |
+
"eval_samples_per_second": 69.638,
|
382 |
+
"eval_steps_per_second": 2.644,
|
383 |
+
"step": 115
|
384 |
+
},
|
385 |
+
{
|
386 |
+
"epoch": 34.0,
|
387 |
+
"eval_accuracy": 0.8227848101265823,
|
388 |
+
"eval_loss": 0.5078614354133606,
|
389 |
+
"eval_runtime": 1.1374,
|
390 |
+
"eval_samples_per_second": 69.459,
|
391 |
+
"eval_steps_per_second": 2.638,
|
392 |
+
"step": 119
|
393 |
+
},
|
394 |
+
{
|
395 |
+
"epoch": 34.285714285714285,
|
396 |
+
"grad_norm": 2.3969502449035645,
|
397 |
+
"learning_rate": 3.3333333333333335e-05,
|
398 |
+
"loss": 0.2093,
|
399 |
+
"step": 120
|
400 |
+
},
|
401 |
+
{
|
402 |
+
"epoch": 34.857142857142854,
|
403 |
+
"eval_accuracy": 0.7974683544303798,
|
404 |
+
"eval_loss": 0.508246660232544,
|
405 |
+
"eval_runtime": 1.135,
|
406 |
+
"eval_samples_per_second": 69.602,
|
407 |
+
"eval_steps_per_second": 2.643,
|
408 |
+
"step": 122
|
409 |
+
},
|
410 |
+
{
|
411 |
+
"epoch": 36.0,
|
412 |
+
"eval_accuracy": 0.7468354430379747,
|
413 |
+
"eval_loss": 0.6112725734710693,
|
414 |
+
"eval_runtime": 1.1521,
|
415 |
+
"eval_samples_per_second": 68.572,
|
416 |
+
"eval_steps_per_second": 2.604,
|
417 |
+
"step": 126
|
418 |
+
},
|
419 |
+
{
|
420 |
+
"epoch": 36.857142857142854,
|
421 |
+
"eval_accuracy": 0.7848101265822784,
|
422 |
+
"eval_loss": 0.5094159841537476,
|
423 |
+
"eval_runtime": 1.1118,
|
424 |
+
"eval_samples_per_second": 71.059,
|
425 |
+
"eval_steps_per_second": 2.698,
|
426 |
+
"step": 129
|
427 |
+
},
|
428 |
+
{
|
429 |
+
"epoch": 37.142857142857146,
|
430 |
+
"grad_norm": 4.463730335235596,
|
431 |
+
"learning_rate": 3.148148148148148e-05,
|
432 |
+
"loss": 0.179,
|
433 |
+
"step": 130
|
434 |
+
},
|
435 |
+
{
|
436 |
+
"epoch": 38.0,
|
437 |
+
"eval_accuracy": 0.810126582278481,
|
438 |
+
"eval_loss": 0.6253436207771301,
|
439 |
+
"eval_runtime": 1.1391,
|
440 |
+
"eval_samples_per_second": 69.35,
|
441 |
+
"eval_steps_per_second": 2.634,
|
442 |
+
"step": 133
|
443 |
+
},
|
444 |
+
{
|
445 |
+
"epoch": 38.857142857142854,
|
446 |
+
"eval_accuracy": 0.7848101265822784,
|
447 |
+
"eval_loss": 0.5406025648117065,
|
448 |
+
"eval_runtime": 1.1575,
|
449 |
+
"eval_samples_per_second": 68.253,
|
450 |
+
"eval_steps_per_second": 2.592,
|
451 |
+
"step": 136
|
452 |
+
},
|
453 |
+
{
|
454 |
+
"epoch": 40.0,
|
455 |
+
"grad_norm": 1.8232030868530273,
|
456 |
+
"learning_rate": 2.962962962962963e-05,
|
457 |
+
"loss": 0.173,
|
458 |
+
"step": 140
|
459 |
+
},
|
460 |
+
{
|
461 |
+
"epoch": 40.0,
|
462 |
+
"eval_accuracy": 0.7721518987341772,
|
463 |
+
"eval_loss": 0.6686509847640991,
|
464 |
+
"eval_runtime": 1.1413,
|
465 |
+
"eval_samples_per_second": 69.221,
|
466 |
+
"eval_steps_per_second": 2.629,
|
467 |
+
"step": 140
|
468 |
+
},
|
469 |
+
{
|
470 |
+
"epoch": 40.857142857142854,
|
471 |
+
"eval_accuracy": 0.7974683544303798,
|
472 |
+
"eval_loss": 0.5690881013870239,
|
473 |
+
"eval_runtime": 1.1296,
|
474 |
+
"eval_samples_per_second": 69.934,
|
475 |
+
"eval_steps_per_second": 2.656,
|
476 |
+
"step": 143
|
477 |
+
},
|
478 |
+
{
|
479 |
+
"epoch": 42.0,
|
480 |
+
"eval_accuracy": 0.7974683544303798,
|
481 |
+
"eval_loss": 0.5776450037956238,
|
482 |
+
"eval_runtime": 1.1443,
|
483 |
+
"eval_samples_per_second": 69.035,
|
484 |
+
"eval_steps_per_second": 2.622,
|
485 |
+
"step": 147
|
486 |
+
},
|
487 |
+
{
|
488 |
+
"epoch": 42.857142857142854,
|
489 |
+
"grad_norm": 4.304912090301514,
|
490 |
+
"learning_rate": 2.777777777777778e-05,
|
491 |
+
"loss": 0.1881,
|
492 |
+
"step": 150
|
493 |
+
},
|
494 |
+
{
|
495 |
+
"epoch": 42.857142857142854,
|
496 |
+
"eval_accuracy": 0.759493670886076,
|
497 |
+
"eval_loss": 0.781569242477417,
|
498 |
+
"eval_runtime": 1.1351,
|
499 |
+
"eval_samples_per_second": 69.599,
|
500 |
+
"eval_steps_per_second": 2.643,
|
501 |
+
"step": 150
|
502 |
+
},
|
503 |
+
{
|
504 |
+
"epoch": 44.0,
|
505 |
+
"eval_accuracy": 0.8354430379746836,
|
506 |
+
"eval_loss": 0.49078744649887085,
|
507 |
+
"eval_runtime": 1.133,
|
508 |
+
"eval_samples_per_second": 69.729,
|
509 |
+
"eval_steps_per_second": 2.648,
|
510 |
+
"step": 154
|
511 |
+
},
|
512 |
+
{
|
513 |
+
"epoch": 44.857142857142854,
|
514 |
+
"eval_accuracy": 0.7848101265822784,
|
515 |
+
"eval_loss": 0.5329591631889343,
|
516 |
+
"eval_runtime": 1.1588,
|
517 |
+
"eval_samples_per_second": 68.172,
|
518 |
+
"eval_steps_per_second": 2.589,
|
519 |
+
"step": 157
|
520 |
+
},
|
521 |
+
{
|
522 |
+
"epoch": 45.714285714285715,
|
523 |
+
"grad_norm": 1.6199402809143066,
|
524 |
+
"learning_rate": 2.5925925925925925e-05,
|
525 |
+
"loss": 0.161,
|
526 |
+
"step": 160
|
527 |
+
},
|
528 |
+
{
|
529 |
+
"epoch": 46.0,
|
530 |
+
"eval_accuracy": 0.7848101265822784,
|
531 |
+
"eval_loss": 0.5350536704063416,
|
532 |
+
"eval_runtime": 1.1191,
|
533 |
+
"eval_samples_per_second": 70.592,
|
534 |
+
"eval_steps_per_second": 2.681,
|
535 |
+
"step": 161
|
536 |
+
},
|
537 |
+
{
|
538 |
+
"epoch": 46.857142857142854,
|
539 |
+
"eval_accuracy": 0.7848101265822784,
|
540 |
+
"eval_loss": 0.66633540391922,
|
541 |
+
"eval_runtime": 1.1167,
|
542 |
+
"eval_samples_per_second": 70.747,
|
543 |
+
"eval_steps_per_second": 2.687,
|
544 |
+
"step": 164
|
545 |
+
},
|
546 |
+
{
|
547 |
+
"epoch": 48.0,
|
548 |
+
"eval_accuracy": 0.7974683544303798,
|
549 |
+
"eval_loss": 0.5382458567619324,
|
550 |
+
"eval_runtime": 1.1327,
|
551 |
+
"eval_samples_per_second": 69.744,
|
552 |
+
"eval_steps_per_second": 2.648,
|
553 |
+
"step": 168
|
554 |
+
},
|
555 |
+
{
|
556 |
+
"epoch": 48.57142857142857,
|
557 |
+
"grad_norm": 3.656897783279419,
|
558 |
+
"learning_rate": 2.4074074074074074e-05,
|
559 |
+
"loss": 0.1446,
|
560 |
+
"step": 170
|
561 |
+
},
|
562 |
+
{
|
563 |
+
"epoch": 48.857142857142854,
|
564 |
+
"eval_accuracy": 0.7848101265822784,
|
565 |
+
"eval_loss": 0.5876865983009338,
|
566 |
+
"eval_runtime": 1.1361,
|
567 |
+
"eval_samples_per_second": 69.536,
|
568 |
+
"eval_steps_per_second": 2.641,
|
569 |
+
"step": 171
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 50.0,
|
573 |
+
"eval_accuracy": 0.8227848101265823,
|
574 |
+
"eval_loss": 0.6855766773223877,
|
575 |
+
"eval_runtime": 1.1183,
|
576 |
+
"eval_samples_per_second": 70.643,
|
577 |
+
"eval_steps_per_second": 2.683,
|
578 |
+
"step": 175
|
579 |
+
},
|
580 |
+
{
|
581 |
+
"epoch": 50.857142857142854,
|
582 |
+
"eval_accuracy": 0.810126582278481,
|
583 |
+
"eval_loss": 0.6043643951416016,
|
584 |
+
"eval_runtime": 1.1585,
|
585 |
+
"eval_samples_per_second": 68.189,
|
586 |
+
"eval_steps_per_second": 2.589,
|
587 |
+
"step": 178
|
588 |
+
},
|
589 |
+
{
|
590 |
+
"epoch": 51.42857142857143,
|
591 |
+
"grad_norm": 3.2771971225738525,
|
592 |
+
"learning_rate": 2.2222222222222223e-05,
|
593 |
+
"loss": 0.1556,
|
594 |
+
"step": 180
|
595 |
+
},
|
596 |
+
{
|
597 |
+
"epoch": 52.0,
|
598 |
+
"eval_accuracy": 0.7848101265822784,
|
599 |
+
"eval_loss": 0.7954353094100952,
|
600 |
+
"eval_runtime": 1.1206,
|
601 |
+
"eval_samples_per_second": 70.5,
|
602 |
+
"eval_steps_per_second": 2.677,
|
603 |
+
"step": 182
|
604 |
+
},
|
605 |
+
{
|
606 |
+
"epoch": 52.857142857142854,
|
607 |
+
"eval_accuracy": 0.7848101265822784,
|
608 |
+
"eval_loss": 0.7432191967964172,
|
609 |
+
"eval_runtime": 1.116,
|
610 |
+
"eval_samples_per_second": 70.79,
|
611 |
+
"eval_steps_per_second": 2.688,
|
612 |
+
"step": 185
|
613 |
+
},
|
614 |
+
{
|
615 |
+
"epoch": 54.0,
|
616 |
+
"eval_accuracy": 0.810126582278481,
|
617 |
+
"eval_loss": 0.7895691394805908,
|
618 |
+
"eval_runtime": 1.1711,
|
619 |
+
"eval_samples_per_second": 67.46,
|
620 |
+
"eval_steps_per_second": 2.562,
|
621 |
+
"step": 189
|
622 |
+
},
|
623 |
+
{
|
624 |
+
"epoch": 54.285714285714285,
|
625 |
+
"grad_norm": 2.9898054599761963,
|
626 |
+
"learning_rate": 2.037037037037037e-05,
|
627 |
+
"loss": 0.1602,
|
628 |
+
"step": 190
|
629 |
+
},
|
630 |
+
{
|
631 |
+
"epoch": 54.857142857142854,
|
632 |
+
"eval_accuracy": 0.810126582278481,
|
633 |
+
"eval_loss": 0.8399502038955688,
|
634 |
+
"eval_runtime": 1.1363,
|
635 |
+
"eval_samples_per_second": 69.523,
|
636 |
+
"eval_steps_per_second": 2.64,
|
637 |
+
"step": 192
|
638 |
+
},
|
639 |
+
{
|
640 |
+
"epoch": 56.0,
|
641 |
+
"eval_accuracy": 0.7848101265822784,
|
642 |
+
"eval_loss": 0.8242945671081543,
|
643 |
+
"eval_runtime": 1.1359,
|
644 |
+
"eval_samples_per_second": 69.546,
|
645 |
+
"eval_steps_per_second": 2.641,
|
646 |
+
"step": 196
|
647 |
+
},
|
648 |
+
{
|
649 |
+
"epoch": 56.857142857142854,
|
650 |
+
"eval_accuracy": 0.8354430379746836,
|
651 |
+
"eval_loss": 0.6864063739776611,
|
652 |
+
"eval_runtime": 1.149,
|
653 |
+
"eval_samples_per_second": 68.757,
|
654 |
+
"eval_steps_per_second": 2.611,
|
655 |
+
"step": 199
|
656 |
+
},
|
657 |
+
{
|
658 |
+
"epoch": 57.142857142857146,
|
659 |
+
"grad_norm": 2.650709390640259,
|
660 |
+
"learning_rate": 1.8518518518518518e-05,
|
661 |
+
"loss": 0.1357,
|
662 |
+
"step": 200
|
663 |
+
},
|
664 |
+
{
|
665 |
+
"epoch": 58.0,
|
666 |
+
"eval_accuracy": 0.7721518987341772,
|
667 |
+
"eval_loss": 0.7131363749504089,
|
668 |
+
"eval_runtime": 1.1228,
|
669 |
+
"eval_samples_per_second": 70.362,
|
670 |
+
"eval_steps_per_second": 2.672,
|
671 |
+
"step": 203
|
672 |
+
},
|
673 |
+
{
|
674 |
+
"epoch": 58.857142857142854,
|
675 |
+
"eval_accuracy": 0.7721518987341772,
|
676 |
+
"eval_loss": 0.9191145300865173,
|
677 |
+
"eval_runtime": 1.1256,
|
678 |
+
"eval_samples_per_second": 70.186,
|
679 |
+
"eval_steps_per_second": 2.665,
|
680 |
+
"step": 206
|
681 |
+
},
|
682 |
+
{
|
683 |
+
"epoch": 60.0,
|
684 |
+
"grad_norm": 2.9468564987182617,
|
685 |
+
"learning_rate": 1.6666666666666667e-05,
|
686 |
+
"loss": 0.1262,
|
687 |
+
"step": 210
|
688 |
+
},
|
689 |
+
{
|
690 |
+
"epoch": 60.0,
|
691 |
+
"eval_accuracy": 0.7721518987341772,
|
692 |
+
"eval_loss": 0.7465390563011169,
|
693 |
+
"eval_runtime": 1.1233,
|
694 |
+
"eval_samples_per_second": 70.329,
|
695 |
+
"eval_steps_per_second": 2.671,
|
696 |
+
"step": 210
|
697 |
+
},
|
698 |
+
{
|
699 |
+
"epoch": 60.857142857142854,
|
700 |
+
"eval_accuracy": 0.7848101265822784,
|
701 |
+
"eval_loss": 0.7127454280853271,
|
702 |
+
"eval_runtime": 1.131,
|
703 |
+
"eval_samples_per_second": 69.847,
|
704 |
+
"eval_steps_per_second": 2.652,
|
705 |
+
"step": 213
|
706 |
+
},
|
707 |
+
{
|
708 |
+
"epoch": 62.0,
|
709 |
+
"eval_accuracy": 0.7974683544303798,
|
710 |
+
"eval_loss": 0.6973403692245483,
|
711 |
+
"eval_runtime": 1.1324,
|
712 |
+
"eval_samples_per_second": 69.763,
|
713 |
+
"eval_steps_per_second": 2.649,
|
714 |
+
"step": 217
|
715 |
+
},
|
716 |
+
{
|
717 |
+
"epoch": 62.857142857142854,
|
718 |
+
"grad_norm": 2.6824991703033447,
|
719 |
+
"learning_rate": 1.4814814814814815e-05,
|
720 |
+
"loss": 0.1323,
|
721 |
+
"step": 220
|
722 |
+
},
|
723 |
+
{
|
724 |
+
"epoch": 62.857142857142854,
|
725 |
+
"eval_accuracy": 0.7848101265822784,
|
726 |
+
"eval_loss": 0.7125291228294373,
|
727 |
+
"eval_runtime": 1.1478,
|
728 |
+
"eval_samples_per_second": 68.828,
|
729 |
+
"eval_steps_per_second": 2.614,
|
730 |
+
"step": 220
|
731 |
+
},
|
732 |
+
{
|
733 |
+
"epoch": 64.0,
|
734 |
+
"eval_accuracy": 0.7848101265822784,
|
735 |
+
"eval_loss": 0.7234599590301514,
|
736 |
+
"eval_runtime": 1.129,
|
737 |
+
"eval_samples_per_second": 69.973,
|
738 |
+
"eval_steps_per_second": 2.657,
|
739 |
+
"step": 224
|
740 |
+
},
|
741 |
+
{
|
742 |
+
"epoch": 64.85714285714286,
|
743 |
+
"eval_accuracy": 0.7974683544303798,
|
744 |
+
"eval_loss": 0.719983696937561,
|
745 |
+
"eval_runtime": 1.1239,
|
746 |
+
"eval_samples_per_second": 70.29,
|
747 |
+
"eval_steps_per_second": 2.669,
|
748 |
+
"step": 227
|
749 |
+
},
|
750 |
+
{
|
751 |
+
"epoch": 65.71428571428571,
|
752 |
+
"grad_norm": 3.247549057006836,
|
753 |
+
"learning_rate": 1.2962962962962962e-05,
|
754 |
+
"loss": 0.1258,
|
755 |
+
"step": 230
|
756 |
+
},
|
757 |
+
{
|
758 |
+
"epoch": 66.0,
|
759 |
+
"eval_accuracy": 0.7848101265822784,
|
760 |
+
"eval_loss": 0.7616466879844666,
|
761 |
+
"eval_runtime": 1.1313,
|
762 |
+
"eval_samples_per_second": 69.831,
|
763 |
+
"eval_steps_per_second": 2.652,
|
764 |
+
"step": 231
|
765 |
+
},
|
766 |
+
{
|
767 |
+
"epoch": 66.85714285714286,
|
768 |
+
"eval_accuracy": 0.7848101265822784,
|
769 |
+
"eval_loss": 0.8537306189537048,
|
770 |
+
"eval_runtime": 1.1248,
|
771 |
+
"eval_samples_per_second": 70.237,
|
772 |
+
"eval_steps_per_second": 2.667,
|
773 |
+
"step": 234
|
774 |
+
},
|
775 |
+
{
|
776 |
+
"epoch": 68.0,
|
777 |
+
"eval_accuracy": 0.7848101265822784,
|
778 |
+
"eval_loss": 0.8222808837890625,
|
779 |
+
"eval_runtime": 1.1211,
|
780 |
+
"eval_samples_per_second": 70.469,
|
781 |
+
"eval_steps_per_second": 2.676,
|
782 |
+
"step": 238
|
783 |
+
},
|
784 |
+
{
|
785 |
+
"epoch": 68.57142857142857,
|
786 |
+
"grad_norm": 1.8724284172058105,
|
787 |
+
"learning_rate": 1.1111111111111112e-05,
|
788 |
+
"loss": 0.1316,
|
789 |
+
"step": 240
|
790 |
+
},
|
791 |
+
{
|
792 |
+
"epoch": 68.85714285714286,
|
793 |
+
"eval_accuracy": 0.8227848101265823,
|
794 |
+
"eval_loss": 0.7751244902610779,
|
795 |
+
"eval_runtime": 1.1296,
|
796 |
+
"eval_samples_per_second": 69.935,
|
797 |
+
"eval_steps_per_second": 2.656,
|
798 |
+
"step": 241
|
799 |
+
},
|
800 |
+
{
|
801 |
+
"epoch": 70.0,
|
802 |
+
"eval_accuracy": 0.8227848101265823,
|
803 |
+
"eval_loss": 0.7689340710639954,
|
804 |
+
"eval_runtime": 1.132,
|
805 |
+
"eval_samples_per_second": 69.785,
|
806 |
+
"eval_steps_per_second": 2.65,
|
807 |
+
"step": 245
|
808 |
+
},
|
809 |
+
{
|
810 |
+
"epoch": 70.85714285714286,
|
811 |
+
"eval_accuracy": 0.810126582278481,
|
812 |
+
"eval_loss": 0.7751415371894836,
|
813 |
+
"eval_runtime": 1.1224,
|
814 |
+
"eval_samples_per_second": 70.386,
|
815 |
+
"eval_steps_per_second": 2.673,
|
816 |
+
"step": 248
|
817 |
+
},
|
818 |
+
{
|
819 |
+
"epoch": 71.42857142857143,
|
820 |
+
"grad_norm": 1.3768597841262817,
|
821 |
+
"learning_rate": 9.259259259259259e-06,
|
822 |
+
"loss": 0.1104,
|
823 |
+
"step": 250
|
824 |
+
},
|
825 |
+
{
|
826 |
+
"epoch": 72.0,
|
827 |
+
"eval_accuracy": 0.810126582278481,
|
828 |
+
"eval_loss": 0.7770127058029175,
|
829 |
+
"eval_runtime": 1.1362,
|
830 |
+
"eval_samples_per_second": 69.53,
|
831 |
+
"eval_steps_per_second": 2.64,
|
832 |
+
"step": 252
|
833 |
+
},
|
834 |
+
{
|
835 |
+
"epoch": 72.85714285714286,
|
836 |
+
"eval_accuracy": 0.7974683544303798,
|
837 |
+
"eval_loss": 0.7916089296340942,
|
838 |
+
"eval_runtime": 1.1361,
|
839 |
+
"eval_samples_per_second": 69.534,
|
840 |
+
"eval_steps_per_second": 2.641,
|
841 |
+
"step": 255
|
842 |
+
},
|
843 |
+
{
|
844 |
+
"epoch": 74.0,
|
845 |
+
"eval_accuracy": 0.7848101265822784,
|
846 |
+
"eval_loss": 0.7994998693466187,
|
847 |
+
"eval_runtime": 1.1345,
|
848 |
+
"eval_samples_per_second": 69.633,
|
849 |
+
"eval_steps_per_second": 2.644,
|
850 |
+
"step": 259
|
851 |
+
},
|
852 |
+
{
|
853 |
+
"epoch": 74.28571428571429,
|
854 |
+
"grad_norm": 1.7055776119232178,
|
855 |
+
"learning_rate": 7.4074074074074075e-06,
|
856 |
+
"loss": 0.1239,
|
857 |
+
"step": 260
|
858 |
+
},
|
859 |
+
{
|
860 |
+
"epoch": 74.85714285714286,
|
861 |
+
"eval_accuracy": 0.7848101265822784,
|
862 |
+
"eval_loss": 0.7980771064758301,
|
863 |
+
"eval_runtime": 1.1415,
|
864 |
+
"eval_samples_per_second": 69.209,
|
865 |
+
"eval_steps_per_second": 2.628,
|
866 |
+
"step": 262
|
867 |
+
},
|
868 |
+
{
|
869 |
+
"epoch": 76.0,
|
870 |
+
"eval_accuracy": 0.8227848101265823,
|
871 |
+
"eval_loss": 0.7889901995658875,
|
872 |
+
"eval_runtime": 1.1265,
|
873 |
+
"eval_samples_per_second": 70.129,
|
874 |
+
"eval_steps_per_second": 2.663,
|
875 |
+
"step": 266
|
876 |
+
},
|
877 |
+
{
|
878 |
+
"epoch": 76.85714285714286,
|
879 |
+
"eval_accuracy": 0.8227848101265823,
|
880 |
+
"eval_loss": 0.7888072729110718,
|
881 |
+
"eval_runtime": 1.1364,
|
882 |
+
"eval_samples_per_second": 69.519,
|
883 |
+
"eval_steps_per_second": 2.64,
|
884 |
+
"step": 269
|
885 |
+
},
|
886 |
+
{
|
887 |
+
"epoch": 77.14285714285714,
|
888 |
+
"grad_norm": 1.5552798509597778,
|
889 |
+
"learning_rate": 5.555555555555556e-06,
|
890 |
+
"loss": 0.1057,
|
891 |
+
"step": 270
|
892 |
+
},
|
893 |
+
{
|
894 |
+
"epoch": 78.0,
|
895 |
+
"eval_accuracy": 0.8227848101265823,
|
896 |
+
"eval_loss": 0.7956684231758118,
|
897 |
+
"eval_runtime": 1.1326,
|
898 |
+
"eval_samples_per_second": 69.75,
|
899 |
+
"eval_steps_per_second": 2.649,
|
900 |
+
"step": 273
|
901 |
+
},
|
902 |
+
{
|
903 |
+
"epoch": 78.85714285714286,
|
904 |
+
"eval_accuracy": 0.8227848101265823,
|
905 |
+
"eval_loss": 0.7978534698486328,
|
906 |
+
"eval_runtime": 1.1385,
|
907 |
+
"eval_samples_per_second": 69.387,
|
908 |
+
"eval_steps_per_second": 2.635,
|
909 |
+
"step": 276
|
910 |
+
},
|
911 |
+
{
|
912 |
+
"epoch": 80.0,
|
913 |
+
"grad_norm": 1.6866562366485596,
|
914 |
+
"learning_rate": 3.7037037037037037e-06,
|
915 |
+
"loss": 0.1015,
|
916 |
+
"step": 280
|
917 |
+
},
|
918 |
+
{
|
919 |
+
"epoch": 80.0,
|
920 |
+
"eval_accuracy": 0.810126582278481,
|
921 |
+
"eval_loss": 0.7958642244338989,
|
922 |
+
"eval_runtime": 1.1663,
|
923 |
+
"eval_samples_per_second": 67.738,
|
924 |
+
"eval_steps_per_second": 2.572,
|
925 |
+
"step": 280
|
926 |
+
},
|
927 |
+
{
|
928 |
+
"epoch": 80.85714285714286,
|
929 |
+
"eval_accuracy": 0.810126582278481,
|
930 |
+
"eval_loss": 0.7934866547584534,
|
931 |
+
"eval_runtime": 1.1341,
|
932 |
+
"eval_samples_per_second": 69.659,
|
933 |
+
"eval_steps_per_second": 2.645,
|
934 |
+
"step": 283
|
935 |
+
},
|
936 |
+
{
|
937 |
+
"epoch": 82.0,
|
938 |
+
"eval_accuracy": 0.8227848101265823,
|
939 |
+
"eval_loss": 0.7973299026489258,
|
940 |
+
"eval_runtime": 1.1543,
|
941 |
+
"eval_samples_per_second": 68.44,
|
942 |
+
"eval_steps_per_second": 2.599,
|
943 |
+
"step": 287
|
944 |
+
},
|
945 |
+
{
|
946 |
+
"epoch": 82.85714285714286,
|
947 |
+
"grad_norm": 2.308314085006714,
|
948 |
+
"learning_rate": 1.8518518518518519e-06,
|
949 |
+
"loss": 0.101,
|
950 |
+
"step": 290
|
951 |
+
},
|
952 |
+
{
|
953 |
+
"epoch": 82.85714285714286,
|
954 |
+
"eval_accuracy": 0.7974683544303798,
|
955 |
+
"eval_loss": 0.8032561540603638,
|
956 |
+
"eval_runtime": 1.1376,
|
957 |
+
"eval_samples_per_second": 69.442,
|
958 |
+
"eval_steps_per_second": 2.637,
|
959 |
+
"step": 290
|
960 |
+
},
|
961 |
+
{
|
962 |
+
"epoch": 84.0,
|
963 |
+
"eval_accuracy": 0.7974683544303798,
|
964 |
+
"eval_loss": 0.8071873784065247,
|
965 |
+
"eval_runtime": 1.1388,
|
966 |
+
"eval_samples_per_second": 69.372,
|
967 |
+
"eval_steps_per_second": 2.634,
|
968 |
+
"step": 294
|
969 |
+
},
|
970 |
+
{
|
971 |
+
"epoch": 84.85714285714286,
|
972 |
+
"eval_accuracy": 0.7974683544303798,
|
973 |
+
"eval_loss": 0.8049139976501465,
|
974 |
+
"eval_runtime": 1.1314,
|
975 |
+
"eval_samples_per_second": 69.822,
|
976 |
+
"eval_steps_per_second": 2.651,
|
977 |
+
"step": 297
|
978 |
+
},
|
979 |
+
{
|
980 |
+
"epoch": 85.71428571428571,
|
981 |
+
"grad_norm": 2.8343937397003174,
|
982 |
+
"learning_rate": 0.0,
|
983 |
+
"loss": 0.1116,
|
984 |
+
"step": 300
|
985 |
+
},
|
986 |
+
{
|
987 |
+
"epoch": 85.71428571428571,
|
988 |
+
"eval_accuracy": 0.810126582278481,
|
989 |
+
"eval_loss": 0.8041517734527588,
|
990 |
+
"eval_runtime": 1.1245,
|
991 |
+
"eval_samples_per_second": 70.256,
|
992 |
+
"eval_steps_per_second": 2.668,
|
993 |
+
"step": 300
|
994 |
+
},
|
995 |
+
{
|
996 |
+
"epoch": 85.71428571428571,
|
997 |
+
"step": 300,
|
998 |
+
"total_flos": 2.9362240500074496e+18,
|
999 |
+
"train_loss": 0.24529865741729737,
|
1000 |
+
"train_runtime": 1655.9157,
|
1001 |
+
"train_samples_per_second": 26.692,
|
1002 |
+
"train_steps_per_second": 0.181
|
1003 |
+
},
|
1004 |
+
{
|
1005 |
+
"epoch": 85.71428571428571,
|
1006 |
+
"eval_accuracy": 0.8354430379746836,
|
1007 |
+
"eval_loss": 0.49078744649887085,
|
1008 |
+
"eval_runtime": 1.1234,
|
1009 |
+
"eval_samples_per_second": 70.325,
|
1010 |
+
"eval_steps_per_second": 2.671,
|
1011 |
+
"step": 300
|
1012 |
+
}
|
1013 |
+
],
|
1014 |
+
"logging_steps": 10,
|
1015 |
+
"max_steps": 300,
|
1016 |
+
"num_input_tokens_seen": 0,
|
1017 |
+
"num_train_epochs": 100,
|
1018 |
+
"save_steps": 500,
|
1019 |
+
"stateful_callbacks": {
|
1020 |
+
"TrainerControl": {
|
1021 |
+
"args": {
|
1022 |
+
"should_epoch_stop": false,
|
1023 |
+
"should_evaluate": false,
|
1024 |
+
"should_log": false,
|
1025 |
+
"should_save": true,
|
1026 |
+
"should_training_stop": true
|
1027 |
+
},
|
1028 |
+
"attributes": {}
|
1029 |
+
}
|
1030 |
+
},
|
1031 |
+
"total_flos": 2.9362240500074496e+18,
|
1032 |
+
"train_batch_size": 32,
|
1033 |
+
"trial_name": null,
|
1034 |
+
"trial_params": null
|
1035 |
+
}
|