PuxAI commited on
Commit
4cfd8f0
·
verified ·
1 Parent(s): 296926c

Upload logs/train_log_T5Gemma-2B_seed789.json with huggingface_hub

Browse files
logs/train_log_T5Gemma-2B_seed789.json ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "loss": 2.0368,
4
+ "grad_norm": 6.189993858337402,
5
+ "learning_rate": 0.00018,
6
+ "epoch": 0.5,
7
+ "step": 10
8
+ },
9
+ {
10
+ "loss": 0.2678,
11
+ "grad_norm": 1.55028235912323,
12
+ "learning_rate": 0.00019510565162951537,
13
+ "epoch": 1.0,
14
+ "step": 20
15
+ },
16
+ {
17
+ "loss": 0.1586,
18
+ "grad_norm": 1.0364527702331543,
19
+ "learning_rate": 0.00017880107536067218,
20
+ "epoch": 1.5,
21
+ "step": 30
22
+ },
23
+ {
24
+ "loss": 0.1385,
25
+ "grad_norm": 1.0755441188812256,
26
+ "learning_rate": 0.0001529919264233205,
27
+ "epoch": 2.0,
28
+ "step": 40
29
+ },
30
+ {
31
+ "loss": 0.0836,
32
+ "grad_norm": 0.8263890743255615,
33
+ "learning_rate": 0.00012079116908177593,
34
+ "epoch": 2.5,
35
+ "step": 50
36
+ },
37
+ {
38
+ "loss": 0.0688,
39
+ "grad_norm": 0.9091669321060181,
40
+ "learning_rate": 8.608268990399349e-05,
41
+ "epoch": 3.0,
42
+ "step": 60
43
+ },
44
+ {
45
+ "loss": 0.0358,
46
+ "grad_norm": 0.4862966537475586,
47
+ "learning_rate": 5.305284372141095e-05,
48
+ "epoch": 3.5,
49
+ "step": 70
50
+ },
51
+ {
52
+ "loss": 0.0353,
53
+ "grad_norm": 1.0672625303268433,
54
+ "learning_rate": 2.5685517452260567e-05,
55
+ "epoch": 4.0,
56
+ "step": 80
57
+ },
58
+ {
59
+ "loss": 0.0212,
60
+ "grad_norm": 0.45949041843414307,
61
+ "learning_rate": 7.281614543321269e-06,
62
+ "epoch": 4.5,
63
+ "step": 90
64
+ },
65
+ {
66
+ "loss": 0.02,
67
+ "grad_norm": 0.47117879986763,
68
+ "learning_rate": 6.09172980904238e-08,
69
+ "epoch": 5.0,
70
+ "step": 100
71
+ },
72
+ {
73
+ "train_runtime": 159.2404,
74
+ "train_samples_per_second": 39.437,
75
+ "train_steps_per_second": 0.628,
76
+ "total_flos": 9716776772935680.0,
77
+ "train_loss": 0.2866386969387531,
78
+ "epoch": 5.0,
79
+ "step": 100
80
+ }
81
+ ]