End of training
Browse files- all_results.json +14 -0
- eval_results.json +9 -0
- train_results.json +8 -0
- trainer_state.json +193 -0
all_results.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 4.0,
|
3 |
+
"eval_accuracy": 0.9036111798169676,
|
4 |
+
"eval_loss": 0.43492814898490906,
|
5 |
+
"eval_runtime": 86.0866,
|
6 |
+
"eval_samples": 40430,
|
7 |
+
"eval_samples_per_second": 469.643,
|
8 |
+
"eval_steps_per_second": 14.683,
|
9 |
+
"train_loss": 0.16138914258944828,
|
10 |
+
"train_runtime": 10416.0445,
|
11 |
+
"train_samples": 363846,
|
12 |
+
"train_samples_per_second": 139.725,
|
13 |
+
"train_steps_per_second": 1.091
|
14 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 4.0,
|
3 |
+
"eval_accuracy": 0.9036111798169676,
|
4 |
+
"eval_loss": 0.43492814898490906,
|
5 |
+
"eval_runtime": 86.0866,
|
6 |
+
"eval_samples": 40430,
|
7 |
+
"eval_samples_per_second": 469.643,
|
8 |
+
"eval_steps_per_second": 14.683
|
9 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 4.0,
|
3 |
+
"train_loss": 0.16138914258944828,
|
4 |
+
"train_runtime": 10416.0445,
|
5 |
+
"train_samples": 363846,
|
6 |
+
"train_samples_per_second": 139.725,
|
7 |
+
"train_steps_per_second": 1.091
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,193 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 3.999120531199156,
|
5 |
+
"global_step": 11368,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.18,
|
12 |
+
"learning_rate": 4.780524278676988e-05,
|
13 |
+
"loss": 0.3883,
|
14 |
+
"step": 500
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 0.35,
|
18 |
+
"learning_rate": 4.56060872624912e-05,
|
19 |
+
"loss": 0.3251,
|
20 |
+
"step": 1000
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.53,
|
24 |
+
"learning_rate": 4.340693173821253e-05,
|
25 |
+
"loss": 0.2984,
|
26 |
+
"step": 1500
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"epoch": 0.7,
|
30 |
+
"learning_rate": 4.121217452498241e-05,
|
31 |
+
"loss": 0.2895,
|
32 |
+
"step": 2000
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.88,
|
36 |
+
"learning_rate": 3.9013019000703735e-05,
|
37 |
+
"loss": 0.2782,
|
38 |
+
"step": 2500
|
39 |
+
},
|
40 |
+
{
|
41 |
+
"epoch": 1.0,
|
42 |
+
"eval_accuracy": 0.8864209745238684,
|
43 |
+
"eval_loss": 0.26690077781677246,
|
44 |
+
"eval_runtime": 86.4948,
|
45 |
+
"eval_samples_per_second": 467.427,
|
46 |
+
"eval_steps_per_second": 14.614,
|
47 |
+
"step": 2842
|
48 |
+
},
|
49 |
+
{
|
50 |
+
"epoch": 1.06,
|
51 |
+
"learning_rate": 3.6813863476425054e-05,
|
52 |
+
"loss": 0.2481,
|
53 |
+
"step": 3000
|
54 |
+
},
|
55 |
+
{
|
56 |
+
"epoch": 1.23,
|
57 |
+
"learning_rate": 3.461470795214638e-05,
|
58 |
+
"loss": 0.1938,
|
59 |
+
"step": 3500
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 1.41,
|
63 |
+
"learning_rate": 3.24155524278677e-05,
|
64 |
+
"loss": 0.1904,
|
65 |
+
"step": 4000
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 1.58,
|
69 |
+
"learning_rate": 3.0220795214637583e-05,
|
70 |
+
"loss": 0.1858,
|
71 |
+
"step": 4500
|
72 |
+
},
|
73 |
+
{
|
74 |
+
"epoch": 1.76,
|
75 |
+
"learning_rate": 2.8021639690358902e-05,
|
76 |
+
"loss": 0.1871,
|
77 |
+
"step": 5000
|
78 |
+
},
|
79 |
+
{
|
80 |
+
"epoch": 1.93,
|
81 |
+
"learning_rate": 2.5822484166080225e-05,
|
82 |
+
"loss": 0.1866,
|
83 |
+
"step": 5500
|
84 |
+
},
|
85 |
+
{
|
86 |
+
"epoch": 2.0,
|
87 |
+
"eval_accuracy": 0.899060103883255,
|
88 |
+
"eval_loss": 0.24605706334114075,
|
89 |
+
"eval_runtime": 86.4559,
|
90 |
+
"eval_samples_per_second": 467.637,
|
91 |
+
"eval_steps_per_second": 14.62,
|
92 |
+
"step": 5685
|
93 |
+
},
|
94 |
+
{
|
95 |
+
"epoch": 2.11,
|
96 |
+
"learning_rate": 2.3623328641801548e-05,
|
97 |
+
"loss": 0.1354,
|
98 |
+
"step": 6000
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 2.29,
|
102 |
+
"learning_rate": 2.1428571428571428e-05,
|
103 |
+
"loss": 0.1054,
|
104 |
+
"step": 6500
|
105 |
+
},
|
106 |
+
{
|
107 |
+
"epoch": 2.46,
|
108 |
+
"learning_rate": 1.923381421534131e-05,
|
109 |
+
"loss": 0.1011,
|
110 |
+
"step": 7000
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 2.64,
|
114 |
+
"learning_rate": 1.703465869106263e-05,
|
115 |
+
"loss": 0.1019,
|
116 |
+
"step": 7500
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 2.81,
|
120 |
+
"learning_rate": 1.4835503166783957e-05,
|
121 |
+
"loss": 0.1007,
|
122 |
+
"step": 8000
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 2.99,
|
126 |
+
"learning_rate": 1.2636347642505278e-05,
|
127 |
+
"loss": 0.0989,
|
128 |
+
"step": 8500
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 3.0,
|
132 |
+
"eval_accuracy": 0.9031412317585951,
|
133 |
+
"eval_loss": 0.3130354583263397,
|
134 |
+
"eval_runtime": 86.6189,
|
135 |
+
"eval_samples_per_second": 466.757,
|
136 |
+
"eval_steps_per_second": 14.593,
|
137 |
+
"step": 8527
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 3.17,
|
141 |
+
"learning_rate": 1.0437192118226603e-05,
|
142 |
+
"loss": 0.0466,
|
143 |
+
"step": 9000
|
144 |
+
},
|
145 |
+
{
|
146 |
+
"epoch": 3.34,
|
147 |
+
"learning_rate": 8.238036593947925e-06,
|
148 |
+
"loss": 0.0474,
|
149 |
+
"step": 9500
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 3.52,
|
153 |
+
"learning_rate": 6.038881069669247e-06,
|
154 |
+
"loss": 0.0431,
|
155 |
+
"step": 10000
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 3.69,
|
159 |
+
"learning_rate": 3.839725545390571e-06,
|
160 |
+
"loss": 0.0445,
|
161 |
+
"step": 10500
|
162 |
+
},
|
163 |
+
{
|
164 |
+
"epoch": 3.87,
|
165 |
+
"learning_rate": 1.640570021111893e-06,
|
166 |
+
"loss": 0.0416,
|
167 |
+
"step": 11000
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 4.0,
|
171 |
+
"eval_accuracy": 0.9036111798169676,
|
172 |
+
"eval_loss": 0.43492814898490906,
|
173 |
+
"eval_runtime": 86.4404,
|
174 |
+
"eval_samples_per_second": 467.721,
|
175 |
+
"eval_steps_per_second": 14.623,
|
176 |
+
"step": 11368
|
177 |
+
},
|
178 |
+
{
|
179 |
+
"epoch": 4.0,
|
180 |
+
"step": 11368,
|
181 |
+
"total_flos": 3.802061080709038e+17,
|
182 |
+
"train_loss": 0.16138914258944828,
|
183 |
+
"train_runtime": 10416.0445,
|
184 |
+
"train_samples_per_second": 139.725,
|
185 |
+
"train_steps_per_second": 1.091
|
186 |
+
}
|
187 |
+
],
|
188 |
+
"max_steps": 11368,
|
189 |
+
"num_train_epochs": 4,
|
190 |
+
"total_flos": 3.802061080709038e+17,
|
191 |
+
"trial_name": null,
|
192 |
+
"trial_params": null
|
193 |
+
}
|