wav2vec2-large-xlsr-georgian / trainer_state.json
1
{
2
  "best_metric": null,
3
  "best_model_checkpoint": null,
4
  "epoch": 59.991596638655466,
5
  "global_step": 3540,
6
  "is_hyper_param_search": false,
7
  "is_local_process_zero": true,
8
  "is_world_process_zero": true,
9
  "log_history": [
10
    {
11
      "epoch": 5.08,
12
      "learning_rate": 0.0003546,
13
      "loss": 6.338,
14
      "step": 300
15
    },
16
    {
17
      "epoch": 5.08,
18
      "eval_loss": 3.1361563205718994,
19
      "eval_runtime": 51.9297,
20
      "eval_samples_per_second": 12.632,
21
      "eval_wer": 1.0,
22
      "step": 300
23
    },
24
    {
25
      "epoch": 10.17,
26
      "learning_rate": 0.0005715592105263158,
27
      "loss": 3.1258,
28
      "step": 600
29
    },
30
    {
31
      "epoch": 10.17,
32
      "eval_loss": 3.0660765171051025,
33
      "eval_runtime": 50.6635,
34
      "eval_samples_per_second": 12.948,
35
      "eval_wer": 1.0,
36
      "step": 600
37
    },
38
    {
39
      "epoch": 15.25,
40
      "learning_rate": 0.0005132368421052633,
41
      "loss": 1.4079,
42
      "step": 900
43
    },
44
    {
45
      "epoch": 15.25,
46
      "eval_loss": 0.5029116272926331,
47
      "eval_runtime": 51.4993,
48
      "eval_samples_per_second": 12.738,
49
      "eval_wer": 0.6116052520118593,
50
      "step": 900
51
    },
52
    {
53
      "epoch": 20.34,
54
      "learning_rate": 0.00045491447368421056,
55
      "loss": 0.1446,
56
      "step": 1200
57
    },
58
    {
59
      "epoch": 20.34,
60
      "eval_loss": 0.5113032460212708,
61
      "eval_runtime": 51.276,
62
      "eval_samples_per_second": 12.794,
63
      "eval_wer": 0.5307073274036426,
64
      "step": 1200
65
    },
66
    {
67
      "epoch": 25.42,
68
      "learning_rate": 0.0003965921052631579,
69
      "loss": 0.0714,
70
      "step": 1500
71
    },
72
    {
73
      "epoch": 25.42,
74
      "eval_loss": 0.5406283140182495,
75
      "eval_runtime": 51.4595,
76
      "eval_samples_per_second": 12.748,
77
      "eval_wer": 0.5006353240152478,
78
      "step": 1500
79
    },
80
    {
81
      "epoch": 30.5,
82
      "learning_rate": 0.00033826973684210527,
83
      "loss": 0.0459,
84
      "step": 1800
85
    },
86
    {
87
      "epoch": 30.5,
88
      "eval_loss": 0.5775919556617737,
89
      "eval_runtime": 51.4074,
90
      "eval_samples_per_second": 12.761,
91
      "eval_wer": 0.49767047861075814,
92
      "step": 1800
93
    },
94
    {
95
      "epoch": 35.59,
96
      "learning_rate": 0.0002799473684210526,
97
      "loss": 0.0372,
98
      "step": 2100
99
    },
100
    {
101
      "epoch": 35.59,
102
      "eval_loss": 0.5846429467201233,
103
      "eval_runtime": 51.8427,
104
      "eval_samples_per_second": 12.654,
105
      "eval_wer": 0.4805167301990682,
106
      "step": 2100
107
    },
108
    {
109
      "epoch": 40.67,
110
      "learning_rate": 0.00022162500000000003,
111
      "loss": 0.0297,
112
      "step": 2400
113
    },
114
    {
115
      "epoch": 40.67,
116
      "eval_loss": 0.5748910903930664,
117
      "eval_runtime": 51.453,
118
      "eval_samples_per_second": 12.75,
119
      "eval_wer": 0.4841168996188056,
120
      "step": 2400
121
    },
122
    {
123
      "epoch": 45.76,
124
      "learning_rate": 0.0001633026315789474,
125
      "loss": 0.0254,
126
      "step": 2700
127
    },
128
    {
129
      "epoch": 45.76,
130
      "eval_loss": 0.592321515083313,
131
      "eval_runtime": 51.3943,
132
      "eval_samples_per_second": 12.764,
133
      "eval_wer": 0.4614570097416349,
134
      "step": 2700
135
    },
136
    {
137
      "epoch": 50.84,
138
      "learning_rate": 0.00010498026315789474,
139
      "loss": 0.0198,
140
      "step": 3000
141
    },
142
    {
143
      "epoch": 50.84,
144
      "eval_loss": 0.5621945261955261,
145
      "eval_runtime": 51.5882,
146
      "eval_samples_per_second": 12.716,
147
      "eval_wer": 0.4404913172384583,
148
      "step": 3000
149
    },
150
    {
151
      "epoch": 55.92,
152
      "learning_rate": 4.665789473684211e-05,
153
      "loss": 0.0193,
154
      "step": 3300
155
    },
156
    {
157
      "epoch": 55.92,
158
      "eval_loss": 0.5657602548599243,
159
      "eval_runtime": 71.8217,
160
      "eval_samples_per_second": 9.134,
161
      "eval_wer": 0.4470563320626853,
162
      "step": 3300
163
    },
164
    {
165
      "epoch": 59.99,
166
      "step": 3540,
167
      "total_flos": 3.4970554953107223e+19,
168
      "train_runtime": 23737.912,
169
      "train_samples_per_second": 0.149
170
    }
171
  ],
172
  "max_steps": 3540,
173
  "num_train_epochs": 60,
174
  "total_flos": 3.4970554953107223e+19,
175
  "trial_name": null,
176
  "trial_params": null
177
}
178