jordyvl commited on
Commit
e2fa4d2
1 Parent(s): 028a07e

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +164 -0
README.md ADDED
@@ -0,0 +1,164 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ metrics:
5
+ - accuracy
6
+ model-index:
7
+ - name: dit-base_tobacco_small_student
8
+ results: []
9
+ ---
10
+
11
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
12
+ should probably proofread and complete it, then remove this comment. -->
13
+
14
+ # dit-base_tobacco_small_student
15
+
16
+ This model is a fine-tuned version of [microsoft/dit-base](https://huggingface.co/microsoft/dit-base) on the None dataset.
17
+ It achieves the following results on the evaluation set:
18
+ - Loss: 4.3305
19
+ - Accuracy: 0.435
20
+ - Brier Loss: 1.0472
21
+ - Nll: 10.3327
22
+ - F1 Micro: 0.435
23
+ - F1 Macro: 0.4299
24
+ - Ece: 0.5115
25
+ - Aurc: 0.4245
26
+
27
+ ## Model description
28
+
29
+ More information needed
30
+
31
+ ## Intended uses & limitations
32
+
33
+ More information needed
34
+
35
+ ## Training and evaluation data
36
+
37
+ More information needed
38
+
39
+ ## Training procedure
40
+
41
+ ### Training hyperparameters
42
+
43
+ The following hyperparameters were used during training:
44
+ - learning_rate: 0.0001
45
+ - train_batch_size: 16
46
+ - eval_batch_size: 16
47
+ - seed: 42
48
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
49
+ - lr_scheduler_type: linear
50
+ - lr_scheduler_warmup_ratio: 0.1
51
+ - num_epochs: 100
52
+
53
+ ### Training results
54
+
55
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
56
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|:----------:|:-------:|:--------:|:--------:|:------:|:------:|
57
+ | No log | 1.0 | 50 | 2.1780 | 0.16 | 0.8745 | 11.2696 | 0.16 | 0.0323 | 0.2326 | 0.8208 |
58
+ | No log | 2.0 | 100 | 2.1761 | 0.19 | 0.8727 | 10.5065 | 0.19 | 0.0548 | 0.2712 | 0.7980 |
59
+ | No log | 3.0 | 150 | 2.1426 | 0.16 | 0.8689 | 8.8915 | 0.16 | 0.0451 | 0.2697 | 0.6322 |
60
+ | No log | 4.0 | 200 | 2.0668 | 0.225 | 0.8434 | 9.6036 | 0.225 | 0.1219 | 0.2680 | 0.6623 |
61
+ | No log | 5.0 | 250 | 2.0633 | 0.21 | 0.8447 | 5.7679 | 0.2100 | 0.1401 | 0.2733 | 0.5765 |
62
+ | No log | 6.0 | 300 | 2.0030 | 0.22 | 0.8351 | 7.1501 | 0.22 | 0.1132 | 0.3000 | 0.6750 |
63
+ | No log | 7.0 | 350 | 1.9273 | 0.32 | 0.8243 | 6.2911 | 0.32 | 0.2612 | 0.2822 | 0.6549 |
64
+ | No log | 8.0 | 400 | 1.7954 | 0.365 | 0.7742 | 4.2641 | 0.3650 | 0.2647 | 0.2630 | 0.5031 |
65
+ | No log | 9.0 | 450 | 1.8070 | 0.36 | 0.7720 | 4.9274 | 0.36 | 0.2914 | 0.2601 | 0.4871 |
66
+ | 1.9795 | 10.0 | 500 | 1.7838 | 0.34 | 0.7857 | 3.3860 | 0.34 | 0.2387 | 0.2902 | 0.5057 |
67
+ | 1.9795 | 11.0 | 550 | 1.7214 | 0.395 | 0.7404 | 4.1630 | 0.395 | 0.2995 | 0.2922 | 0.4210 |
68
+ | 1.9795 | 12.0 | 600 | 1.6834 | 0.445 | 0.7284 | 3.7081 | 0.445 | 0.3444 | 0.2700 | 0.3914 |
69
+ | 1.9795 | 13.0 | 650 | 1.6992 | 0.38 | 0.7641 | 4.1246 | 0.38 | 0.3045 | 0.3375 | 0.4155 |
70
+ | 1.9795 | 14.0 | 700 | 1.8695 | 0.395 | 0.7711 | 5.6899 | 0.395 | 0.3432 | 0.3224 | 0.4425 |
71
+ | 1.9795 | 15.0 | 750 | 1.8757 | 0.38 | 0.7939 | 5.1099 | 0.38 | 0.3879 | 0.3102 | 0.4313 |
72
+ | 1.9795 | 16.0 | 800 | 2.0457 | 0.405 | 0.8184 | 5.6034 | 0.405 | 0.3957 | 0.3256 | 0.4414 |
73
+ | 1.9795 | 17.0 | 850 | 2.2243 | 0.395 | 0.8653 | 7.7124 | 0.395 | 0.3567 | 0.3887 | 0.3997 |
74
+ | 1.9795 | 18.0 | 900 | 1.9309 | 0.45 | 0.7794 | 5.2698 | 0.45 | 0.3763 | 0.3626 | 0.3767 |
75
+ | 1.9795 | 19.0 | 950 | 2.2285 | 0.415 | 0.8319 | 6.7127 | 0.415 | 0.4153 | 0.3667 | 0.3942 |
76
+ | 0.6717 | 20.0 | 1000 | 2.3745 | 0.445 | 0.8643 | 7.4432 | 0.445 | 0.4290 | 0.3859 | 0.4046 |
77
+ | 0.6717 | 21.0 | 1050 | 2.5389 | 0.41 | 0.9148 | 7.6865 | 0.41 | 0.3994 | 0.4351 | 0.4054 |
78
+ | 0.6717 | 22.0 | 1100 | 2.5537 | 0.465 | 0.8500 | 8.1266 | 0.465 | 0.4623 | 0.4070 | 0.3900 |
79
+ | 0.6717 | 23.0 | 1150 | 2.8355 | 0.42 | 0.9426 | 8.8542 | 0.4200 | 0.3930 | 0.4508 | 0.4201 |
80
+ | 0.6717 | 24.0 | 1200 | 2.8575 | 0.4 | 0.9962 | 7.6428 | 0.4000 | 0.3502 | 0.4994 | 0.4119 |
81
+ | 0.6717 | 25.0 | 1250 | 2.8704 | 0.445 | 0.9418 | 9.2600 | 0.445 | 0.4570 | 0.4309 | 0.4021 |
82
+ | 0.6717 | 26.0 | 1300 | 3.4702 | 0.43 | 0.9641 | 12.1621 | 0.4300 | 0.3977 | 0.4590 | 0.3597 |
83
+ | 0.6717 | 27.0 | 1350 | 3.1484 | 0.475 | 0.9518 | 8.1474 | 0.4750 | 0.4641 | 0.4809 | 0.4088 |
84
+ | 0.6717 | 28.0 | 1400 | 3.2299 | 0.455 | 0.9673 | 9.6161 | 0.455 | 0.4205 | 0.4711 | 0.3806 |
85
+ | 0.6717 | 29.0 | 1450 | 3.4968 | 0.425 | 1.0136 | 10.5614 | 0.425 | 0.3992 | 0.4743 | 0.3773 |
86
+ | 0.0268 | 30.0 | 1500 | 3.1340 | 0.46 | 0.9443 | 8.5023 | 0.46 | 0.4296 | 0.4557 | 0.3735 |
87
+ | 0.0268 | 31.0 | 1550 | 3.4297 | 0.435 | 1.0058 | 8.2428 | 0.435 | 0.3979 | 0.4967 | 0.3848 |
88
+ | 0.0268 | 32.0 | 1600 | 3.6922 | 0.4 | 1.0488 | 10.8019 | 0.4000 | 0.3880 | 0.5223 | 0.4017 |
89
+ | 0.0268 | 33.0 | 1650 | 3.6009 | 0.445 | 0.9964 | 10.1007 | 0.445 | 0.4204 | 0.4924 | 0.3981 |
90
+ | 0.0268 | 34.0 | 1700 | 3.6678 | 0.425 | 1.0494 | 9.1369 | 0.425 | 0.3896 | 0.5159 | 0.4192 |
91
+ | 0.0268 | 35.0 | 1750 | 3.5743 | 0.45 | 0.9953 | 9.5996 | 0.45 | 0.4182 | 0.4934 | 0.4030 |
92
+ | 0.0268 | 36.0 | 1800 | 3.5551 | 0.465 | 0.9877 | 9.6080 | 0.465 | 0.4221 | 0.5033 | 0.3977 |
93
+ | 0.0268 | 37.0 | 1850 | 3.7424 | 0.435 | 1.0191 | 9.9258 | 0.435 | 0.4292 | 0.4955 | 0.4120 |
94
+ | 0.0268 | 38.0 | 1900 | 3.7093 | 0.45 | 1.0051 | 9.7038 | 0.45 | 0.4033 | 0.4966 | 0.3857 |
95
+ | 0.0268 | 39.0 | 1950 | 3.7240 | 0.45 | 1.0076 | 9.8462 | 0.45 | 0.4027 | 0.4953 | 0.3962 |
96
+ | 0.0022 | 40.0 | 2000 | 3.7503 | 0.455 | 1.0090 | 9.9967 | 0.455 | 0.4076 | 0.5056 | 0.3968 |
97
+ | 0.0022 | 41.0 | 2050 | 3.5545 | 0.44 | 1.0007 | 8.7616 | 0.44 | 0.4285 | 0.4894 | 0.4008 |
98
+ | 0.0022 | 42.0 | 2100 | 3.7452 | 0.435 | 1.0142 | 9.4376 | 0.435 | 0.4135 | 0.5032 | 0.4022 |
99
+ | 0.0022 | 43.0 | 2150 | 3.5980 | 0.47 | 0.9532 | 8.2333 | 0.47 | 0.4441 | 0.4650 | 0.4113 |
100
+ | 0.0022 | 44.0 | 2200 | 3.7055 | 0.45 | 0.9946 | 9.0121 | 0.45 | 0.4327 | 0.4817 | 0.3688 |
101
+ | 0.0022 | 45.0 | 2250 | 3.8500 | 0.435 | 1.0161 | 9.2035 | 0.435 | 0.4164 | 0.5128 | 0.3723 |
102
+ | 0.0022 | 46.0 | 2300 | 3.8806 | 0.435 | 1.0261 | 10.7033 | 0.435 | 0.4323 | 0.5008 | 0.3812 |
103
+ | 0.0022 | 47.0 | 2350 | 3.8114 | 0.455 | 1.0128 | 9.6784 | 0.455 | 0.4236 | 0.5025 | 0.3873 |
104
+ | 0.0022 | 48.0 | 2400 | 3.8743 | 0.435 | 1.0294 | 8.7193 | 0.435 | 0.3734 | 0.5109 | 0.3783 |
105
+ | 0.0022 | 49.0 | 2450 | 3.9281 | 0.43 | 1.0414 | 9.9489 | 0.4300 | 0.4296 | 0.5047 | 0.4049 |
106
+ | 0.0047 | 50.0 | 2500 | 3.7824 | 0.45 | 0.9956 | 10.7814 | 0.45 | 0.4467 | 0.4975 | 0.3949 |
107
+ | 0.0047 | 51.0 | 2550 | 4.0089 | 0.475 | 0.9668 | 11.9005 | 0.4750 | 0.4253 | 0.4637 | 0.4501 |
108
+ | 0.0047 | 52.0 | 2600 | 3.7248 | 0.43 | 0.9909 | 10.6449 | 0.4300 | 0.4064 | 0.4750 | 0.4023 |
109
+ | 0.0047 | 53.0 | 2650 | 3.7911 | 0.415 | 1.0491 | 9.1188 | 0.415 | 0.3608 | 0.5130 | 0.4173 |
110
+ | 0.0047 | 54.0 | 2700 | 3.6925 | 0.44 | 1.0000 | 8.9655 | 0.44 | 0.3970 | 0.4826 | 0.4168 |
111
+ | 0.0047 | 55.0 | 2750 | 3.6214 | 0.46 | 0.9590 | 9.5422 | 0.46 | 0.4440 | 0.4636 | 0.3829 |
112
+ | 0.0047 | 56.0 | 2800 | 4.3545 | 0.405 | 1.0811 | 10.6531 | 0.405 | 0.4090 | 0.5439 | 0.4533 |
113
+ | 0.0047 | 57.0 | 2850 | 3.6835 | 0.46 | 0.9717 | 8.2408 | 0.46 | 0.4367 | 0.4950 | 0.4118 |
114
+ | 0.0047 | 58.0 | 2900 | 4.0080 | 0.465 | 1.0011 | 9.3764 | 0.465 | 0.4579 | 0.4927 | 0.4234 |
115
+ | 0.0047 | 59.0 | 2950 | 4.0141 | 0.45 | 1.0014 | 9.7100 | 0.45 | 0.4443 | 0.4987 | 0.4220 |
116
+ | 0.0118 | 60.0 | 3000 | 3.7963 | 0.43 | 1.0135 | 9.4040 | 0.4300 | 0.4007 | 0.5007 | 0.3979 |
117
+ | 0.0118 | 61.0 | 3050 | 4.0609 | 0.43 | 1.0426 | 9.3533 | 0.4300 | 0.3825 | 0.5266 | 0.4285 |
118
+ | 0.0118 | 62.0 | 3100 | 4.0150 | 0.47 | 1.0002 | 9.3307 | 0.47 | 0.4490 | 0.5030 | 0.4052 |
119
+ | 0.0118 | 63.0 | 3150 | 3.7982 | 0.47 | 0.9660 | 8.5060 | 0.47 | 0.4581 | 0.4716 | 0.3988 |
120
+ | 0.0118 | 64.0 | 3200 | 4.3553 | 0.44 | 1.0428 | 10.3840 | 0.44 | 0.4218 | 0.5163 | 0.4312 |
121
+ | 0.0118 | 65.0 | 3250 | 3.7142 | 0.44 | 0.9900 | 8.5049 | 0.44 | 0.4298 | 0.4849 | 0.3735 |
122
+ | 0.0118 | 66.0 | 3300 | 3.7411 | 0.47 | 0.9661 | 8.1935 | 0.47 | 0.4497 | 0.4789 | 0.3812 |
123
+ | 0.0118 | 67.0 | 3350 | 3.7858 | 0.49 | 0.9574 | 8.8397 | 0.49 | 0.4799 | 0.4616 | 0.3895 |
124
+ | 0.0118 | 68.0 | 3400 | 3.7927 | 0.495 | 0.9459 | 8.6915 | 0.495 | 0.4870 | 0.4577 | 0.3883 |
125
+ | 0.0118 | 69.0 | 3450 | 3.8348 | 0.5 | 0.9454 | 8.8298 | 0.5 | 0.4889 | 0.4715 | 0.3891 |
126
+ | 0.0004 | 70.0 | 3500 | 3.8551 | 0.48 | 0.9500 | 8.9827 | 0.48 | 0.4755 | 0.4691 | 0.3913 |
127
+ | 0.0004 | 71.0 | 3550 | 3.8432 | 0.48 | 0.9622 | 9.1404 | 0.48 | 0.4691 | 0.4690 | 0.3885 |
128
+ | 0.0004 | 72.0 | 3600 | 3.8594 | 0.48 | 0.9617 | 8.8182 | 0.48 | 0.4691 | 0.4805 | 0.3854 |
129
+ | 0.0004 | 73.0 | 3650 | 3.8855 | 0.485 | 0.9622 | 8.8248 | 0.485 | 0.4760 | 0.4809 | 0.3881 |
130
+ | 0.0004 | 74.0 | 3700 | 3.8996 | 0.49 | 0.9610 | 8.9750 | 0.49 | 0.4818 | 0.4634 | 0.3892 |
131
+ | 0.0004 | 75.0 | 3750 | 3.9921 | 0.475 | 0.9642 | 9.5409 | 0.4750 | 0.4597 | 0.4666 | 0.4185 |
132
+ | 0.0004 | 76.0 | 3800 | 4.1128 | 0.43 | 1.0429 | 9.9966 | 0.4300 | 0.3844 | 0.5187 | 0.4056 |
133
+ | 0.0004 | 77.0 | 3850 | 4.0783 | 0.44 | 1.0172 | 9.3016 | 0.44 | 0.4205 | 0.5051 | 0.3988 |
134
+ | 0.0004 | 78.0 | 3900 | 4.0804 | 0.445 | 1.0254 | 8.9753 | 0.445 | 0.4246 | 0.5089 | 0.3982 |
135
+ | 0.0004 | 79.0 | 3950 | 4.0892 | 0.445 | 1.0269 | 8.8290 | 0.445 | 0.4246 | 0.5069 | 0.4000 |
136
+ | 0.0002 | 80.0 | 4000 | 4.1013 | 0.445 | 1.0258 | 9.1363 | 0.445 | 0.4246 | 0.5129 | 0.4033 |
137
+ | 0.0002 | 81.0 | 4050 | 4.0985 | 0.44 | 1.0287 | 9.1459 | 0.44 | 0.4213 | 0.5074 | 0.4054 |
138
+ | 0.0002 | 82.0 | 4100 | 4.1029 | 0.44 | 1.0263 | 9.3107 | 0.44 | 0.4211 | 0.5125 | 0.4066 |
139
+ | 0.0002 | 83.0 | 4150 | 4.1075 | 0.44 | 1.0248 | 9.4604 | 0.44 | 0.4224 | 0.5164 | 0.4061 |
140
+ | 0.0002 | 84.0 | 4200 | 4.1087 | 0.44 | 1.0225 | 9.7739 | 0.44 | 0.4221 | 0.5090 | 0.4055 |
141
+ | 0.0002 | 85.0 | 4250 | 4.1248 | 0.44 | 1.0262 | 9.7747 | 0.44 | 0.4259 | 0.5032 | 0.4065 |
142
+ | 0.0002 | 86.0 | 4300 | 4.1527 | 0.445 | 1.0263 | 9.4647 | 0.445 | 0.4299 | 0.5128 | 0.4066 |
143
+ | 0.0002 | 87.0 | 4350 | 4.0529 | 0.475 | 0.9810 | 9.1439 | 0.4750 | 0.4488 | 0.4910 | 0.3938 |
144
+ | 0.0002 | 88.0 | 4400 | 4.1405 | 0.455 | 1.0091 | 9.5149 | 0.455 | 0.4230 | 0.4966 | 0.4147 |
145
+ | 0.0002 | 89.0 | 4450 | 4.3483 | 0.41 | 1.0724 | 9.8421 | 0.41 | 0.4083 | 0.5384 | 0.4090 |
146
+ | 0.0008 | 90.0 | 4500 | 4.5574 | 0.39 | 1.1077 | 11.2517 | 0.39 | 0.3940 | 0.5618 | 0.4405 |
147
+ | 0.0008 | 91.0 | 4550 | 4.5104 | 0.41 | 1.0890 | 10.8687 | 0.41 | 0.4173 | 0.5411 | 0.4350 |
148
+ | 0.0008 | 92.0 | 4600 | 4.3791 | 0.425 | 1.0672 | 10.7198 | 0.425 | 0.4202 | 0.5233 | 0.4306 |
149
+ | 0.0008 | 93.0 | 4650 | 4.3608 | 0.43 | 1.0553 | 10.8428 | 0.4300 | 0.4236 | 0.5196 | 0.4284 |
150
+ | 0.0008 | 94.0 | 4700 | 4.3469 | 0.44 | 1.0474 | 10.6774 | 0.44 | 0.4428 | 0.5020 | 0.4280 |
151
+ | 0.0008 | 95.0 | 4750 | 4.3420 | 0.44 | 1.0487 | 10.5138 | 0.44 | 0.4385 | 0.5260 | 0.4270 |
152
+ | 0.0008 | 96.0 | 4800 | 4.3385 | 0.435 | 1.0491 | 10.3448 | 0.435 | 0.4312 | 0.5170 | 0.4266 |
153
+ | 0.0008 | 97.0 | 4850 | 4.3341 | 0.435 | 1.0485 | 10.3378 | 0.435 | 0.4312 | 0.5136 | 0.4261 |
154
+ | 0.0008 | 98.0 | 4900 | 4.3336 | 0.435 | 1.0480 | 10.3350 | 0.435 | 0.4312 | 0.5184 | 0.4253 |
155
+ | 0.0008 | 99.0 | 4950 | 4.3306 | 0.435 | 1.0472 | 10.3328 | 0.435 | 0.4299 | 0.5116 | 0.4245 |
156
+ | 0.0001 | 100.0 | 5000 | 4.3305 | 0.435 | 1.0472 | 10.3327 | 0.435 | 0.4299 | 0.5115 | 0.4245 |
157
+
158
+
159
+ ### Framework versions
160
+
161
+ - Transformers 4.28.0.dev0
162
+ - Pytorch 1.12.1+cu113
163
+ - Datasets 2.12.0
164
+ - Tokenizers 0.12.1