End of training
Browse files- README.md +10 -8
- all_results.json +16 -0
- eval_results.json +11 -0
- logs/events.out.tfevents.1675059329.serv-3334.1016902.11 +3 -0
- train_results.json +8 -0
- trainer_state.json +297 -0
README.md
CHANGED
@@ -1,4 +1,6 @@
|
|
1 |
---
|
|
|
|
|
2 |
license: apache-2.0
|
3 |
tags:
|
4 |
- generated_from_trainer
|
@@ -14,7 +16,7 @@ model-index:
|
|
14 |
name: Text Classification
|
15 |
type: text-classification
|
16 |
dataset:
|
17 |
-
name:
|
18 |
type: glue
|
19 |
config: qqp
|
20 |
split: validation
|
@@ -22,10 +24,10 @@ model-index:
|
|
22 |
metrics:
|
23 |
- name: Accuracy
|
24 |
type: accuracy
|
25 |
-
value: 0.
|
26 |
- name: F1
|
27 |
type: f1
|
28 |
-
value: 0.
|
29 |
---
|
30 |
|
31 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -33,12 +35,12 @@ should probably proofread and complete it, then remove this comment. -->
|
|
33 |
|
34 |
# mobilebert_sa_GLUE_Experiment_logit_kd_qqp_256
|
35 |
|
36 |
-
This model is a fine-tuned version of [google/mobilebert-uncased](https://huggingface.co/google/mobilebert-uncased) on the
|
37 |
It achieves the following results on the evaluation set:
|
38 |
-
- Loss: 0.
|
39 |
-
- Accuracy: 0.
|
40 |
-
- F1: 0.
|
41 |
-
- Combined Score: 0.
|
42 |
|
43 |
## Model description
|
44 |
|
|
|
1 |
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
license: apache-2.0
|
5 |
tags:
|
6 |
- generated_from_trainer
|
|
|
16 |
name: Text Classification
|
17 |
type: text-classification
|
18 |
dataset:
|
19 |
+
name: GLUE QQP
|
20 |
type: glue
|
21 |
config: qqp
|
22 |
split: validation
|
|
|
24 |
metrics:
|
25 |
- name: Accuracy
|
26 |
type: accuracy
|
27 |
+
value: 0.794855305466238
|
28 |
- name: F1
|
29 |
type: f1
|
30 |
+
value: 0.7224044447419508
|
31 |
---
|
32 |
|
33 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
35 |
|
36 |
# mobilebert_sa_GLUE_Experiment_logit_kd_qqp_256
|
37 |
|
38 |
+
This model is a fine-tuned version of [google/mobilebert-uncased](https://huggingface.co/google/mobilebert-uncased) on the GLUE QQP dataset.
|
39 |
It achieves the following results on the evaluation set:
|
40 |
+
- Loss: 0.6619
|
41 |
+
- Accuracy: 0.7949
|
42 |
+
- F1: 0.7224
|
43 |
+
- Combined Score: 0.7586
|
44 |
|
45 |
## Model description
|
46 |
|
all_results.json
ADDED
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 16.0,
|
3 |
+
"eval_accuracy": 0.794855305466238,
|
4 |
+
"eval_combined_score": 0.7586298751040943,
|
5 |
+
"eval_f1": 0.7224044447419508,
|
6 |
+
"eval_loss": 0.6618520021438599,
|
7 |
+
"eval_runtime": 72.2937,
|
8 |
+
"eval_samples": 40430,
|
9 |
+
"eval_samples_per_second": 559.246,
|
10 |
+
"eval_steps_per_second": 4.371,
|
11 |
+
"train_loss": 0.5890673631687546,
|
12 |
+
"train_runtime": 25425.7527,
|
13 |
+
"train_samples": 363846,
|
14 |
+
"train_samples_per_second": 715.507,
|
15 |
+
"train_steps_per_second": 5.591
|
16 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 16.0,
|
3 |
+
"eval_accuracy": 0.794855305466238,
|
4 |
+
"eval_combined_score": 0.7586298751040943,
|
5 |
+
"eval_f1": 0.7224044447419508,
|
6 |
+
"eval_loss": 0.6618520021438599,
|
7 |
+
"eval_runtime": 72.2937,
|
8 |
+
"eval_samples": 40430,
|
9 |
+
"eval_samples_per_second": 559.246,
|
10 |
+
"eval_steps_per_second": 4.371
|
11 |
+
}
|
logs/events.out.tfevents.1675059329.serv-3334.1016902.11
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c6ec17e84c161c68c06ca1514d71d8f4121d67fa9e5b4fc8f8c9d092142530f7
|
3 |
+
size 475
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 16.0,
|
3 |
+
"train_loss": 0.5890673631687546,
|
4 |
+
"train_runtime": 25425.7527,
|
5 |
+
"train_samples": 363846,
|
6 |
+
"train_samples_per_second": 715.507,
|
7 |
+
"train_steps_per_second": 5.591
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,297 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.6618520021438599,
|
3 |
+
"best_model_checkpoint": "mobilebert_sa_GLUE_Experiment_logit_kd_qqp_256/checkpoint-31273",
|
4 |
+
"epoch": 16.0,
|
5 |
+
"global_step": 45488,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 1.0,
|
12 |
+
"learning_rate": 4.9e-05,
|
13 |
+
"loss": 0.9454,
|
14 |
+
"step": 2843
|
15 |
+
},
|
16 |
+
{
|
17 |
+
"epoch": 1.0,
|
18 |
+
"eval_accuracy": 0.755552807321296,
|
19 |
+
"eval_combined_score": 0.7059041209574377,
|
20 |
+
"eval_f1": 0.6562554345935794,
|
21 |
+
"eval_loss": 0.8257145881652832,
|
22 |
+
"eval_runtime": 72.5003,
|
23 |
+
"eval_samples_per_second": 557.653,
|
24 |
+
"eval_steps_per_second": 4.359,
|
25 |
+
"step": 2843
|
26 |
+
},
|
27 |
+
{
|
28 |
+
"epoch": 2.0,
|
29 |
+
"learning_rate": 4.8e-05,
|
30 |
+
"loss": 0.8165,
|
31 |
+
"step": 5686
|
32 |
+
},
|
33 |
+
{
|
34 |
+
"epoch": 2.0,
|
35 |
+
"eval_accuracy": 0.7589661142715806,
|
36 |
+
"eval_combined_score": 0.6819519426913909,
|
37 |
+
"eval_f1": 0.6049377711112012,
|
38 |
+
"eval_loss": 0.7681939005851746,
|
39 |
+
"eval_runtime": 72.4829,
|
40 |
+
"eval_samples_per_second": 557.787,
|
41 |
+
"eval_steps_per_second": 4.36,
|
42 |
+
"step": 5686
|
43 |
+
},
|
44 |
+
{
|
45 |
+
"epoch": 3.0,
|
46 |
+
"learning_rate": 4.7e-05,
|
47 |
+
"loss": 0.7741,
|
48 |
+
"step": 8529
|
49 |
+
},
|
50 |
+
{
|
51 |
+
"epoch": 3.0,
|
52 |
+
"eval_accuracy": 0.7638139995053178,
|
53 |
+
"eval_combined_score": 0.6920433977325372,
|
54 |
+
"eval_f1": 0.6202727959597566,
|
55 |
+
"eval_loss": 0.7514052987098694,
|
56 |
+
"eval_runtime": 72.7836,
|
57 |
+
"eval_samples_per_second": 555.482,
|
58 |
+
"eval_steps_per_second": 4.342,
|
59 |
+
"step": 8529
|
60 |
+
},
|
61 |
+
{
|
62 |
+
"epoch": 4.0,
|
63 |
+
"learning_rate": 4.600000000000001e-05,
|
64 |
+
"loss": 0.7325,
|
65 |
+
"step": 11372
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 4.0,
|
69 |
+
"eval_accuracy": 0.7674746475389562,
|
70 |
+
"eval_combined_score": 0.698115556996189,
|
71 |
+
"eval_f1": 0.6287564664534218,
|
72 |
+
"eval_loss": 0.7353646755218506,
|
73 |
+
"eval_runtime": 72.4082,
|
74 |
+
"eval_samples_per_second": 558.362,
|
75 |
+
"eval_steps_per_second": 4.364,
|
76 |
+
"step": 11372
|
77 |
+
},
|
78 |
+
{
|
79 |
+
"epoch": 5.0,
|
80 |
+
"learning_rate": 4.5e-05,
|
81 |
+
"loss": 0.6849,
|
82 |
+
"step": 14215
|
83 |
+
},
|
84 |
+
{
|
85 |
+
"epoch": 5.0,
|
86 |
+
"eval_accuracy": 0.7785060598565422,
|
87 |
+
"eval_combined_score": 0.7301661583702679,
|
88 |
+
"eval_f1": 0.6818262568839936,
|
89 |
+
"eval_loss": 0.7063168883323669,
|
90 |
+
"eval_runtime": 72.761,
|
91 |
+
"eval_samples_per_second": 555.655,
|
92 |
+
"eval_steps_per_second": 4.343,
|
93 |
+
"step": 14215
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 6.0,
|
97 |
+
"learning_rate": 4.4000000000000006e-05,
|
98 |
+
"loss": 0.6399,
|
99 |
+
"step": 17058
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 6.0,
|
103 |
+
"eval_accuracy": 0.78276032649023,
|
104 |
+
"eval_combined_score": 0.7351824040418427,
|
105 |
+
"eval_f1": 0.6876044815934554,
|
106 |
+
"eval_loss": 0.6906189918518066,
|
107 |
+
"eval_runtime": 72.4222,
|
108 |
+
"eval_samples_per_second": 558.254,
|
109 |
+
"eval_steps_per_second": 4.363,
|
110 |
+
"step": 17058
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 7.0,
|
114 |
+
"learning_rate": 4.3e-05,
|
115 |
+
"loss": 0.6005,
|
116 |
+
"step": 19901
|
117 |
+
},
|
118 |
+
{
|
119 |
+
"epoch": 7.0,
|
120 |
+
"eval_accuracy": 0.7867919861488993,
|
121 |
+
"eval_combined_score": 0.7430436592992455,
|
122 |
+
"eval_f1": 0.6992953324495919,
|
123 |
+
"eval_loss": 0.677066445350647,
|
124 |
+
"eval_runtime": 72.5064,
|
125 |
+
"eval_samples_per_second": 557.606,
|
126 |
+
"eval_steps_per_second": 4.358,
|
127 |
+
"step": 19901
|
128 |
+
},
|
129 |
+
{
|
130 |
+
"epoch": 8.0,
|
131 |
+
"learning_rate": 4.2e-05,
|
132 |
+
"loss": 0.5666,
|
133 |
+
"step": 22744
|
134 |
+
},
|
135 |
+
{
|
136 |
+
"epoch": 8.0,
|
137 |
+
"eval_accuracy": 0.7897106109324759,
|
138 |
+
"eval_combined_score": 0.7517336301581778,
|
139 |
+
"eval_f1": 0.71375664938388,
|
140 |
+
"eval_loss": 0.6808561086654663,
|
141 |
+
"eval_runtime": 72.3704,
|
142 |
+
"eval_samples_per_second": 558.654,
|
143 |
+
"eval_steps_per_second": 4.366,
|
144 |
+
"step": 22744
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 9.0,
|
148 |
+
"learning_rate": 4.1e-05,
|
149 |
+
"loss": 0.5365,
|
150 |
+
"step": 25587
|
151 |
+
},
|
152 |
+
{
|
153 |
+
"epoch": 9.0,
|
154 |
+
"eval_accuracy": 0.7885975760573831,
|
155 |
+
"eval_combined_score": 0.7403265307777458,
|
156 |
+
"eval_f1": 0.6920554854981085,
|
157 |
+
"eval_loss": 0.6807268857955933,
|
158 |
+
"eval_runtime": 72.5072,
|
159 |
+
"eval_samples_per_second": 557.6,
|
160 |
+
"eval_steps_per_second": 4.358,
|
161 |
+
"step": 25587
|
162 |
+
},
|
163 |
+
{
|
164 |
+
"epoch": 10.0,
|
165 |
+
"learning_rate": 4e-05,
|
166 |
+
"loss": 0.5097,
|
167 |
+
"step": 28430
|
168 |
+
},
|
169 |
+
{
|
170 |
+
"epoch": 10.0,
|
171 |
+
"eval_accuracy": 0.7872619342072719,
|
172 |
+
"eval_combined_score": 0.7566418293977779,
|
173 |
+
"eval_f1": 0.726021724588284,
|
174 |
+
"eval_loss": 0.6826764345169067,
|
175 |
+
"eval_runtime": 72.5186,
|
176 |
+
"eval_samples_per_second": 557.512,
|
177 |
+
"eval_steps_per_second": 4.358,
|
178 |
+
"step": 28430
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"epoch": 11.0,
|
182 |
+
"learning_rate": 3.9000000000000006e-05,
|
183 |
+
"loss": 0.4856,
|
184 |
+
"step": 31273
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 11.0,
|
188 |
+
"eval_accuracy": 0.794855305466238,
|
189 |
+
"eval_combined_score": 0.7586298751040943,
|
190 |
+
"eval_f1": 0.7224044447419508,
|
191 |
+
"eval_loss": 0.6618520021438599,
|
192 |
+
"eval_runtime": 72.4772,
|
193 |
+
"eval_samples_per_second": 557.83,
|
194 |
+
"eval_steps_per_second": 4.36,
|
195 |
+
"step": 31273
|
196 |
+
},
|
197 |
+
{
|
198 |
+
"epoch": 12.0,
|
199 |
+
"learning_rate": 3.8e-05,
|
200 |
+
"loss": 0.4653,
|
201 |
+
"step": 34116
|
202 |
+
},
|
203 |
+
{
|
204 |
+
"epoch": 12.0,
|
205 |
+
"eval_accuracy": 0.7947563690328964,
|
206 |
+
"eval_combined_score": 0.7572282047839796,
|
207 |
+
"eval_f1": 0.7197000405350629,
|
208 |
+
"eval_loss": 0.7001789212226868,
|
209 |
+
"eval_runtime": 72.5909,
|
210 |
+
"eval_samples_per_second": 556.957,
|
211 |
+
"eval_steps_per_second": 4.353,
|
212 |
+
"step": 34116
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 13.0,
|
216 |
+
"learning_rate": 3.7e-05,
|
217 |
+
"loss": 0.4438,
|
218 |
+
"step": 36959
|
219 |
+
},
|
220 |
+
{
|
221 |
+
"epoch": 13.0,
|
222 |
+
"eval_accuracy": 0.7965372248330448,
|
223 |
+
"eval_combined_score": 0.7584372396008563,
|
224 |
+
"eval_f1": 0.7203372543686679,
|
225 |
+
"eval_loss": 0.689961850643158,
|
226 |
+
"eval_runtime": 72.5363,
|
227 |
+
"eval_samples_per_second": 557.376,
|
228 |
+
"eval_steps_per_second": 4.356,
|
229 |
+
"step": 36959
|
230 |
+
},
|
231 |
+
{
|
232 |
+
"epoch": 14.0,
|
233 |
+
"learning_rate": 3.6e-05,
|
234 |
+
"loss": 0.4255,
|
235 |
+
"step": 39802
|
236 |
+
},
|
237 |
+
{
|
238 |
+
"epoch": 14.0,
|
239 |
+
"eval_accuracy": 0.7980707395498392,
|
240 |
+
"eval_combined_score": 0.7632222209859922,
|
241 |
+
"eval_f1": 0.7283737024221453,
|
242 |
+
"eval_loss": 0.6847425699234009,
|
243 |
+
"eval_runtime": 72.6156,
|
244 |
+
"eval_samples_per_second": 556.767,
|
245 |
+
"eval_steps_per_second": 4.352,
|
246 |
+
"step": 39802
|
247 |
+
},
|
248 |
+
{
|
249 |
+
"epoch": 15.0,
|
250 |
+
"learning_rate": 3.5e-05,
|
251 |
+
"loss": 0.4072,
|
252 |
+
"step": 42645
|
253 |
+
},
|
254 |
+
{
|
255 |
+
"epoch": 15.0,
|
256 |
+
"eval_accuracy": 0.7917388078159783,
|
257 |
+
"eval_combined_score": 0.7626668474313707,
|
258 |
+
"eval_f1": 0.7335948870467632,
|
259 |
+
"eval_loss": 0.6892824769020081,
|
260 |
+
"eval_runtime": 72.7761,
|
261 |
+
"eval_samples_per_second": 555.539,
|
262 |
+
"eval_steps_per_second": 4.342,
|
263 |
+
"step": 42645
|
264 |
+
},
|
265 |
+
{
|
266 |
+
"epoch": 16.0,
|
267 |
+
"learning_rate": 3.4000000000000007e-05,
|
268 |
+
"loss": 0.391,
|
269 |
+
"step": 45488
|
270 |
+
},
|
271 |
+
{
|
272 |
+
"epoch": 16.0,
|
273 |
+
"eval_accuracy": 0.7957209992579768,
|
274 |
+
"eval_combined_score": 0.7628521627538454,
|
275 |
+
"eval_f1": 0.729983326249714,
|
276 |
+
"eval_loss": 0.7136414647102356,
|
277 |
+
"eval_runtime": 72.54,
|
278 |
+
"eval_samples_per_second": 557.348,
|
279 |
+
"eval_steps_per_second": 4.356,
|
280 |
+
"step": 45488
|
281 |
+
},
|
282 |
+
{
|
283 |
+
"epoch": 16.0,
|
284 |
+
"step": 45488,
|
285 |
+
"total_flos": 1.5334594896828826e+17,
|
286 |
+
"train_loss": 0.5890673631687546,
|
287 |
+
"train_runtime": 25425.7527,
|
288 |
+
"train_samples_per_second": 715.507,
|
289 |
+
"train_steps_per_second": 5.591
|
290 |
+
}
|
291 |
+
],
|
292 |
+
"max_steps": 142150,
|
293 |
+
"num_train_epochs": 50,
|
294 |
+
"total_flos": 1.5334594896828826e+17,
|
295 |
+
"trial_name": null,
|
296 |
+
"trial_params": null
|
297 |
+
}
|