Extra epochs training
Browse files
all_results.json
CHANGED
@@ -1,13 +1,13 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_accuracy": 0.
|
4 |
-
"eval_loss": 0.
|
5 |
-
"eval_runtime":
|
6 |
-
"eval_samples_per_second":
|
7 |
-
"eval_steps_per_second": 0.
|
8 |
-
"total_flos":
|
9 |
-
"train_loss": 0.
|
10 |
-
"train_runtime":
|
11 |
-
"train_samples_per_second":
|
12 |
-
"train_steps_per_second": 0.
|
13 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 6.98,
|
3 |
+
"eval_accuracy": 0.87375,
|
4 |
+
"eval_loss": 0.4391254186630249,
|
5 |
+
"eval_runtime": 43.2356,
|
6 |
+
"eval_samples_per_second": 37.007,
|
7 |
+
"eval_steps_per_second": 0.578,
|
8 |
+
"total_flos": 2.784658737197482e+18,
|
9 |
+
"train_loss": 0.45341705913601676,
|
10 |
+
"train_runtime": 2059.1149,
|
11 |
+
"train_samples_per_second": 41.015,
|
12 |
+
"train_steps_per_second": 0.16
|
13 |
}
|
eval_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"eval_accuracy": 0.
|
4 |
-
"eval_loss": 0.
|
5 |
-
"eval_runtime":
|
6 |
-
"eval_samples_per_second":
|
7 |
-
"eval_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 6.98,
|
3 |
+
"eval_accuracy": 0.87375,
|
4 |
+
"eval_loss": 0.4391254186630249,
|
5 |
+
"eval_runtime": 43.2356,
|
6 |
+
"eval_samples_per_second": 37.007,
|
7 |
+
"eval_steps_per_second": 0.578
|
8 |
}
|
runs/Apr22_01-57-25_25733e00f35f/events.out.tfevents.1682128678.25733e00f35f.544.12
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:12b3f2e0499f66f2356028ac5033c7ff56d5eaced70bb265cb43c2fa977a683c
|
3 |
+
size 18880
|
runs/Apr22_01-57-25_25733e00f35f/events.out.tfevents.1682131388.25733e00f35f.544.15
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1056d1d2a8358a8416e282fbbe0a5d21654f1017027460e8141275c83c610170
|
3 |
+
size 411
|
train_results.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"epoch":
|
3 |
-
"total_flos":
|
4 |
-
"train_loss": 0.
|
5 |
-
"train_runtime":
|
6 |
-
"train_samples_per_second":
|
7 |
-
"train_steps_per_second": 0.
|
8 |
}
|
|
|
1 |
{
|
2 |
+
"epoch": 6.98,
|
3 |
+
"total_flos": 2.784658737197482e+18,
|
4 |
+
"train_loss": 0.45341705913601676,
|
5 |
+
"train_runtime": 2059.1149,
|
6 |
+
"train_samples_per_second": 41.015,
|
7 |
+
"train_steps_per_second": 0.16
|
8 |
}
|
trainer_state.json
CHANGED
@@ -1,8 +1,8 @@
|
|
1 |
{
|
2 |
-
"best_metric": 0.
|
3 |
-
"best_model_checkpoint": "swinv2-tiny-patch4-window8-256-finetuned-thai/checkpoint-
|
4 |
-
"epoch":
|
5 |
-
"global_step":
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
@@ -120,89 +120,161 @@
|
|
120 |
},
|
121 |
{
|
122 |
"epoch": 3.19,
|
123 |
-
"learning_rate":
|
124 |
-
"loss": 1.
|
125 |
"step": 150
|
126 |
},
|
127 |
{
|
128 |
"epoch": 3.4,
|
129 |
-
"learning_rate":
|
130 |
-
"loss": 1.
|
131 |
"step": 160
|
132 |
},
|
133 |
{
|
134 |
"epoch": 3.61,
|
135 |
-
"learning_rate":
|
136 |
-
"loss": 1.
|
137 |
"step": 170
|
138 |
},
|
139 |
{
|
140 |
"epoch": 3.83,
|
141 |
-
"learning_rate":
|
142 |
-
"loss":
|
143 |
"step": 180
|
144 |
},
|
145 |
{
|
146 |
"epoch": 3.99,
|
147 |
-
"eval_accuracy": 0.
|
148 |
-
"eval_loss": 0.
|
149 |
-
"eval_runtime": 43.
|
150 |
-
"eval_samples_per_second": 36.
|
151 |
-
"eval_steps_per_second": 0.
|
152 |
"step": 188
|
153 |
},
|
154 |
{
|
155 |
"epoch": 4.04,
|
156 |
-
"learning_rate":
|
157 |
-
"loss":
|
158 |
"step": 190
|
159 |
},
|
160 |
{
|
161 |
"epoch": 4.25,
|
162 |
-
"learning_rate":
|
163 |
-
"loss":
|
164 |
"step": 200
|
165 |
},
|
166 |
{
|
167 |
"epoch": 4.46,
|
168 |
-
"learning_rate":
|
169 |
-
"loss":
|
170 |
"step": 210
|
171 |
},
|
172 |
{
|
173 |
"epoch": 4.67,
|
174 |
-
"learning_rate":
|
175 |
-
"loss":
|
176 |
"step": 220
|
177 |
},
|
178 |
{
|
179 |
"epoch": 4.88,
|
180 |
-
"learning_rate":
|
181 |
-
"loss":
|
182 |
"step": 230
|
183 |
},
|
184 |
{
|
185 |
"epoch": 4.99,
|
186 |
-
"eval_accuracy": 0.
|
187 |
-
"eval_loss": 0.
|
188 |
-
"eval_runtime":
|
189 |
-
"eval_samples_per_second": 36.
|
190 |
-
"eval_steps_per_second": 0.
|
191 |
"step": 235
|
192 |
},
|
193 |
{
|
194 |
-
"epoch":
|
195 |
-
"
|
196 |
-
"
|
197 |
-
"
|
198 |
-
|
199 |
-
|
200 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
201 |
}
|
202 |
],
|
203 |
-
"max_steps":
|
204 |
-
"num_train_epochs":
|
205 |
-
"total_flos":
|
206 |
"trial_name": null,
|
207 |
"trial_params": null
|
208 |
}
|
|
|
1 |
{
|
2 |
+
"best_metric": 0.87375,
|
3 |
+
"best_model_checkpoint": "swinv2-tiny-patch4-window8-256-finetuned-thai/checkpoint-329",
|
4 |
+
"epoch": 6.978835978835979,
|
5 |
+
"global_step": 329,
|
6 |
"is_hyper_param_search": false,
|
7 |
"is_local_process_zero": true,
|
8 |
"is_world_process_zero": true,
|
|
|
120 |
},
|
121 |
{
|
122 |
"epoch": 3.19,
|
123 |
+
"learning_rate": 3.0236486486486488e-05,
|
124 |
+
"loss": 1.0802,
|
125 |
"step": 150
|
126 |
},
|
127 |
{
|
128 |
"epoch": 3.4,
|
129 |
+
"learning_rate": 2.8547297297297298e-05,
|
130 |
+
"loss": 1.018,
|
131 |
"step": 160
|
132 |
},
|
133 |
{
|
134 |
"epoch": 3.61,
|
135 |
+
"learning_rate": 2.685810810810811e-05,
|
136 |
+
"loss": 1.0061,
|
137 |
"step": 170
|
138 |
},
|
139 |
{
|
140 |
"epoch": 3.83,
|
141 |
+
"learning_rate": 2.516891891891892e-05,
|
142 |
+
"loss": 0.9125,
|
143 |
"step": 180
|
144 |
},
|
145 |
{
|
146 |
"epoch": 3.99,
|
147 |
+
"eval_accuracy": 0.840625,
|
148 |
+
"eval_loss": 0.6059834361076355,
|
149 |
+
"eval_runtime": 43.4292,
|
150 |
+
"eval_samples_per_second": 36.842,
|
151 |
+
"eval_steps_per_second": 0.576,
|
152 |
"step": 188
|
153 |
},
|
154 |
{
|
155 |
"epoch": 4.04,
|
156 |
+
"learning_rate": 2.347972972972973e-05,
|
157 |
+
"loss": 0.8931,
|
158 |
"step": 190
|
159 |
},
|
160 |
{
|
161 |
"epoch": 4.25,
|
162 |
+
"learning_rate": 2.179054054054054e-05,
|
163 |
+
"loss": 0.8468,
|
164 |
"step": 200
|
165 |
},
|
166 |
{
|
167 |
"epoch": 4.46,
|
168 |
+
"learning_rate": 2.0101351351351353e-05,
|
169 |
+
"loss": 0.8092,
|
170 |
"step": 210
|
171 |
},
|
172 |
{
|
173 |
"epoch": 4.67,
|
174 |
+
"learning_rate": 1.8412162162162163e-05,
|
175 |
+
"loss": 0.7857,
|
176 |
"step": 220
|
177 |
},
|
178 |
{
|
179 |
"epoch": 4.88,
|
180 |
+
"learning_rate": 1.6722972972972976e-05,
|
181 |
+
"loss": 0.7527,
|
182 |
"step": 230
|
183 |
},
|
184 |
{
|
185 |
"epoch": 4.99,
|
186 |
+
"eval_accuracy": 0.8575,
|
187 |
+
"eval_loss": 0.5213606953620911,
|
188 |
+
"eval_runtime": 44.0663,
|
189 |
+
"eval_samples_per_second": 36.309,
|
190 |
+
"eval_steps_per_second": 0.567,
|
191 |
"step": 235
|
192 |
},
|
193 |
{
|
194 |
+
"epoch": 5.1,
|
195 |
+
"learning_rate": 1.5033783783783784e-05,
|
196 |
+
"loss": 0.7727,
|
197 |
+
"step": 240
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 5.31,
|
201 |
+
"learning_rate": 1.3344594594594596e-05,
|
202 |
+
"loss": 0.731,
|
203 |
+
"step": 250
|
204 |
+
},
|
205 |
+
{
|
206 |
+
"epoch": 5.52,
|
207 |
+
"learning_rate": 1.1655405405405405e-05,
|
208 |
+
"loss": 0.7291,
|
209 |
+
"step": 260
|
210 |
+
},
|
211 |
+
{
|
212 |
+
"epoch": 5.73,
|
213 |
+
"learning_rate": 9.966216216216217e-06,
|
214 |
+
"loss": 0.6753,
|
215 |
+
"step": 270
|
216 |
+
},
|
217 |
+
{
|
218 |
+
"epoch": 5.94,
|
219 |
+
"learning_rate": 8.277027027027027e-06,
|
220 |
+
"loss": 0.6852,
|
221 |
+
"step": 280
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 5.98,
|
225 |
+
"eval_accuracy": 0.865625,
|
226 |
+
"eval_loss": 0.45880234241485596,
|
227 |
+
"eval_runtime": 43.8008,
|
228 |
+
"eval_samples_per_second": 36.529,
|
229 |
+
"eval_steps_per_second": 0.571,
|
230 |
+
"step": 282
|
231 |
+
},
|
232 |
+
{
|
233 |
+
"epoch": 6.15,
|
234 |
+
"learning_rate": 6.587837837837837e-06,
|
235 |
+
"loss": 0.7279,
|
236 |
+
"step": 290
|
237 |
+
},
|
238 |
+
{
|
239 |
+
"epoch": 6.37,
|
240 |
+
"learning_rate": 4.898648648648649e-06,
|
241 |
+
"loss": 0.6565,
|
242 |
+
"step": 300
|
243 |
+
},
|
244 |
+
{
|
245 |
+
"epoch": 6.58,
|
246 |
+
"learning_rate": 3.2094594594594594e-06,
|
247 |
+
"loss": 0.6914,
|
248 |
+
"step": 310
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 6.79,
|
252 |
+
"learning_rate": 1.5202702702702704e-06,
|
253 |
+
"loss": 0.6233,
|
254 |
+
"step": 320
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 6.98,
|
258 |
+
"eval_accuracy": 0.87375,
|
259 |
+
"eval_loss": 0.4391254186630249,
|
260 |
+
"eval_runtime": 43.5089,
|
261 |
+
"eval_samples_per_second": 36.774,
|
262 |
+
"eval_steps_per_second": 0.575,
|
263 |
+
"step": 329
|
264 |
+
},
|
265 |
+
{
|
266 |
+
"epoch": 6.98,
|
267 |
+
"step": 329,
|
268 |
+
"total_flos": 2.784658737197482e+18,
|
269 |
+
"train_loss": 0.45341705913601676,
|
270 |
+
"train_runtime": 2059.1149,
|
271 |
+
"train_samples_per_second": 41.015,
|
272 |
+
"train_steps_per_second": 0.16
|
273 |
}
|
274 |
],
|
275 |
+
"max_steps": 329,
|
276 |
+
"num_train_epochs": 7,
|
277 |
+
"total_flos": 2.784658737197482e+18,
|
278 |
"trial_name": null,
|
279 |
"trial_params": null
|
280 |
}
|