ShengdingHu commited on
Commit
6d12a1b
1 Parent(s): 16da2dd

Training in progress, step 200

Browse files
all_results.json CHANGED
@@ -1,18 +1,18 @@
1
  {
2
  "epoch": 3.0,
3
- "eval_accuracy": 97.5,
4
- "eval_average_metrics": 97.5,
5
- "eval_loss": 0.03714486584067345,
6
- "eval_runtime": 5.7314,
7
- "eval_samples_per_second": 174.477,
8
- "test_accuracy": 93.34862385321101,
9
- "test_average_metrics": 93.34862385321101,
10
- "test_loss": 0.07569558918476105,
11
- "test_runtime": 5.6934,
12
- "test_samples_per_second": 153.159,
13
- "train_loss": 0.058686515289424354,
14
- "train_runtime": 1966.0928,
15
  "train_samples": 66349,
16
- "train_samples_per_second": 101.24,
17
- "train_steps_per_second": 3.165
18
  }
 
1
  {
2
  "epoch": 3.0,
3
+ "eval_accuracy": 60.099999999999994,
4
+ "eval_average_metrics": 60.099999999999994,
5
+ "eval_loss": 0.2783910036087036,
6
+ "eval_runtime": 3.2954,
7
+ "eval_samples_per_second": 303.458,
8
+ "test_accuracy": 58.142201834862384,
9
+ "test_average_metrics": 58.142201834862384,
10
+ "test_loss": 0.2752912938594818,
11
+ "test_runtime": 3.1391,
12
+ "test_samples_per_second": 277.783,
13
+ "train_loss": 0.5804944922785328,
14
+ "train_runtime": 1254.7674,
15
  "train_samples": 66349,
16
+ "train_samples_per_second": 158.633,
17
+ "train_steps_per_second": 4.959
18
  }
eval_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 3.0,
3
- "eval_accuracy": 97.5,
4
- "eval_average_metrics": 97.5,
5
- "eval_loss": 0.03714486584067345,
6
- "eval_runtime": 5.7314,
7
- "eval_samples_per_second": 174.477
8
  }
 
1
  {
2
  "epoch": 3.0,
3
+ "eval_accuracy": 60.099999999999994,
4
+ "eval_average_metrics": 60.099999999999994,
5
+ "eval_loss": 0.2783910036087036,
6
+ "eval_runtime": 3.2954,
7
+ "eval_samples_per_second": 303.458
8
  }
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:313a8aa2ad882fe8925fb538257ec103af71608407bbe84799eb3b904abaa5aa
3
  size 335851
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca578edc26c679322afc4e794b748861809ce462f741454b4a5d1327e399320f
3
  size 335851
runs/Feb04_09-56-16_node1/events.out.tfevents.1643940000.node1 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:a091178aa1969a4b4a220994b83ce23c3e06f9259d106e3b6071afc59671639a
3
- size 15536
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19b60bae55358e79ee8cc62539e971324e17914693622f66af818a2d245fce4b
3
+ size 16534
runs/Feb04_09-56-16_node1/events.out.tfevents.1643941260.node1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3fbc496fbb0b017946d09a318fa8b81191c585c2053d8ff1f9d5474c4e9c1776
3
+ size 684
runs/Feb04_10-46-17_node1/1643942904.4796145/events.out.tfevents.1643942904.node1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e663a68a3ea6d4ba8bf840e42067622d741437998789d2407af56650d036e12f
3
+ size 5026
runs/Feb04_10-46-17_node1/events.out.tfevents.1643942904.node1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1f95558cca914d35629334a365c56bd84a73822d96a0efeb17cb5e14af2e46bb
3
+ size 4312
test_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 3.0,
3
- "test_accuracy": 93.34862385321101,
4
- "test_average_metrics": 93.34862385321101,
5
- "test_loss": 0.07569558918476105,
6
- "test_runtime": 5.6934,
7
- "test_samples_per_second": 153.159
8
  }
 
1
  {
2
  "epoch": 3.0,
3
+ "test_accuracy": 58.142201834862384,
4
+ "test_average_metrics": 58.142201834862384,
5
+ "test_loss": 0.2752912938594818,
6
+ "test_runtime": 3.1391,
7
+ "test_samples_per_second": 277.783
8
  }
tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff
 
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 3.0,
3
- "train_loss": 0.058686515289424354,
4
- "train_runtime": 1966.0928,
5
  "train_samples": 66349,
6
- "train_samples_per_second": 101.24,
7
- "train_steps_per_second": 3.165
8
  }
 
1
  {
2
  "epoch": 3.0,
3
+ "train_loss": 0.5804944922785328,
4
+ "train_runtime": 1254.7674,
5
  "train_samples": 66349,
6
+ "train_samples_per_second": 158.633,
7
+ "train_steps_per_second": 4.959
8
  }
trainer_state.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
- "best_metric": 97.5,
3
- "best_model_checkpoint": "outputs/bitfit/t5-base/sst2/checkpoint-3000",
4
  "epoch": 3.0,
5
  "global_step": 6222,
6
  "is_hyper_param_search": false,
@@ -9,368 +9,368 @@
9
  "log_history": [
10
  {
11
  "epoch": 0.1,
12
- "eval_accuracy": 97.3,
13
- "eval_average_metrics": 97.3,
14
- "eval_loss": 0.04369654133915901,
15
- "eval_runtime": 5.1897,
16
- "eval_samples_per_second": 192.69,
17
  "step": 200
18
  },
19
  {
20
  "epoch": 0.19,
21
- "eval_accuracy": 97.3,
22
- "eval_average_metrics": 97.3,
23
- "eval_loss": 0.042765092104673386,
24
- "eval_runtime": 5.6341,
25
- "eval_samples_per_second": 177.492,
26
  "step": 400
27
  },
28
  {
29
  "epoch": 0.24,
30
  "learning_rate": 0.00027589199614271937,
31
- "loss": 0.1281,
32
  "step": 500
33
  },
34
  {
35
  "epoch": 0.29,
36
- "eval_accuracy": 97.2,
37
- "eval_average_metrics": 97.2,
38
- "eval_loss": 0.037586040794849396,
39
- "eval_runtime": 6.2353,
40
- "eval_samples_per_second": 160.377,
41
  "step": 600
42
  },
43
  {
44
  "epoch": 0.39,
45
- "eval_accuracy": 96.7,
46
- "eval_average_metrics": 96.7,
47
- "eval_loss": 0.04588627442717552,
48
- "eval_runtime": 5.966,
49
- "eval_samples_per_second": 167.617,
50
  "step": 800
51
  },
52
  {
53
  "epoch": 0.48,
54
  "learning_rate": 0.0002517839922854387,
55
- "loss": 0.0649,
56
  "step": 1000
57
  },
58
  {
59
  "epoch": 0.48,
60
- "eval_accuracy": 97.1,
61
- "eval_average_metrics": 97.1,
62
- "eval_loss": 0.040396977216005325,
63
- "eval_runtime": 5.3919,
64
- "eval_samples_per_second": 185.463,
65
  "step": 1000
66
  },
67
  {
68
  "epoch": 0.58,
69
- "eval_accuracy": 97.2,
70
- "eval_average_metrics": 97.2,
71
- "eval_loss": 0.03829416632652283,
72
- "eval_runtime": 5.3561,
73
- "eval_samples_per_second": 186.702,
74
  "step": 1200
75
  },
76
  {
77
  "epoch": 0.68,
78
- "eval_accuracy": 97.3,
79
- "eval_average_metrics": 97.3,
80
- "eval_loss": 0.039076462388038635,
81
- "eval_runtime": 5.9241,
82
- "eval_samples_per_second": 168.803,
83
  "step": 1400
84
  },
85
  {
86
  "epoch": 0.72,
87
  "learning_rate": 0.00022767598842815813,
88
- "loss": 0.0578,
89
  "step": 1500
90
  },
91
  {
92
  "epoch": 0.77,
93
- "eval_accuracy": 97.1,
94
- "eval_average_metrics": 97.1,
95
- "eval_loss": 0.04230912774801254,
96
- "eval_runtime": 5.153,
97
- "eval_samples_per_second": 194.06,
98
  "step": 1600
99
  },
100
  {
101
  "epoch": 0.87,
102
- "eval_accuracy": 97.3,
103
- "eval_average_metrics": 97.3,
104
- "eval_loss": 0.03823951631784439,
105
- "eval_runtime": 5.1279,
106
- "eval_samples_per_second": 195.011,
107
  "step": 1800
108
  },
109
  {
110
  "epoch": 0.96,
111
  "learning_rate": 0.00020356798457087753,
112
- "loss": 0.0582,
113
  "step": 2000
114
  },
115
  {
116
  "epoch": 0.96,
117
- "eval_accuracy": 97.1,
118
- "eval_average_metrics": 97.1,
119
- "eval_loss": 0.0395108200609684,
120
- "eval_runtime": 5.8042,
121
- "eval_samples_per_second": 172.29,
122
  "step": 2000
123
  },
124
  {
125
  "epoch": 1.06,
126
- "eval_accuracy": 97.1,
127
- "eval_average_metrics": 97.1,
128
- "eval_loss": 0.04088559374213219,
129
- "eval_runtime": 5.4894,
130
- "eval_samples_per_second": 182.168,
131
  "step": 2200
132
  },
133
  {
134
  "epoch": 1.16,
135
- "eval_accuracy": 97.2,
136
- "eval_average_metrics": 97.2,
137
- "eval_loss": 0.038739945739507675,
138
- "eval_runtime": 4.6293,
139
- "eval_samples_per_second": 216.018,
140
  "step": 2400
141
  },
142
  {
143
  "epoch": 1.21,
144
  "learning_rate": 0.0001794599807135969,
145
- "loss": 0.054,
146
  "step": 2500
147
  },
148
  {
149
  "epoch": 1.25,
150
- "eval_accuracy": 97.3,
151
- "eval_average_metrics": 97.3,
152
- "eval_loss": 0.04564524069428444,
153
- "eval_runtime": 5.2048,
154
- "eval_samples_per_second": 192.132,
155
  "step": 2600
156
  },
157
  {
158
  "epoch": 1.35,
159
- "eval_accuracy": 97.3,
160
- "eval_average_metrics": 97.3,
161
- "eval_loss": 0.03781759738922119,
162
- "eval_runtime": 5.6194,
163
- "eval_samples_per_second": 177.956,
164
  "step": 2800
165
  },
166
  {
167
  "epoch": 1.45,
168
  "learning_rate": 0.00015535197685631627,
169
- "loss": 0.0533,
170
  "step": 3000
171
  },
172
  {
173
  "epoch": 1.45,
174
- "eval_accuracy": 97.5,
175
- "eval_average_metrics": 97.5,
176
- "eval_loss": 0.03714486584067345,
177
- "eval_runtime": 5.328,
178
- "eval_samples_per_second": 187.688,
179
  "step": 3000
180
  },
181
  {
182
  "epoch": 1.54,
183
- "eval_accuracy": 97.5,
184
- "eval_average_metrics": 97.5,
185
- "eval_loss": 0.0382937453687191,
186
- "eval_runtime": 4.0581,
187
- "eval_samples_per_second": 246.419,
188
  "step": 3200
189
  },
190
  {
191
  "epoch": 1.64,
192
- "eval_accuracy": 97.5,
193
- "eval_average_metrics": 97.5,
194
- "eval_loss": 0.037162039428949356,
195
- "eval_runtime": 5.3134,
196
- "eval_samples_per_second": 188.202,
197
  "step": 3400
198
  },
199
  {
200
  "epoch": 1.69,
201
  "learning_rate": 0.00013124397299903566,
202
- "loss": 0.0539,
203
  "step": 3500
204
  },
205
  {
206
  "epoch": 1.74,
207
- "eval_accuracy": 97.5,
208
- "eval_average_metrics": 97.5,
209
- "eval_loss": 0.03954707458615303,
210
- "eval_runtime": 5.9646,
211
- "eval_samples_per_second": 167.655,
212
  "step": 3600
213
  },
214
  {
215
  "epoch": 1.83,
216
- "eval_accuracy": 97.2,
217
- "eval_average_metrics": 97.2,
218
- "eval_loss": 0.037706729024648666,
219
- "eval_runtime": 4.5321,
220
- "eval_samples_per_second": 220.649,
221
  "step": 3800
222
  },
223
  {
224
  "epoch": 1.93,
225
  "learning_rate": 0.00010713596914175504,
226
- "loss": 0.0531,
227
  "step": 4000
228
  },
229
  {
230
  "epoch": 1.93,
231
- "eval_accuracy": 97.2,
232
- "eval_average_metrics": 97.2,
233
- "eval_loss": 0.03779396042227745,
234
- "eval_runtime": 6.6618,
235
- "eval_samples_per_second": 150.108,
236
  "step": 4000
237
  },
238
  {
239
  "epoch": 2.03,
240
- "eval_accuracy": 97.3,
241
- "eval_average_metrics": 97.3,
242
- "eval_loss": 0.04000015929341316,
243
- "eval_runtime": 6.7004,
244
- "eval_samples_per_second": 149.245,
245
  "step": 4200
246
  },
247
  {
248
  "epoch": 2.12,
249
- "eval_accuracy": 97.1,
250
- "eval_average_metrics": 97.1,
251
- "eval_loss": 0.0396127812564373,
252
- "eval_runtime": 5.9828,
253
- "eval_samples_per_second": 167.145,
254
  "step": 4400
255
  },
256
  {
257
  "epoch": 2.17,
258
  "learning_rate": 8.302796528447444e-05,
259
- "loss": 0.049,
260
  "step": 4500
261
  },
262
  {
263
  "epoch": 2.22,
264
- "eval_accuracy": 97.2,
265
- "eval_average_metrics": 97.2,
266
- "eval_loss": 0.0382530614733696,
267
- "eval_runtime": 5.4179,
268
- "eval_samples_per_second": 184.574,
269
  "step": 4600
270
  },
271
  {
272
  "epoch": 2.31,
273
- "eval_accuracy": 97.0,
274
- "eval_average_metrics": 97.0,
275
- "eval_loss": 0.04122977331280708,
276
- "eval_runtime": 4.9766,
277
- "eval_samples_per_second": 200.94,
278
  "step": 4800
279
  },
280
  {
281
  "epoch": 2.41,
282
  "learning_rate": 5.891996142719383e-05,
283
- "loss": 0.0484,
284
  "step": 5000
285
  },
286
  {
287
  "epoch": 2.41,
288
- "eval_accuracy": 97.3,
289
- "eval_average_metrics": 97.3,
290
- "eval_loss": 0.03976716473698616,
291
- "eval_runtime": 5.1121,
292
- "eval_samples_per_second": 195.615,
293
  "step": 5000
294
  },
295
  {
296
  "epoch": 2.51,
297
- "eval_accuracy": 97.2,
298
- "eval_average_metrics": 97.2,
299
- "eval_loss": 0.03887654095888138,
300
- "eval_runtime": 5.3028,
301
- "eval_samples_per_second": 188.578,
302
  "step": 5200
303
  },
304
  {
305
  "epoch": 2.6,
306
- "eval_accuracy": 97.3,
307
- "eval_average_metrics": 97.3,
308
- "eval_loss": 0.03931749612092972,
309
- "eval_runtime": 6.019,
310
- "eval_samples_per_second": 166.14,
311
  "step": 5400
312
  },
313
  {
314
  "epoch": 2.65,
315
  "learning_rate": 3.481195756991321e-05,
316
- "loss": 0.0478,
317
  "step": 5500
318
  },
319
  {
320
  "epoch": 2.7,
321
- "eval_accuracy": 97.1,
322
- "eval_average_metrics": 97.1,
323
- "eval_loss": 0.03971054032444954,
324
- "eval_runtime": 5.7951,
325
- "eval_samples_per_second": 172.558,
326
  "step": 5600
327
  },
328
  {
329
  "epoch": 2.8,
330
- "eval_accuracy": 97.1,
331
- "eval_average_metrics": 97.1,
332
- "eval_loss": 0.04016176983714104,
333
- "eval_runtime": 5.5419,
334
- "eval_samples_per_second": 180.443,
335
  "step": 5800
336
  },
337
  {
338
  "epoch": 2.89,
339
  "learning_rate": 1.0703953712632592e-05,
340
- "loss": 0.0444,
341
  "step": 6000
342
  },
343
  {
344
  "epoch": 2.89,
345
- "eval_accuracy": 97.0,
346
- "eval_average_metrics": 97.0,
347
- "eval_loss": 0.04050704091787338,
348
- "eval_runtime": 6.1769,
349
- "eval_samples_per_second": 161.894,
350
  "step": 6000
351
  },
352
  {
353
  "epoch": 2.99,
354
- "eval_accuracy": 97.2,
355
- "eval_average_metrics": 97.2,
356
- "eval_loss": 0.04039543867111206,
357
- "eval_runtime": 5.4148,
358
- "eval_samples_per_second": 184.68,
359
  "step": 6200
360
  },
361
  {
362
  "epoch": 3.0,
363
  "step": 6222,
364
- "total_flos": 1.3280955617597184e+16,
365
- "train_loss": 0.058686515289424354,
366
- "train_runtime": 1966.0928,
367
- "train_samples_per_second": 101.24,
368
- "train_steps_per_second": 3.165
369
  }
370
  ],
371
  "max_steps": 6222,
372
  "num_train_epochs": 3,
373
- "total_flos": 1.3280955617597184e+16,
374
  "trial_name": null,
375
  "trial_params": null
376
  }
 
1
  {
2
+ "best_metric": 60.099999999999994,
3
+ "best_model_checkpoint": "outputs/soft_prompt/t5-base/sst2/checkpoint-5800",
4
  "epoch": 3.0,
5
  "global_step": 6222,
6
  "is_hyper_param_search": false,
 
9
  "log_history": [
10
  {
11
  "epoch": 0.1,
12
+ "eval_accuracy": 0.0,
13
+ "eval_average_metrics": 0.0,
14
+ "eval_loss": 1.9322336912155151,
15
+ "eval_runtime": 9.7016,
16
+ "eval_samples_per_second": 103.076,
17
  "step": 200
18
  },
19
  {
20
  "epoch": 0.19,
21
+ "eval_accuracy": 51.0,
22
+ "eval_average_metrics": 51.0,
23
+ "eval_loss": 0.7396745681762695,
24
+ "eval_runtime": 3.3022,
25
+ "eval_samples_per_second": 302.832,
26
  "step": 400
27
  },
28
  {
29
  "epoch": 0.24,
30
  "learning_rate": 0.00027589199614271937,
31
+ "loss": 2.5059,
32
  "step": 500
33
  },
34
  {
35
  "epoch": 0.29,
36
+ "eval_accuracy": 55.2,
37
+ "eval_average_metrics": 55.2,
38
+ "eval_loss": 0.5560445785522461,
39
+ "eval_runtime": 3.2155,
40
+ "eval_samples_per_second": 310.993,
41
  "step": 600
42
  },
43
  {
44
  "epoch": 0.39,
45
+ "eval_accuracy": 55.300000000000004,
46
+ "eval_average_metrics": 55.300000000000004,
47
+ "eval_loss": 0.463209867477417,
48
+ "eval_runtime": 3.1314,
49
+ "eval_samples_per_second": 319.345,
50
  "step": 800
51
  },
52
  {
53
  "epoch": 0.48,
54
  "learning_rate": 0.0002517839922854387,
55
+ "loss": 0.7329,
56
  "step": 1000
57
  },
58
  {
59
  "epoch": 0.48,
60
+ "eval_accuracy": 55.300000000000004,
61
+ "eval_average_metrics": 55.300000000000004,
62
+ "eval_loss": 0.4116092920303345,
63
+ "eval_runtime": 3.3786,
64
+ "eval_samples_per_second": 295.977,
65
  "step": 1000
66
  },
67
  {
68
  "epoch": 0.58,
69
+ "eval_accuracy": 55.300000000000004,
70
+ "eval_average_metrics": 55.300000000000004,
71
+ "eval_loss": 0.38179224729537964,
72
+ "eval_runtime": 3.2857,
73
+ "eval_samples_per_second": 304.351,
74
  "step": 1200
75
  },
76
  {
77
  "epoch": 0.68,
78
+ "eval_accuracy": 55.300000000000004,
79
+ "eval_average_metrics": 55.300000000000004,
80
+ "eval_loss": 0.35022690892219543,
81
+ "eval_runtime": 2.8293,
82
+ "eval_samples_per_second": 353.443,
83
  "step": 1400
84
  },
85
  {
86
  "epoch": 0.72,
87
  "learning_rate": 0.00022767598842815813,
88
+ "loss": 0.4828,
89
  "step": 1500
90
  },
91
  {
92
  "epoch": 0.77,
93
+ "eval_accuracy": 55.300000000000004,
94
+ "eval_average_metrics": 55.300000000000004,
95
+ "eval_loss": 0.3595990538597107,
96
+ "eval_runtime": 3.2567,
97
+ "eval_samples_per_second": 307.055,
98
  "step": 1600
99
  },
100
  {
101
  "epoch": 0.87,
102
+ "eval_accuracy": 55.300000000000004,
103
+ "eval_average_metrics": 55.300000000000004,
104
+ "eval_loss": 0.3511555790901184,
105
+ "eval_runtime": 3.0816,
106
+ "eval_samples_per_second": 324.506,
107
  "step": 1800
108
  },
109
  {
110
  "epoch": 0.96,
111
  "learning_rate": 0.00020356798457087753,
112
+ "loss": 0.421,
113
  "step": 2000
114
  },
115
  {
116
  "epoch": 0.96,
117
+ "eval_accuracy": 55.300000000000004,
118
+ "eval_average_metrics": 55.300000000000004,
119
+ "eval_loss": 0.3314475417137146,
120
+ "eval_runtime": 2.522,
121
+ "eval_samples_per_second": 396.513,
122
  "step": 2000
123
  },
124
  {
125
  "epoch": 1.06,
126
+ "eval_accuracy": 55.300000000000004,
127
+ "eval_average_metrics": 55.300000000000004,
128
+ "eval_loss": 0.3108561336994171,
129
+ "eval_runtime": 2.7063,
130
+ "eval_samples_per_second": 369.512,
131
  "step": 2200
132
  },
133
  {
134
  "epoch": 1.16,
135
+ "eval_accuracy": 55.7,
136
+ "eval_average_metrics": 55.7,
137
+ "eval_loss": 0.2974016070365906,
138
+ "eval_runtime": 3.5599,
139
+ "eval_samples_per_second": 280.91,
140
  "step": 2400
141
  },
142
  {
143
  "epoch": 1.21,
144
  "learning_rate": 0.0001794599807135969,
145
+ "loss": 0.3931,
146
  "step": 2500
147
  },
148
  {
149
  "epoch": 1.25,
150
+ "eval_accuracy": 55.7,
151
+ "eval_average_metrics": 55.7,
152
+ "eval_loss": 0.3047039210796356,
153
+ "eval_runtime": 3.0837,
154
+ "eval_samples_per_second": 324.282,
155
  "step": 2600
156
  },
157
  {
158
  "epoch": 1.35,
159
+ "eval_accuracy": 55.900000000000006,
160
+ "eval_average_metrics": 55.900000000000006,
161
+ "eval_loss": 0.2895060181617737,
162
+ "eval_runtime": 3.9994,
163
+ "eval_samples_per_second": 250.039,
164
  "step": 2800
165
  },
166
  {
167
  "epoch": 1.45,
168
  "learning_rate": 0.00015535197685631627,
169
+ "loss": 0.3777,
170
  "step": 3000
171
  },
172
  {
173
  "epoch": 1.45,
174
+ "eval_accuracy": 56.2,
175
+ "eval_average_metrics": 56.2,
176
+ "eval_loss": 0.28209826350212097,
177
+ "eval_runtime": 3.5924,
178
+ "eval_samples_per_second": 278.366,
179
  "step": 3000
180
  },
181
  {
182
  "epoch": 1.54,
183
+ "eval_accuracy": 57.4,
184
+ "eval_average_metrics": 57.4,
185
+ "eval_loss": 0.2806238532066345,
186
+ "eval_runtime": 3.1947,
187
+ "eval_samples_per_second": 313.019,
188
  "step": 3200
189
  },
190
  {
191
  "epoch": 1.64,
192
+ "eval_accuracy": 56.699999999999996,
193
+ "eval_average_metrics": 56.699999999999996,
194
+ "eval_loss": 0.2809857428073883,
195
+ "eval_runtime": 3.3046,
196
+ "eval_samples_per_second": 302.612,
197
  "step": 3400
198
  },
199
  {
200
  "epoch": 1.69,
201
  "learning_rate": 0.00013124397299903566,
202
+ "loss": 0.372,
203
  "step": 3500
204
  },
205
  {
206
  "epoch": 1.74,
207
+ "eval_accuracy": 56.89999999999999,
208
+ "eval_average_metrics": 56.89999999999999,
209
+ "eval_loss": 0.2799268066883087,
210
+ "eval_runtime": 3.5652,
211
+ "eval_samples_per_second": 280.491,
212
  "step": 3600
213
  },
214
  {
215
  "epoch": 1.83,
216
+ "eval_accuracy": 56.599999999999994,
217
+ "eval_average_metrics": 56.599999999999994,
218
+ "eval_loss": 0.2795256972312927,
219
+ "eval_runtime": 3.5541,
220
+ "eval_samples_per_second": 281.363,
221
  "step": 3800
222
  },
223
  {
224
  "epoch": 1.93,
225
  "learning_rate": 0.00010713596914175504,
226
+ "loss": 0.3629,
227
  "step": 4000
228
  },
229
  {
230
  "epoch": 1.93,
231
+ "eval_accuracy": 57.099999999999994,
232
+ "eval_average_metrics": 57.099999999999994,
233
+ "eval_loss": 0.27900928258895874,
234
+ "eval_runtime": 3.42,
235
+ "eval_samples_per_second": 292.396,
236
  "step": 4000
237
  },
238
  {
239
  "epoch": 2.03,
240
+ "eval_accuracy": 57.699999999999996,
241
+ "eval_average_metrics": 57.699999999999996,
242
+ "eval_loss": 0.27878034114837646,
243
+ "eval_runtime": 3.38,
244
+ "eval_samples_per_second": 295.859,
245
  "step": 4200
246
  },
247
  {
248
  "epoch": 2.12,
249
+ "eval_accuracy": 57.699999999999996,
250
+ "eval_average_metrics": 57.699999999999996,
251
+ "eval_loss": 0.27947279810905457,
252
+ "eval_runtime": 3.1427,
253
+ "eval_samples_per_second": 318.196,
254
  "step": 4400
255
  },
256
  {
257
  "epoch": 2.17,
258
  "learning_rate": 8.302796528447444e-05,
259
+ "loss": 0.3597,
260
  "step": 4500
261
  },
262
  {
263
  "epoch": 2.22,
264
+ "eval_accuracy": 57.99999999999999,
265
+ "eval_average_metrics": 57.99999999999999,
266
+ "eval_loss": 0.2773045599460602,
267
+ "eval_runtime": 3.1388,
268
+ "eval_samples_per_second": 318.593,
269
  "step": 4600
270
  },
271
  {
272
  "epoch": 2.31,
273
+ "eval_accuracy": 59.8,
274
+ "eval_average_metrics": 59.8,
275
+ "eval_loss": 0.2791491448879242,
276
+ "eval_runtime": 3.8832,
277
+ "eval_samples_per_second": 257.519,
278
  "step": 4800
279
  },
280
  {
281
  "epoch": 2.41,
282
  "learning_rate": 5.891996142719383e-05,
283
+ "loss": 0.3555,
284
  "step": 5000
285
  },
286
  {
287
  "epoch": 2.41,
288
+ "eval_accuracy": 59.4,
289
+ "eval_average_metrics": 59.4,
290
+ "eval_loss": 0.2792259454727173,
291
+ "eval_runtime": 3.3994,
292
+ "eval_samples_per_second": 294.169,
293
  "step": 5000
294
  },
295
  {
296
  "epoch": 2.51,
297
+ "eval_accuracy": 59.3,
298
+ "eval_average_metrics": 59.3,
299
+ "eval_loss": 0.27831217646598816,
300
+ "eval_runtime": 3.2184,
301
+ "eval_samples_per_second": 310.718,
302
  "step": 5200
303
  },
304
  {
305
  "epoch": 2.6,
306
+ "eval_accuracy": 59.199999999999996,
307
+ "eval_average_metrics": 59.199999999999996,
308
+ "eval_loss": 0.2772988975048065,
309
+ "eval_runtime": 3.2951,
310
+ "eval_samples_per_second": 303.485,
311
  "step": 5400
312
  },
313
  {
314
  "epoch": 2.65,
315
  "learning_rate": 3.481195756991321e-05,
316
+ "loss": 0.3528,
317
  "step": 5500
318
  },
319
  {
320
  "epoch": 2.7,
321
+ "eval_accuracy": 59.9,
322
+ "eval_average_metrics": 59.9,
323
+ "eval_loss": 0.2777373790740967,
324
+ "eval_runtime": 3.3995,
325
+ "eval_samples_per_second": 294.161,
326
  "step": 5600
327
  },
328
  {
329
  "epoch": 2.8,
330
+ "eval_accuracy": 60.099999999999994,
331
+ "eval_average_metrics": 60.099999999999994,
332
+ "eval_loss": 0.2783910036087036,
333
+ "eval_runtime": 3.8322,
334
+ "eval_samples_per_second": 260.948,
335
  "step": 5800
336
  },
337
  {
338
  "epoch": 2.89,
339
  "learning_rate": 1.0703953712632592e-05,
340
+ "loss": 0.3503,
341
  "step": 6000
342
  },
343
  {
344
  "epoch": 2.89,
345
+ "eval_accuracy": 59.699999999999996,
346
+ "eval_average_metrics": 59.699999999999996,
347
+ "eval_loss": 0.27744776010513306,
348
+ "eval_runtime": 3.3726,
349
+ "eval_samples_per_second": 296.511,
350
  "step": 6000
351
  },
352
  {
353
  "epoch": 2.99,
354
+ "eval_accuracy": 59.9,
355
+ "eval_average_metrics": 59.9,
356
+ "eval_loss": 0.27743667364120483,
357
+ "eval_runtime": 3.6434,
358
+ "eval_samples_per_second": 274.47,
359
  "step": 6200
360
  },
361
  {
362
  "epoch": 3.0,
363
  "step": 6222,
364
+ "total_flos": 1.316605032958464e+16,
365
+ "train_loss": 0.5804944922785328,
366
+ "train_runtime": 1254.7674,
367
+ "train_samples_per_second": 158.633,
368
+ "train_steps_per_second": 4.959
369
  }
370
  ],
371
  "max_steps": 6222,
372
  "num_train_epochs": 3,
373
+ "total_flos": 1.316605032958464e+16,
374
  "trial_name": null,
375
  "trial_params": null
376
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5f396571923132052a008693cef3a6acfe656e9be0d56e2e10ea73893ffe1b46
3
  size 3183
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:011028b4522e6a361729bdf6448cb8ad949df04a35d5783ce734197da2e28eb4
3
  size 3183
training_config.json CHANGED
@@ -1 +1 @@
1
- {"dataset_config_name": ["en"], "delta_type": "soft_prompt", "do_eval": true, "do_test": true, "do_train": true, "eval_dataset_config_name": ["en"], "eval_dataset_name": "sst2", "eval_steps": 200, "evaluation_strategy": "steps", "greater_is_better": true, "learning_rate": 0.0003, "load_best_model_at_end": true, "max_source_length": 128, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 3, "output_dir": "outputs/soft_prompt/t5-base/sst2", "overwrite_output_dir": true, "per_device_eval_batch_size": 32, "per_device_train_batch_size": 32, "predict_with_generate": true, "push_to_hub": true, "save_steps": 200, "save_strategy": "steps", "save_total_limit": 1, "seed": 42, "soft_token_num": 100, "split_validation_test": true, "task_name": "sst2", "test_dataset_config_name": ["en"], "test_dataset_name": "sst2", "tokenizer_name": "../../../../plm_cache/t5-base", "unfrozen_modules": ["deltas"], "warmup_steps": 0}
 
1
+ {"dataset_config_name": ["en"], "delta_type": "soft_prompt", "do_eval": true, "do_test": true, "do_train": true, "eval_dataset_config_name": ["en"], "eval_dataset_name": "sst2", "eval_steps": 200, "evaluation_strategy": "steps", "greater_is_better": true, "learning_rate": 0.03, "load_best_model_at_end": true, "max_source_length": 128, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 3, "output_dir": "outputs/soft_prompt/t5-base/sst2", "overwrite_output_dir": true, "per_device_eval_batch_size": 32, "per_device_train_batch_size": 32, "predict_with_generate": true, "push_to_hub": true, "save_steps": 200, "save_strategy": "steps", "save_total_limit": 1, "seed": 42, "soft_token_num": 100, "split_validation_test": true, "task_name": "sst2", "test_dataset_config_name": ["en"], "test_dataset_name": "sst2", "tokenizer_name": "../../../../plm_cache/t5-base", "unfrozen_modules": ["deltas"], "warmup_steps": 0}