Training completed
Browse files- all_results.json +4 -4
- train_results.json +4 -4
- trainer_state.json +10 -10
all_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 0.0547945205479452,
|
3 |
"total_flos": 1.6697353660111258e+17,
|
4 |
-
"train_loss": 2.
|
5 |
-
"train_runtime":
|
6 |
-
"train_samples_per_second": 10.
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 0.0547945205479452,
|
3 |
"total_flos": 1.6697353660111258e+17,
|
4 |
+
"train_loss": 2.0328386863072714,
|
5 |
+
"train_runtime": 442.7248,
|
6 |
+
"train_samples_per_second": 10.842,
|
7 |
+
"train_steps_per_second": 0.068
|
8 |
}
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 0.0547945205479452,
|
3 |
"total_flos": 1.6697353660111258e+17,
|
4 |
-
"train_loss": 2.
|
5 |
-
"train_runtime":
|
6 |
-
"train_samples_per_second": 10.
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 0.0547945205479452,
|
3 |
"total_flos": 1.6697353660111258e+17,
|
4 |
+
"train_loss": 2.0328386863072714,
|
5 |
+
"train_runtime": 442.7248,
|
6 |
+
"train_samples_per_second": 10.842,
|
7 |
+
"train_steps_per_second": 0.068
|
8 |
}
|
trainer_state.json
CHANGED
@@ -10,33 +10,33 @@
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0182648401826484,
|
13 |
-
"grad_norm": 1.
|
14 |
"learning_rate": 0.0001,
|
15 |
-
"loss": 4.
|
16 |
"step": 10
|
17 |
},
|
18 |
{
|
19 |
"epoch": 0.0365296803652968,
|
20 |
-
"grad_norm": 4.
|
21 |
"learning_rate": 0.0001,
|
22 |
-
"loss": 1.
|
23 |
"step": 20
|
24 |
},
|
25 |
{
|
26 |
"epoch": 0.0547945205479452,
|
27 |
-
"grad_norm": 0.
|
28 |
"learning_rate": 0.0001,
|
29 |
-
"loss": 0.
|
30 |
"step": 30
|
31 |
},
|
32 |
{
|
33 |
"epoch": 0.0547945205479452,
|
34 |
"step": 30,
|
35 |
"total_flos": 1.6697353660111258e+17,
|
36 |
-
"train_loss": 2.
|
37 |
-
"train_runtime":
|
38 |
-
"train_samples_per_second": 10.
|
39 |
-
"train_steps_per_second": 0.
|
40 |
}
|
41 |
],
|
42 |
"logging_steps": 10,
|
|
|
10 |
"log_history": [
|
11 |
{
|
12 |
"epoch": 0.0182648401826484,
|
13 |
+
"grad_norm": 1.46377694606781,
|
14 |
"learning_rate": 0.0001,
|
15 |
+
"loss": 4.6773,
|
16 |
"step": 10
|
17 |
},
|
18 |
{
|
19 |
"epoch": 0.0365296803652968,
|
20 |
+
"grad_norm": 4.111562728881836,
|
21 |
"learning_rate": 0.0001,
|
22 |
+
"loss": 1.0524,
|
23 |
"step": 20
|
24 |
},
|
25 |
{
|
26 |
"epoch": 0.0547945205479452,
|
27 |
+
"grad_norm": 0.28213319182395935,
|
28 |
"learning_rate": 0.0001,
|
29 |
+
"loss": 0.3689,
|
30 |
"step": 30
|
31 |
},
|
32 |
{
|
33 |
"epoch": 0.0547945205479452,
|
34 |
"step": 30,
|
35 |
"total_flos": 1.6697353660111258e+17,
|
36 |
+
"train_loss": 2.0328386863072714,
|
37 |
+
"train_runtime": 442.7248,
|
38 |
+
"train_samples_per_second": 10.842,
|
39 |
+
"train_steps_per_second": 0.068
|
40 |
}
|
41 |
],
|
42 |
"logging_steps": 10,
|