HorcruxNo13
commited on
Commit
•
b5125ea
1
Parent(s):
b8df258
Training in progress, epoch 1
Browse files- all_results.json +4 -4
- pytorch_model.bin +1 -1
- train_results.json +4 -4
- trainer_state.json +19 -19
- training_args.bin +1 -1
all_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 3.0,
|
3 |
"total_flos": 7.4567966957568e+16,
|
4 |
-
"train_loss": 0.
|
5 |
-
"train_runtime":
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 3.0,
|
3 |
"total_flos": 7.4567966957568e+16,
|
4 |
+
"train_loss": 0.5926340123017629,
|
5 |
+
"train_runtime": 76.7748,
|
6 |
+
"train_samples_per_second": 39.075,
|
7 |
+
"train_steps_per_second": 0.313
|
8 |
}
|
pytorch_model.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 110394865
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:db8deb248d3eebbc81fd58dec941edb204e79e84985e479801439fbd753a0752
|
3 |
size 110394865
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
"epoch": 3.0,
|
3 |
"total_flos": 7.4567966957568e+16,
|
4 |
-
"train_loss": 0.
|
5 |
-
"train_runtime":
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
"epoch": 3.0,
|
3 |
"total_flos": 7.4567966957568e+16,
|
4 |
+
"train_loss": 0.5926340123017629,
|
5 |
+
"train_runtime": 76.7748,
|
6 |
+
"train_samples_per_second": 39.075,
|
7 |
+
"train_steps_per_second": 0.313
|
8 |
}
|
trainer_state.json
CHANGED
@@ -11,50 +11,50 @@
|
|
11 |
{
|
12 |
"epoch": 1.0,
|
13 |
"eval_accuracy": 0.7333333333333333,
|
14 |
-
"eval_loss": 0.
|
15 |
-
"eval_runtime":
|
16 |
-
"eval_samples_per_second":
|
17 |
-
"eval_steps_per_second":
|
18 |
"step": 8
|
19 |
},
|
20 |
{
|
21 |
"epoch": 1.25,
|
22 |
"learning_rate": 3.3333333333333335e-05,
|
23 |
-
"loss": 0.
|
24 |
"step": 10
|
25 |
},
|
26 |
{
|
27 |
"epoch": 2.0,
|
28 |
"eval_accuracy": 0.7333333333333333,
|
29 |
-
"eval_loss": 0.
|
30 |
-
"eval_runtime":
|
31 |
-
"eval_samples_per_second":
|
32 |
-
"eval_steps_per_second":
|
33 |
"step": 16
|
34 |
},
|
35 |
{
|
36 |
"epoch": 2.5,
|
37 |
"learning_rate": 9.523809523809523e-06,
|
38 |
-
"loss": 0.
|
39 |
"step": 20
|
40 |
},
|
41 |
{
|
42 |
"epoch": 3.0,
|
43 |
-
"eval_accuracy": 0.
|
44 |
-
"eval_loss": 0.
|
45 |
-
"eval_runtime":
|
46 |
-
"eval_samples_per_second":
|
47 |
-
"eval_steps_per_second":
|
48 |
"step": 24
|
49 |
},
|
50 |
{
|
51 |
"epoch": 3.0,
|
52 |
"step": 24,
|
53 |
"total_flos": 7.4567966957568e+16,
|
54 |
-
"train_loss": 0.
|
55 |
-
"train_runtime":
|
56 |
-
"train_samples_per_second":
|
57 |
-
"train_steps_per_second": 0.
|
58 |
}
|
59 |
],
|
60 |
"logging_steps": 10,
|
|
|
11 |
{
|
12 |
"epoch": 1.0,
|
13 |
"eval_accuracy": 0.7333333333333333,
|
14 |
+
"eval_loss": 0.6174909472465515,
|
15 |
+
"eval_runtime": 1.9368,
|
16 |
+
"eval_samples_per_second": 123.919,
|
17 |
+
"eval_steps_per_second": 4.131,
|
18 |
"step": 8
|
19 |
},
|
20 |
{
|
21 |
"epoch": 1.25,
|
22 |
"learning_rate": 3.3333333333333335e-05,
|
23 |
+
"loss": 0.6122,
|
24 |
"step": 10
|
25 |
},
|
26 |
{
|
27 |
"epoch": 2.0,
|
28 |
"eval_accuracy": 0.7333333333333333,
|
29 |
+
"eval_loss": 0.578750729560852,
|
30 |
+
"eval_runtime": 1.969,
|
31 |
+
"eval_samples_per_second": 121.891,
|
32 |
+
"eval_steps_per_second": 4.063,
|
33 |
"step": 16
|
34 |
},
|
35 |
{
|
36 |
"epoch": 2.5,
|
37 |
"learning_rate": 9.523809523809523e-06,
|
38 |
+
"loss": 0.5704,
|
39 |
"step": 20
|
40 |
},
|
41 |
{
|
42 |
"epoch": 3.0,
|
43 |
+
"eval_accuracy": 0.7291666666666666,
|
44 |
+
"eval_loss": 0.5844461917877197,
|
45 |
+
"eval_runtime": 1.7247,
|
46 |
+
"eval_samples_per_second": 139.155,
|
47 |
+
"eval_steps_per_second": 4.639,
|
48 |
"step": 24
|
49 |
},
|
50 |
{
|
51 |
"epoch": 3.0,
|
52 |
"step": 24,
|
53 |
"total_flos": 7.4567966957568e+16,
|
54 |
+
"train_loss": 0.5926340123017629,
|
55 |
+
"train_runtime": 76.7748,
|
56 |
+
"train_samples_per_second": 39.075,
|
57 |
+
"train_steps_per_second": 0.313
|
58 |
}
|
59 |
],
|
60 |
"logging_steps": 10,
|
training_args.bin
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
size 4091
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:17c84ea59c2c834a63bf732c8f4ef9ae4ec3c6983fa813f621b98c510d0b86fe
|
3 |
size 4091
|