PuxAI commited on
Commit
8a112a7
·
verified ·
1 Parent(s): b98b42c

Upload logs/train_log_Gemma3-4B-IT_seed123.json with huggingface_hub

Browse files
logs/train_log_Gemma3-4B-IT_seed123.json ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "loss": 4.6726,
4
+ "grad_norm": 5.59173059463501,
5
+ "learning_rate": 0.00018,
6
+ "epoch": 0.5,
7
+ "step": 10
8
+ },
9
+ {
10
+ "loss": 0.6936,
11
+ "grad_norm": 2.593961000442505,
12
+ "learning_rate": 0.00018,
13
+ "epoch": 1.0,
14
+ "step": 20
15
+ },
16
+ {
17
+ "loss": 0.2555,
18
+ "grad_norm": 0.7065151929855347,
19
+ "learning_rate": 0.0001577777777777778,
20
+ "epoch": 1.5,
21
+ "step": 30
22
+ },
23
+ {
24
+ "loss": 0.1841,
25
+ "grad_norm": 1.0202727317810059,
26
+ "learning_rate": 0.00013555555555555556,
27
+ "epoch": 2.0,
28
+ "step": 40
29
+ },
30
+ {
31
+ "loss": 0.1344,
32
+ "grad_norm": 0.9817218780517578,
33
+ "learning_rate": 0.00011333333333333334,
34
+ "epoch": 2.5,
35
+ "step": 50
36
+ },
37
+ {
38
+ "loss": 0.1296,
39
+ "grad_norm": 0.8052711486816406,
40
+ "learning_rate": 9.111111111111112e-05,
41
+ "epoch": 3.0,
42
+ "step": 60
43
+ },
44
+ {
45
+ "loss": 0.1017,
46
+ "grad_norm": 0.5530523657798767,
47
+ "learning_rate": 6.88888888888889e-05,
48
+ "epoch": 3.5,
49
+ "step": 70
50
+ },
51
+ {
52
+ "loss": 0.0789,
53
+ "grad_norm": 0.7885369062423706,
54
+ "learning_rate": 4.666666666666667e-05,
55
+ "epoch": 4.0,
56
+ "step": 80
57
+ },
58
+ {
59
+ "loss": 0.0671,
60
+ "grad_norm": 0.7642138004302979,
61
+ "learning_rate": 2.4444444444444445e-05,
62
+ "epoch": 4.5,
63
+ "step": 90
64
+ },
65
+ {
66
+ "loss": 0.0659,
67
+ "grad_norm": 0.976900041103363,
68
+ "learning_rate": 2.2222222222222225e-06,
69
+ "epoch": 5.0,
70
+ "step": 100
71
+ },
72
+ {
73
+ "train_runtime": 268.9097,
74
+ "train_samples_per_second": 23.354,
75
+ "train_steps_per_second": 0.372,
76
+ "total_flos": 2.585479150123776e+16,
77
+ "train_loss": 0.6383443474769592,
78
+ "epoch": 5.0,
79
+ "step": 100
80
+ }
81
+ ]