infinitejoy
commited on
Commit
•
ea85cc4
1
Parent(s):
3f166bb
End of training
Browse files- all_results.json +14 -0
- eval_results.json +9 -0
- train_results.json +8 -0
- trainer_state.json +205 -0
all_results.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 100.0,
|
3 |
+
"eval_loss": 0.817560613155365,
|
4 |
+
"eval_runtime": 10.3871,
|
5 |
+
"eval_samples": 213,
|
6 |
+
"eval_samples_per_second": 20.506,
|
7 |
+
"eval_steps_per_second": 1.348,
|
8 |
+
"eval_wer": 0.5818422796554009,
|
9 |
+
"train_loss": 1.675173250834147,
|
10 |
+
"train_runtime": 5040.7553,
|
11 |
+
"train_samples": 753,
|
12 |
+
"train_samples_per_second": 14.938,
|
13 |
+
"train_steps_per_second": 0.476
|
14 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 100.0,
|
3 |
+
"eval_loss": 0.817560613155365,
|
4 |
+
"eval_runtime": 10.3871,
|
5 |
+
"eval_samples": 213,
|
6 |
+
"eval_samples_per_second": 20.506,
|
7 |
+
"eval_steps_per_second": 1.348,
|
8 |
+
"eval_wer": 0.5818422796554009
|
9 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 100.0,
|
3 |
+
"train_loss": 1.675173250834147,
|
4 |
+
"train_runtime": 5040.7553,
|
5 |
+
"train_samples": 753,
|
6 |
+
"train_samples_per_second": 14.938,
|
7 |
+
"train_steps_per_second": 0.476
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,205 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 100.0,
|
5 |
+
"global_step": 2400,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 4.17,
|
12 |
+
"learning_rate": 5.82e-05,
|
13 |
+
"loss": 11.8246,
|
14 |
+
"step": 100
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 8.33,
|
18 |
+
"learning_rate": 0.0001182,
|
19 |
+
"loss": 3.9245,
|
20 |
+
"step": 200
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 12.5,
|
24 |
+
"learning_rate": 0.00017819999999999997,
|
25 |
+
"loss": 3.3755,
|
26 |
+
"step": 300
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 16.67,
|
30 |
+
"learning_rate": 0.0002382,
|
31 |
+
"loss": 3.2775,
|
32 |
+
"step": 400
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 20.83,
|
36 |
+
"learning_rate": 0.0002982,
|
37 |
+
"loss": 2.3957,
|
38 |
+
"step": 500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 20.83,
|
42 |
+
"eval_loss": 1.0925403833389282,
|
43 |
+
"eval_runtime": 10.3822,
|
44 |
+
"eval_samples_per_second": 20.516,
|
45 |
+
"eval_steps_per_second": 1.348,
|
46 |
+
"eval_wer": 0.8111332007952287,
|
47 |
+
"step": 500
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"epoch": 25.0,
|
51 |
+
"learning_rate": 0.0002846842105263158,
|
52 |
+
"loss": 1.5517,
|
53 |
+
"step": 600
|
54 |
+
},
|
55 |
+
{
|
56 |
+
"epoch": 29.17,
|
57 |
+
"learning_rate": 0.0002688947368421052,
|
58 |
+
"loss": 1.3096,
|
59 |
+
"step": 700
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 33.33,
|
63 |
+
"learning_rate": 0.0002531052631578947,
|
64 |
+
"loss": 1.2066,
|
65 |
+
"step": 800
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 37.5,
|
69 |
+
"learning_rate": 0.0002373157894736842,
|
70 |
+
"loss": 1.1194,
|
71 |
+
"step": 900
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 41.67,
|
75 |
+
"learning_rate": 0.00022152631578947366,
|
76 |
+
"loss": 1.0351,
|
77 |
+
"step": 1000
|
78 |
+
},
|
79 |
+
{
|
80 |
+
"epoch": 41.67,
|
81 |
+
"eval_loss": 0.7836905717849731,
|
82 |
+
"eval_runtime": 11.7781,
|
83 |
+
"eval_samples_per_second": 18.084,
|
84 |
+
"eval_steps_per_second": 1.189,
|
85 |
+
"eval_wer": 0.6573889993373094,
|
86 |
+
"step": 1000
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 45.83,
|
90 |
+
"learning_rate": 0.00020573684210526314,
|
91 |
+
"loss": 0.955,
|
92 |
+
"step": 1100
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 50.0,
|
96 |
+
"learning_rate": 0.00018994736842105263,
|
97 |
+
"loss": 0.8928,
|
98 |
+
"step": 1200
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 54.17,
|
102 |
+
"learning_rate": 0.00017415789473684208,
|
103 |
+
"loss": 0.8239,
|
104 |
+
"step": 1300
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 58.33,
|
108 |
+
"learning_rate": 0.00015836842105263158,
|
109 |
+
"loss": 0.7782,
|
110 |
+
"step": 1400
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 62.5,
|
114 |
+
"learning_rate": 0.00014257894736842105,
|
115 |
+
"loss": 0.7396,
|
116 |
+
"step": 1500
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 62.5,
|
120 |
+
"eval_loss": 0.7674338221549988,
|
121 |
+
"eval_runtime": 10.3819,
|
122 |
+
"eval_samples_per_second": 20.516,
|
123 |
+
"eval_steps_per_second": 1.348,
|
124 |
+
"eval_wer": 0.6083499005964215,
|
125 |
+
"step": 1500
|
126 |
+
},
|
127 |
+
{
|
128 |
+
"epoch": 66.67,
|
129 |
+
"learning_rate": 0.00012678947368421052,
|
130 |
+
"loss": 0.6811,
|
131 |
+
"step": 1600
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 70.83,
|
135 |
+
"learning_rate": 0.00011099999999999999,
|
136 |
+
"loss": 0.6438,
|
137 |
+
"step": 1700
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 75.0,
|
141 |
+
"learning_rate": 9.521052631578946e-05,
|
142 |
+
"loss": 0.6074,
|
143 |
+
"step": 1800
|
144 |
+
},
|
145 |
+
{
|
146 |
+
"epoch": 79.17,
|
147 |
+
"learning_rate": 7.942105263157894e-05,
|
148 |
+
"loss": 0.5791,
|
149 |
+
"step": 1900
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 83.33,
|
153 |
+
"learning_rate": 6.363157894736841e-05,
|
154 |
+
"loss": 0.5385,
|
155 |
+
"step": 2000
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 83.33,
|
159 |
+
"eval_loss": 0.801542341709137,
|
160 |
+
"eval_runtime": 10.3837,
|
161 |
+
"eval_samples_per_second": 20.513,
|
162 |
+
"eval_steps_per_second": 1.348,
|
163 |
+
"eval_wer": 0.5811795891318754,
|
164 |
+
"step": 2000
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 87.5,
|
168 |
+
"learning_rate": 4.784210526315789e-05,
|
169 |
+
"loss": 0.5167,
|
170 |
+
"step": 2100
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 91.67,
|
174 |
+
"learning_rate": 3.205263157894736e-05,
|
175 |
+
"loss": 0.495,
|
176 |
+
"step": 2200
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 95.83,
|
180 |
+
"learning_rate": 1.626315789473684e-05,
|
181 |
+
"loss": 0.4769,
|
182 |
+
"step": 2300
|
183 |
+
},
|
184 |
+
{
|
185 |
+
"epoch": 100.0,
|
186 |
+
"learning_rate": 4.7368421052631574e-07,
|
187 |
+
"loss": 0.4558,
|
188 |
+
"step": 2400
|
189 |
+
},
|
190 |
+
{
|
191 |
+
"epoch": 100.0,
|
192 |
+
"step": 2400,
|
193 |
+
"total_flos": 1.3881399754298573e+19,
|
194 |
+
"train_loss": 1.675173250834147,
|
195 |
+
"train_runtime": 5040.7553,
|
196 |
+
"train_samples_per_second": 14.938,
|
197 |
+
"train_steps_per_second": 0.476
|
198 |
+
}
|
199 |
+
],
|
200 |
+
"max_steps": 2400,
|
201 |
+
"num_train_epochs": 100,
|
202 |
+
"total_flos": 1.3881399754298573e+19,
|
203 |
+
"trial_name": null,
|
204 |
+
"trial_params": null
|
205 |
+
}
|