End of training
Browse files- all_results.json +11 -11
- eval_results.json +6 -6
- runs/Jan09_09-56-32_b4ebfdaa0ca7/events.out.tfevents.1673260552.b4ebfdaa0ca7.23.5 +3 -0
- train_results.json +6 -6
- trainer_state.json +351 -63
all_results.json
CHANGED
@@ -1,13 +1,13 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_accuracy": 0.
|
4 |
-
"eval_loss": 0.
|
5 |
-
"eval_runtime": 62.
|
6 |
-
"eval_samples_per_second": 115.
|
7 |
-
"eval_steps_per_second": 3.
|
8 |
-
"total_flos":
|
9 |
-
"train_loss": 1.
|
10 |
-
"train_runtime":
|
11 |
-
"train_samples_per_second":
|
12 |
-
"train_steps_per_second": 0.
|
13 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"eval_accuracy": 0.6840345500139314,
|
4 |
+
"eval_loss": 0.8481202721595764,
|
5 |
+
"eval_runtime": 62.1775,
|
6 |
+
"eval_samples_per_second": 115.444,
|
7 |
+
"eval_steps_per_second": 3.619,
|
8 |
+
"total_flos": 6.668732964123095e+18,
|
9 |
+
"train_loss": 1.1477188680853163,
|
10 |
+
"train_runtime": 2102.1014,
|
11 |
+
"train_samples_per_second": 40.972,
|
12 |
+
"train_steps_per_second": 0.32
|
13 |
}
|
eval_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_accuracy": 0.
|
4 |
-
"eval_loss": 0.
|
5 |
-
"eval_runtime": 62.
|
6 |
-
"eval_samples_per_second": 115.
|
7 |
-
"eval_steps_per_second": 3.
|
8 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"eval_accuracy": 0.6840345500139314,
|
4 |
+
"eval_loss": 0.8481202721595764,
|
5 |
+
"eval_runtime": 62.1775,
|
6 |
+
"eval_samples_per_second": 115.444,
|
7 |
+
"eval_steps_per_second": 3.619
|
8 |
}
|
runs/Jan09_09-56-32_b4ebfdaa0ca7/events.out.tfevents.1673260552.b4ebfdaa0ca7.23.5
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b61fce306217fb8a82dccfaa82eed720ca59083c2cde04d26d7c6f6f40d1ea0f
|
3 |
+
size 363
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"total_flos":
|
4 |
-
"train_loss": 1.
|
5 |
-
"train_runtime":
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 3.0,
|
3 |
+
"total_flos": 6.668732964123095e+18,
|
4 |
+
"train_loss": 1.1477188680853163,
|
5 |
+
"train_runtime": 2102.1014,
|
6 |
+
"train_samples_per_second": 40.972,
|
7 |
+
"train_steps_per_second": 0.32
|
8 |
}
|
trainer_state.json
CHANGED
@@ -1,166 +1,454 @@
|
|
1 |
{
|
2 |
-
"best_metric": 0.
|
3 |
-
"best_model_checkpoint": "beit-base-patch16-224-pt22k-ft22k-finetuned-FER2013-9e-05/checkpoint-
|
4 |
-
"epoch":
|
5 |
-
"global_step":
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
9 |
"log_history": [
|
10 |
{
|
11 |
"epoch": 0.04,
|
12 |
-
"learning_rate":
|
13 |
-
"loss": 1.
|
14 |
"step": 10
|
15 |
},
|
16 |
{
|
17 |
"epoch": 0.09,
|
18 |
-
"learning_rate":
|
19 |
-
"loss": 1.
|
20 |
"step": 20
|
21 |
},
|
22 |
{
|
23 |
"epoch": 0.13,
|
24 |
-
"learning_rate":
|
25 |
-
"loss": 1.
|
26 |
"step": 30
|
27 |
},
|
28 |
{
|
29 |
"epoch": 0.18,
|
30 |
-
"learning_rate":
|
31 |
-
"loss": 1.
|
32 |
"step": 40
|
33 |
},
|
34 |
{
|
35 |
"epoch": 0.22,
|
36 |
-
"learning_rate":
|
37 |
-
"loss": 1.
|
38 |
"step": 50
|
39 |
},
|
40 |
{
|
41 |
"epoch": 0.27,
|
42 |
-
"learning_rate": 7.
|
43 |
-
"loss": 1.
|
44 |
"step": 60
|
45 |
},
|
46 |
{
|
47 |
"epoch": 0.31,
|
48 |
-
"learning_rate":
|
49 |
-
"loss": 1.
|
50 |
"step": 70
|
51 |
},
|
52 |
{
|
53 |
"epoch": 0.36,
|
54 |
-
"learning_rate":
|
55 |
-
"loss": 1.
|
56 |
"step": 80
|
57 |
},
|
58 |
{
|
59 |
"epoch": 0.4,
|
60 |
-
"learning_rate":
|
61 |
-
"loss": 1.
|
62 |
"step": 90
|
63 |
},
|
64 |
{
|
65 |
"epoch": 0.45,
|
66 |
-
"learning_rate":
|
67 |
-
"loss": 1.
|
68 |
"step": 100
|
69 |
},
|
70 |
{
|
71 |
"epoch": 0.49,
|
72 |
-
"learning_rate":
|
73 |
-
"loss": 1.
|
74 |
"step": 110
|
75 |
},
|
76 |
{
|
77 |
"epoch": 0.53,
|
78 |
-
"learning_rate":
|
79 |
-
"loss": 1.
|
80 |
"step": 120
|
81 |
},
|
82 |
{
|
83 |
"epoch": 0.58,
|
84 |
-
"learning_rate":
|
85 |
-
"loss": 1.
|
86 |
"step": 130
|
87 |
},
|
88 |
{
|
89 |
"epoch": 0.62,
|
90 |
-
"learning_rate":
|
91 |
-
"loss": 1.
|
92 |
"step": 140
|
93 |
},
|
94 |
{
|
95 |
"epoch": 0.67,
|
96 |
-
"learning_rate":
|
97 |
-
"loss": 1.
|
98 |
"step": 150
|
99 |
},
|
100 |
{
|
101 |
"epoch": 0.71,
|
102 |
-
"learning_rate":
|
103 |
-
"loss": 1.
|
104 |
"step": 160
|
105 |
},
|
106 |
{
|
107 |
"epoch": 0.76,
|
108 |
-
"learning_rate":
|
109 |
-
"loss": 1.
|
110 |
"step": 170
|
111 |
},
|
112 |
{
|
113 |
"epoch": 0.8,
|
114 |
-
"learning_rate":
|
115 |
-
"loss": 1.
|
116 |
"step": 180
|
117 |
},
|
118 |
{
|
119 |
"epoch": 0.85,
|
120 |
-
"learning_rate":
|
121 |
-
"loss": 1.
|
122 |
"step": 190
|
123 |
},
|
124 |
{
|
125 |
"epoch": 0.89,
|
126 |
-
"learning_rate":
|
127 |
-
"loss": 1.
|
128 |
"step": 200
|
129 |
},
|
130 |
{
|
131 |
"epoch": 0.94,
|
132 |
-
"learning_rate": 6.
|
133 |
-
"loss": 1.
|
134 |
"step": 210
|
135 |
},
|
136 |
{
|
137 |
"epoch": 0.98,
|
138 |
-
"learning_rate":
|
139 |
-
"loss": 1.
|
140 |
"step": 220
|
141 |
},
|
142 |
{
|
143 |
"epoch": 1.0,
|
144 |
-
"eval_accuracy": 0.
|
145 |
-
"eval_loss":
|
146 |
-
"eval_runtime":
|
147 |
-
"eval_samples_per_second":
|
148 |
-
"eval_steps_per_second":
|
149 |
"step": 224
|
150 |
},
|
151 |
{
|
152 |
-
"epoch": 1.
|
153 |
-
"
|
154 |
-
"
|
155 |
-
"
|
156 |
-
|
157 |
-
|
158 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
159 |
}
|
160 |
],
|
161 |
-
"max_steps":
|
162 |
-
"num_train_epochs":
|
163 |
-
"total_flos":
|
164 |
"trial_name": null,
|
165 |
"trial_params": null
|
166 |
}
|
|
|
1 |
{
|
2 |
+
"best_metric": 0.6840345500139314,
|
3 |
+
"best_model_checkpoint": "beit-base-patch16-224-pt22k-ft22k-finetuned-FER2013-9e-05/checkpoint-672",
|
4 |
+
"epoch": 2.997772828507795,
|
5 |
+
"global_step": 672,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
9 |
"log_history": [
|
10 |
{
|
11 |
"epoch": 0.04,
|
12 |
+
"learning_rate": 1.323529411764706e-05,
|
13 |
+
"loss": 1.9205,
|
14 |
"step": 10
|
15 |
},
|
16 |
{
|
17 |
"epoch": 0.09,
|
18 |
+
"learning_rate": 2.647058823529412e-05,
|
19 |
+
"loss": 1.7213,
|
20 |
"step": 20
|
21 |
},
|
22 |
{
|
23 |
"epoch": 0.13,
|
24 |
+
"learning_rate": 3.970588235294118e-05,
|
25 |
+
"loss": 1.5455,
|
26 |
"step": 30
|
27 |
},
|
28 |
{
|
29 |
"epoch": 0.18,
|
30 |
+
"learning_rate": 5.294117647058824e-05,
|
31 |
+
"loss": 1.4512,
|
32 |
"step": 40
|
33 |
},
|
34 |
{
|
35 |
"epoch": 0.22,
|
36 |
+
"learning_rate": 6.61764705882353e-05,
|
37 |
+
"loss": 1.4027,
|
38 |
"step": 50
|
39 |
},
|
40 |
{
|
41 |
"epoch": 0.27,
|
42 |
+
"learning_rate": 7.941176470588236e-05,
|
43 |
+
"loss": 1.3752,
|
44 |
"step": 60
|
45 |
},
|
46 |
{
|
47 |
"epoch": 0.31,
|
48 |
+
"learning_rate": 8.97019867549669e-05,
|
49 |
+
"loss": 1.3787,
|
50 |
"step": 70
|
51 |
},
|
52 |
{
|
53 |
"epoch": 0.36,
|
54 |
+
"learning_rate": 8.821192052980134e-05,
|
55 |
+
"loss": 1.275,
|
56 |
"step": 80
|
57 |
},
|
58 |
{
|
59 |
"epoch": 0.4,
|
60 |
+
"learning_rate": 8.672185430463577e-05,
|
61 |
+
"loss": 1.3066,
|
62 |
"step": 90
|
63 |
},
|
64 |
{
|
65 |
"epoch": 0.45,
|
66 |
+
"learning_rate": 8.52317880794702e-05,
|
67 |
+
"loss": 1.3146,
|
68 |
"step": 100
|
69 |
},
|
70 |
{
|
71 |
"epoch": 0.49,
|
72 |
+
"learning_rate": 8.374172185430464e-05,
|
73 |
+
"loss": 1.2811,
|
74 |
"step": 110
|
75 |
},
|
76 |
{
|
77 |
"epoch": 0.53,
|
78 |
+
"learning_rate": 8.225165562913908e-05,
|
79 |
+
"loss": 1.2602,
|
80 |
"step": 120
|
81 |
},
|
82 |
{
|
83 |
"epoch": 0.58,
|
84 |
+
"learning_rate": 8.076158940397351e-05,
|
85 |
+
"loss": 1.1923,
|
86 |
"step": 130
|
87 |
},
|
88 |
{
|
89 |
"epoch": 0.62,
|
90 |
+
"learning_rate": 7.927152317880794e-05,
|
91 |
+
"loss": 1.2981,
|
92 |
"step": 140
|
93 |
},
|
94 |
{
|
95 |
"epoch": 0.67,
|
96 |
+
"learning_rate": 7.778145695364238e-05,
|
97 |
+
"loss": 1.2406,
|
98 |
"step": 150
|
99 |
},
|
100 |
{
|
101 |
"epoch": 0.71,
|
102 |
+
"learning_rate": 7.629139072847683e-05,
|
103 |
+
"loss": 1.3138,
|
104 |
"step": 160
|
105 |
},
|
106 |
{
|
107 |
"epoch": 0.76,
|
108 |
+
"learning_rate": 7.480132450331127e-05,
|
109 |
+
"loss": 1.2256,
|
110 |
"step": 170
|
111 |
},
|
112 |
{
|
113 |
"epoch": 0.8,
|
114 |
+
"learning_rate": 7.33112582781457e-05,
|
115 |
+
"loss": 1.1977,
|
116 |
"step": 180
|
117 |
},
|
118 |
{
|
119 |
"epoch": 0.85,
|
120 |
+
"learning_rate": 7.182119205298013e-05,
|
121 |
+
"loss": 1.2222,
|
122 |
"step": 190
|
123 |
},
|
124 |
{
|
125 |
"epoch": 0.89,
|
126 |
+
"learning_rate": 7.033112582781457e-05,
|
127 |
+
"loss": 1.1974,
|
128 |
"step": 200
|
129 |
},
|
130 |
{
|
131 |
"epoch": 0.94,
|
132 |
+
"learning_rate": 6.884105960264902e-05,
|
133 |
+
"loss": 1.1744,
|
134 |
"step": 210
|
135 |
},
|
136 |
{
|
137 |
"epoch": 0.98,
|
138 |
+
"learning_rate": 6.735099337748346e-05,
|
139 |
+
"loss": 1.1839,
|
140 |
"step": 220
|
141 |
},
|
142 |
{
|
143 |
"epoch": 1.0,
|
144 |
+
"eval_accuracy": 0.6120089161326274,
|
145 |
+
"eval_loss": 1.026611089706421,
|
146 |
+
"eval_runtime": 62.7487,
|
147 |
+
"eval_samples_per_second": 114.393,
|
148 |
+
"eval_steps_per_second": 3.586,
|
149 |
"step": 224
|
150 |
},
|
151 |
{
|
152 |
+
"epoch": 1.03,
|
153 |
+
"learning_rate": 6.586092715231789e-05,
|
154 |
+
"loss": 1.228,
|
155 |
+
"step": 230
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 1.07,
|
159 |
+
"learning_rate": 6.437086092715232e-05,
|
160 |
+
"loss": 1.1007,
|
161 |
+
"step": 240
|
162 |
+
},
|
163 |
+
{
|
164 |
+
"epoch": 1.12,
|
165 |
+
"learning_rate": 6.288079470198676e-05,
|
166 |
+
"loss": 1.1172,
|
167 |
+
"step": 250
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 1.16,
|
171 |
+
"learning_rate": 6.13907284768212e-05,
|
172 |
+
"loss": 1.1499,
|
173 |
+
"step": 260
|
174 |
+
},
|
175 |
+
{
|
176 |
+
"epoch": 1.2,
|
177 |
+
"learning_rate": 5.9900662251655627e-05,
|
178 |
+
"loss": 1.1565,
|
179 |
+
"step": 270
|
180 |
+
},
|
181 |
+
{
|
182 |
+
"epoch": 1.25,
|
183 |
+
"learning_rate": 5.841059602649007e-05,
|
184 |
+
"loss": 1.1459,
|
185 |
+
"step": 280
|
186 |
+
},
|
187 |
+
{
|
188 |
+
"epoch": 1.29,
|
189 |
+
"learning_rate": 5.6920529801324506e-05,
|
190 |
+
"loss": 1.1547,
|
191 |
+
"step": 290
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 1.34,
|
195 |
+
"learning_rate": 5.543046357615895e-05,
|
196 |
+
"loss": 1.0678,
|
197 |
+
"step": 300
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 1.38,
|
201 |
+
"learning_rate": 5.394039735099338e-05,
|
202 |
+
"loss": 1.13,
|
203 |
+
"step": 310
|
204 |
+
},
|
205 |
+
{
|
206 |
+
"epoch": 1.43,
|
207 |
+
"learning_rate": 5.2450331125827815e-05,
|
208 |
+
"loss": 1.1797,
|
209 |
+
"step": 320
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 1.47,
|
213 |
+
"learning_rate": 5.096026490066226e-05,
|
214 |
+
"loss": 1.0447,
|
215 |
+
"step": 330
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"epoch": 1.52,
|
219 |
+
"learning_rate": 4.9470198675496694e-05,
|
220 |
+
"loss": 1.0492,
|
221 |
+
"step": 340
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 1.56,
|
225 |
+
"learning_rate": 4.7980132450331124e-05,
|
226 |
+
"loss": 1.0945,
|
227 |
+
"step": 350
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 1.61,
|
231 |
+
"learning_rate": 4.6490066225165567e-05,
|
232 |
+
"loss": 1.0536,
|
233 |
+
"step": 360
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 1.65,
|
237 |
+
"learning_rate": 4.5e-05,
|
238 |
+
"loss": 1.075,
|
239 |
+
"step": 370
|
240 |
+
},
|
241 |
+
{
|
242 |
+
"epoch": 1.69,
|
243 |
+
"learning_rate": 4.350993377483444e-05,
|
244 |
+
"loss": 1.148,
|
245 |
+
"step": 380
|
246 |
+
},
|
247 |
+
{
|
248 |
+
"epoch": 1.74,
|
249 |
+
"learning_rate": 4.2019867549668875e-05,
|
250 |
+
"loss": 1.1247,
|
251 |
+
"step": 390
|
252 |
+
},
|
253 |
+
{
|
254 |
+
"epoch": 1.78,
|
255 |
+
"learning_rate": 4.052980132450332e-05,
|
256 |
+
"loss": 1.0691,
|
257 |
+
"step": 400
|
258 |
+
},
|
259 |
+
{
|
260 |
+
"epoch": 1.83,
|
261 |
+
"learning_rate": 3.903973509933775e-05,
|
262 |
+
"loss": 1.051,
|
263 |
+
"step": 410
|
264 |
+
},
|
265 |
+
{
|
266 |
+
"epoch": 1.87,
|
267 |
+
"learning_rate": 3.754966887417219e-05,
|
268 |
+
"loss": 1.1058,
|
269 |
+
"step": 420
|
270 |
+
},
|
271 |
+
{
|
272 |
+
"epoch": 1.92,
|
273 |
+
"learning_rate": 3.605960264900663e-05,
|
274 |
+
"loss": 1.0431,
|
275 |
+
"step": 430
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 1.96,
|
279 |
+
"learning_rate": 3.4569536423841064e-05,
|
280 |
+
"loss": 1.0333,
|
281 |
+
"step": 440
|
282 |
+
},
|
283 |
+
{
|
284 |
+
"epoch": 2.0,
|
285 |
+
"eval_accuracy": 0.6607690164391196,
|
286 |
+
"eval_loss": 0.9063372015953064,
|
287 |
+
"eval_runtime": 61.8075,
|
288 |
+
"eval_samples_per_second": 116.135,
|
289 |
+
"eval_steps_per_second": 3.64,
|
290 |
+
"step": 448
|
291 |
+
},
|
292 |
+
{
|
293 |
+
"epoch": 2.01,
|
294 |
+
"learning_rate": 3.30794701986755e-05,
|
295 |
+
"loss": 1.0725,
|
296 |
+
"step": 450
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 2.05,
|
300 |
+
"learning_rate": 3.1589403973509936e-05,
|
301 |
+
"loss": 0.9962,
|
302 |
+
"step": 460
|
303 |
+
},
|
304 |
+
{
|
305 |
+
"epoch": 2.1,
|
306 |
+
"learning_rate": 3.0099337748344372e-05,
|
307 |
+
"loss": 1.0569,
|
308 |
+
"step": 470
|
309 |
+
},
|
310 |
+
{
|
311 |
+
"epoch": 2.14,
|
312 |
+
"learning_rate": 2.8609271523178812e-05,
|
313 |
+
"loss": 0.9909,
|
314 |
+
"step": 480
|
315 |
+
},
|
316 |
+
{
|
317 |
+
"epoch": 2.19,
|
318 |
+
"learning_rate": 2.7119205298013245e-05,
|
319 |
+
"loss": 1.0458,
|
320 |
+
"step": 490
|
321 |
+
},
|
322 |
+
{
|
323 |
+
"epoch": 2.23,
|
324 |
+
"learning_rate": 2.5629139072847685e-05,
|
325 |
+
"loss": 1.0103,
|
326 |
+
"step": 500
|
327 |
+
},
|
328 |
+
{
|
329 |
+
"epoch": 2.28,
|
330 |
+
"learning_rate": 2.4139072847682117e-05,
|
331 |
+
"loss": 1.0123,
|
332 |
+
"step": 510
|
333 |
+
},
|
334 |
+
{
|
335 |
+
"epoch": 2.32,
|
336 |
+
"learning_rate": 2.2649006622516557e-05,
|
337 |
+
"loss": 1.005,
|
338 |
+
"step": 520
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 2.37,
|
342 |
+
"learning_rate": 2.1158940397350997e-05,
|
343 |
+
"loss": 1.0186,
|
344 |
+
"step": 530
|
345 |
+
},
|
346 |
+
{
|
347 |
+
"epoch": 2.41,
|
348 |
+
"learning_rate": 1.9668874172185433e-05,
|
349 |
+
"loss": 0.9912,
|
350 |
+
"step": 540
|
351 |
+
},
|
352 |
+
{
|
353 |
+
"epoch": 2.45,
|
354 |
+
"learning_rate": 1.817880794701987e-05,
|
355 |
+
"loss": 0.971,
|
356 |
+
"step": 550
|
357 |
+
},
|
358 |
+
{
|
359 |
+
"epoch": 2.5,
|
360 |
+
"learning_rate": 1.6688741721854306e-05,
|
361 |
+
"loss": 1.0455,
|
362 |
+
"step": 560
|
363 |
+
},
|
364 |
+
{
|
365 |
+
"epoch": 2.54,
|
366 |
+
"learning_rate": 1.5198675496688744e-05,
|
367 |
+
"loss": 0.9897,
|
368 |
+
"step": 570
|
369 |
+
},
|
370 |
+
{
|
371 |
+
"epoch": 2.59,
|
372 |
+
"learning_rate": 1.370860927152318e-05,
|
373 |
+
"loss": 1.0151,
|
374 |
+
"step": 580
|
375 |
+
},
|
376 |
+
{
|
377 |
+
"epoch": 2.63,
|
378 |
+
"learning_rate": 1.2218543046357616e-05,
|
379 |
+
"loss": 1.0069,
|
380 |
+
"step": 590
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 2.68,
|
384 |
+
"learning_rate": 1.0728476821192054e-05,
|
385 |
+
"loss": 1.0242,
|
386 |
+
"step": 600
|
387 |
+
},
|
388 |
+
{
|
389 |
+
"epoch": 2.72,
|
390 |
+
"learning_rate": 9.23841059602649e-06,
|
391 |
+
"loss": 1.0147,
|
392 |
+
"step": 610
|
393 |
+
},
|
394 |
+
{
|
395 |
+
"epoch": 2.77,
|
396 |
+
"learning_rate": 7.748344370860927e-06,
|
397 |
+
"loss": 0.9892,
|
398 |
+
"step": 620
|
399 |
+
},
|
400 |
+
{
|
401 |
+
"epoch": 2.81,
|
402 |
+
"learning_rate": 6.258278145695364e-06,
|
403 |
+
"loss": 0.9651,
|
404 |
+
"step": 630
|
405 |
+
},
|
406 |
+
{
|
407 |
+
"epoch": 2.86,
|
408 |
+
"learning_rate": 4.768211920529802e-06,
|
409 |
+
"loss": 0.9636,
|
410 |
+
"step": 640
|
411 |
+
},
|
412 |
+
{
|
413 |
+
"epoch": 2.9,
|
414 |
+
"learning_rate": 3.2781456953642384e-06,
|
415 |
+
"loss": 0.9405,
|
416 |
+
"step": 650
|
417 |
+
},
|
418 |
+
{
|
419 |
+
"epoch": 2.94,
|
420 |
+
"learning_rate": 1.7880794701986758e-06,
|
421 |
+
"loss": 1.0278,
|
422 |
+
"step": 660
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 2.99,
|
426 |
+
"learning_rate": 2.980132450331126e-07,
|
427 |
+
"loss": 0.9655,
|
428 |
+
"step": 670
|
429 |
+
},
|
430 |
+
{
|
431 |
+
"epoch": 3.0,
|
432 |
+
"eval_accuracy": 0.6840345500139314,
|
433 |
+
"eval_loss": 0.8481202721595764,
|
434 |
+
"eval_runtime": 64.028,
|
435 |
+
"eval_samples_per_second": 112.107,
|
436 |
+
"eval_steps_per_second": 3.514,
|
437 |
+
"step": 672
|
438 |
+
},
|
439 |
+
{
|
440 |
+
"epoch": 3.0,
|
441 |
+
"step": 672,
|
442 |
+
"total_flos": 6.668732964123095e+18,
|
443 |
+
"train_loss": 1.1477188680853163,
|
444 |
+
"train_runtime": 2102.1014,
|
445 |
+
"train_samples_per_second": 40.972,
|
446 |
+
"train_steps_per_second": 0.32
|
447 |
}
|
448 |
],
|
449 |
+
"max_steps": 672,
|
450 |
+
"num_train_epochs": 3,
|
451 |
+
"total_flos": 6.668732964123095e+18,
|
452 |
"trial_name": null,
|
453 |
"trial_params": null
|
454 |
}
|