Hugging Face's logo

Yova
/
baseline

Text2Text Generation
Transformers
TensorBoard
Safetensors
t5
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Metrics Community
Yova commited on
Commit
696c7be
·
1 Parent(s): ca79586

End of training

Browse files
Files changed (3) hide show
  1. README.md +111 -107
  2. generation_config.json +1 -0
  3. model.safetensors +1 -1
README.md CHANGED
@@ -13,8 +13,8 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
15
  It achieves the following results on the evaluation set:
16
- - Loss: 1.6338
17
- - Exact Match: 0.142
18
 
19
  ## Model description
20
 
@@ -34,117 +34,121 @@ More information needed
34
 
35
  The following hyperparameters were used during training:
36
  - learning_rate: 0.001
37
- - train_batch_size: 32
38
  - eval_batch_size: 8
39
  - seed: 42
40
- - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
41
- - lr_scheduler_type: linear
 
 
 
42
  - num_epochs: 100
 
43
 
44
  ### Training results
45
 
46
- | Training Loss | Epoch | Step | Validation Loss | Exact Match |
47
- |:-------------:|:-----:|:-----:|:---------------:|:-----------:|
48
- | 2.6989 | 1.0 | 313 | 1.8586 | 0.0 |
49
- | 1.834 | 2.0 | 626 | 1.5284 | 0.003 |
50
- | 1.517 | 3.0 | 939 | 1.3632 | 0.005 |
51
- | 1.2977 | 4.0 | 1252 | 1.2077 | 0.021 |
52
- | 1.124 | 5.0 | 1565 | 1.1030 | 0.037 |
53
- | 0.9885 | 6.0 | 1878 | 1.0607 | 0.05 |
54
- | 0.8762 | 7.0 | 2191 | 1.0329 | 0.047 |
55
- | 0.7698 | 8.0 | 2504 | 1.0087 | 0.063 |
56
- | 0.6983 | 9.0 | 2817 | 0.9963 | 0.046 |
57
- | 0.6297 | 10.0 | 3130 | 0.9754 | 0.076 |
58
- | 0.5719 | 11.0 | 3443 | 0.9907 | 0.075 |
59
- | 0.5247 | 12.0 | 3756 | 0.9777 | 0.069 |
60
- | 0.4776 | 13.0 | 4069 | 0.9766 | 0.055 |
61
- | 0.442 | 14.0 | 4382 | 0.9953 | 0.091 |
62
- | 0.4081 | 15.0 | 4695 | 1.0005 | 0.098 |
63
- | 0.3783 | 16.0 | 5008 | 1.0274 | 0.093 |
64
- | 0.3545 | 17.0 | 5321 | 1.0516 | 0.087 |
65
- | 0.3243 | 18.0 | 5634 | 1.0339 | 0.09 |
66
- | 0.3045 | 19.0 | 5947 | 1.0419 | 0.078 |
67
- | 0.2841 | 20.0 | 6260 | 1.0640 | 0.087 |
68
- | 0.2692 | 21.0 | 6573 | 1.0839 | 0.105 |
69
- | 0.2543 | 22.0 | 6886 | 1.1608 | 0.064 |
70
- | 0.2346 | 23.0 | 7199 | 1.1046 | 0.113 |
71
- | 0.2245 | 24.0 | 7512 | 1.1569 | 0.128 |
72
- | 0.2135 | 25.0 | 7825 | 1.1242 | 0.108 |
73
- | 0.2029 | 26.0 | 8138 | 1.1436 | 0.118 |
74
- | 0.1902 | 27.0 | 8451 | 1.2023 | 0.095 |
75
- | 0.1832 | 28.0 | 8764 | 1.1556 | 0.115 |
76
- | 0.171 | 29.0 | 9077 | 1.2068 | 0.094 |
77
- | 0.1639 | 30.0 | 9390 | 1.2101 | 0.151 |
78
- | 0.1581 | 31.0 | 9703 | 1.2299 | 0.112 |
79
- | 0.1504 | 32.0 | 10016 | 1.3153 | 0.1 |
80
- | 0.1463 | 33.0 | 10329 | 1.2785 | 0.091 |
81
- | 0.1405 | 34.0 | 10642 | 1.2662 | 0.111 |
82
- | 0.1349 | 35.0 | 10955 | 1.2805 | 0.134 |
83
- | 0.1291 | 36.0 | 11268 | 1.2516 | 0.137 |
84
- | 0.126 | 37.0 | 11581 | 1.3312 | 0.141 |
85
- | 0.1204 | 38.0 | 11894 | 1.2776 | 0.116 |
86
- | 0.1163 | 39.0 | 12207 | 1.3203 | 0.11 |
87
- | 0.114 | 40.0 | 12520 | 1.3212 | 0.129 |
88
- | 0.1056 | 41.0 | 12833 | 1.3291 | 0.127 |
89
- | 0.1033 | 42.0 | 13146 | 1.3010 | 0.125 |
90
- | 0.1034 | 43.0 | 13459 | 1.3206 | 0.135 |
91
- | 0.098 | 44.0 | 13772 | 1.3879 | 0.127 |
92
- | 0.0951 | 45.0 | 14085 | 1.3693 | 0.111 |
93
- | 0.089 | 46.0 | 14398 | 1.4261 | 0.124 |
94
- | 0.0913 | 47.0 | 14711 | 1.3644 | 0.122 |
95
- | 0.0863 | 48.0 | 15024 | 1.4392 | 0.108 |
96
- | 0.0809 | 49.0 | 15337 | 1.3726 | 0.098 |
97
- | 0.0795 | 50.0 | 15650 | 1.3791 | 0.084 |
98
- | 0.0763 | 51.0 | 15963 | 1.3911 | 0.134 |
99
- | 0.0768 | 52.0 | 16276 | 1.4202 | 0.104 |
100
- | 0.076 | 53.0 | 16589 | 1.4594 | 0.122 |
101
- | 0.0734 | 54.0 | 16902 | 1.4541 | 0.129 |
102
- | 0.0714 | 55.0 | 17215 | 1.4032 | 0.133 |
103
- | 0.0696 | 56.0 | 17528 | 1.4467 | 0.128 |
104
- | 0.0674 | 57.0 | 17841 | 1.4952 | 0.103 |
105
- | 0.0657 | 58.0 | 18154 | 1.4582 | 0.14 |
106
- | 0.0658 | 59.0 | 18467 | 1.4619 | 0.121 |
107
- | 0.061 | 60.0 | 18780 | 1.5447 | 0.111 |
108
- | 0.0609 | 61.0 | 19093 | 1.4233 | 0.16 |
109
- | 0.0596 | 62.0 | 19406 | 1.4705 | 0.134 |
110
- | 0.058 | 63.0 | 19719 | 1.4721 | 0.144 |
111
- | 0.0555 | 64.0 | 20032 | 1.4377 | 0.156 |
112
- | 0.0532 | 65.0 | 20345 | 1.5016 | 0.125 |
113
- | 0.0559 | 66.0 | 20658 | 1.5405 | 0.156 |
114
- | 0.0517 | 67.0 | 20971 | 1.5166 | 0.133 |
115
- | 0.0499 | 68.0 | 21284 | 1.4787 | 0.139 |
116
- | 0.0477 | 69.0 | 21597 | 1.5063 | 0.124 |
117
- | 0.0491 | 70.0 | 21910 | 1.5287 | 0.147 |
118
- | 0.0464 | 71.0 | 22223 | 1.5428 | 0.131 |
119
- | 0.0456 | 72.0 | 22536 | 1.5434 | 0.132 |
120
- | 0.0449 | 73.0 | 22849 | 1.5364 | 0.116 |
121
- | 0.0432 | 74.0 | 23162 | 1.5830 | 0.12 |
122
- | 0.042 | 75.0 | 23475 | 1.5508 | 0.113 |
123
- | 0.0403 | 76.0 | 23788 | 1.5146 | 0.134 |
124
- | 0.0398 | 77.0 | 24101 | 1.5955 | 0.111 |
125
- | 0.0412 | 78.0 | 24414 | 1.5759 | 0.132 |
126
- | 0.0391 | 79.0 | 24727 | 1.5588 | 0.136 |
127
- | 0.0383 | 80.0 | 25040 | 1.5580 | 0.141 |
128
- | 0.0366 | 81.0 | 25353 | 1.5895 | 0.143 |
129
- | 0.0365 | 82.0 | 25666 | 1.5637 | 0.148 |
130
- | 0.035 | 83.0 | 25979 | 1.6012 | 0.155 |
131
- | 0.0359 | 84.0 | 26292 | 1.6130 | 0.118 |
132
- | 0.0343 | 85.0 | 26605 | 1.6038 | 0.158 |
133
- | 0.0333 | 86.0 | 26918 | 1.6300 | 0.124 |
134
- | 0.0318 | 87.0 | 27231 | 1.6259 | 0.145 |
135
- | 0.0309 | 88.0 | 27544 | 1.6178 | 0.139 |
136
- | 0.0303 | 89.0 | 27857 | 1.6166 | 0.143 |
137
- | 0.0302 | 90.0 | 28170 | 1.6394 | 0.141 |
138
- | 0.0293 | 91.0 | 28483 | 1.6408 | 0.154 |
139
- | 0.0281 | 92.0 | 28796 | 1.6424 | 0.13 |
140
- | 0.0288 | 93.0 | 29109 | 1.6426 | 0.136 |
141
- | 0.0272 | 94.0 | 29422 | 1.6477 | 0.131 |
142
- | 0.0278 | 95.0 | 29735 | 1.6288 | 0.142 |
143
- | 0.0264 | 96.0 | 30048 | 1.6251 | 0.142 |
144
- | 0.0268 | 97.0 | 30361 | 1.6340 | 0.142 |
145
- | 0.0255 | 98.0 | 30674 | 1.6353 | 0.145 |
146
- | 0.0263 | 99.0 | 30987 | 1.6333 | 0.143 |
147
- | 0.0259 | 100.0 | 31300 | 1.6338 | 0.142 |
148
 
149
 
150
  ### Framework versions
 
13
 
14
  This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
15
  It achieves the following results on the evaluation set:
16
+ - Loss: 1.8173
17
+ - Exact Match: 0.024
18
 
19
  ## Model description
20
 
 
34
 
35
  The following hyperparameters were used during training:
36
  - learning_rate: 0.001
37
+ - train_batch_size: 100
38
  - eval_batch_size: 8
39
  - seed: 42
40
+ - gradient_accumulation_steps: 4
41
+ - total_train_batch_size: 400
42
+ - optimizer: Adam with betas=(0.98,0.999) and epsilon=1e-08
43
+ - lr_scheduler_type: inverse_sqrt
44
+ - lr_scheduler_warmup_steps: 4000
45
  - num_epochs: 100
46
+ - label_smoothing_factor: 0.1
47
 
48
  ### Training results
49
 
50
+ | Training Loss | Epoch | Step | Validation Loss | Exact Match |
51
+ |:-------------:|:-----:|:----:|:---------------:|:-----------:|
52
+ | 6.2658 | 1.0 | 25 | 5.8355 | 0.0 |
53
+ | 5.8196 | 2.0 | 50 | 4.9580 | 0.0 |
54
+ | 5.1965 | 3.0 | 75 | 4.4378 | 0.0 |
55
+ | 4.7235 | 4.0 | 100 | 4.0961 | 0.0 |
56
+ | 4.4214 | 5.0 | 125 | 3.9076 | 0.0 |
57
+ | 4.2363 | 6.0 | 150 | 3.7785 | 0.0 |
58
+ | 4.1005 | 7.0 | 175 | 3.6651 | 0.0 |
59
+ | 3.9843 | 8.0 | 200 | 3.5463 | 0.0 |
60
+ | 3.8646 | 9.0 | 225 | 3.4428 | 0.0 |
61
+ | 3.741 | 10.0 | 250 | 3.3358 | 0.0 |
62
+ | 3.6147 | 11.0 | 275 | 3.2427 | 0.0 |
63
+ | 3.495 | 12.0 | 300 | 3.1311 | 0.0 |
64
+ | 3.3903 | 13.0 | 325 | 3.0348 | 0.0 |
65
+ | 3.2919 | 14.0 | 350 | 2.9669 | 0.0 |
66
+ | 3.2062 | 15.0 | 375 | 2.9071 | 0.0 |
67
+ | 3.137 | 16.0 | 400 | 2.8355 | 0.0 |
68
+ | 3.0572 | 17.0 | 425 | 2.7804 | 0.0 |
69
+ | 2.9925 | 18.0 | 450 | 2.7326 | 0.0 |
70
+ | 2.9289 | 19.0 | 475 | 2.6681 | 0.0 |
71
+ | 2.8768 | 20.0 | 500 | 2.6168 | 0.0 |
72
+ | 2.8239 | 21.0 | 525 | 2.5649 | 0.0 |
73
+ | 2.7689 | 22.0 | 550 | 2.5171 | 0.0 |
74
+ | 2.7125 | 23.0 | 575 | 2.4704 | 0.0 |
75
+ | 2.6607 | 24.0 | 600 | 2.4268 | 0.0 |
76
+ | 2.6149 | 25.0 | 625 | 2.4030 | 0.0 |
77
+ | 2.56 | 26.0 | 650 | 2.3599 | 0.001 |
78
+ | 2.5085 | 27.0 | 675 | 2.3096 | 0.006 |
79
+ | 2.4572 | 28.0 | 700 | 2.2606 | 0.008 |
80
+ | 2.4133 | 29.0 | 725 | 2.2313 | 0.009 |
81
+ | 2.3613 | 30.0 | 750 | 2.2011 | 0.011 |
82
+ | 2.3141 | 31.0 | 775 | 2.1757 | 0.011 |
83
+ | 2.2714 | 32.0 | 800 | 2.1362 | 0.013 |
84
+ | 2.2226 | 33.0 | 825 | 2.1176 | 0.019 |
85
+ | 2.1821 | 34.0 | 850 | 2.0885 | 0.022 |
86
+ | 2.1445 | 35.0 | 875 | 2.0810 | 0.022 |
87
+ | 2.1082 | 36.0 | 900 | 2.0501 | 0.034 |
88
+ | 2.0683 | 37.0 | 925 | 2.0571 | 0.029 |
89
+ | 2.0365 | 38.0 | 950 | 2.0318 | 0.028 |
90
+ | 2.0003 | 39.0 | 975 | 2.0227 | 0.033 |
91
+ | 1.9654 | 40.0 | 1000 | 2.0141 | 0.042 |
92
+ | 1.9358 | 41.0 | 1025 | 2.0042 | 0.045 |
93
+ | 1.905 | 42.0 | 1050 | 1.9903 | 0.055 |
94
+ | 1.878 | 43.0 | 1075 | 2.0076 | 0.057 |
95
+ | 1.8517 | 44.0 | 1100 | 1.9761 | 0.057 |
96
+ | 1.8233 | 45.0 | 1125 | 1.9952 | 0.063 |
97
+ | 1.7934 | 46.0 | 1150 | 1.9562 | 0.062 |
98
+ | 1.7636 | 47.0 | 1175 | 1.9776 | 0.068 |
99
+ | 1.7445 | 48.0 | 1200 | 1.9503 | 0.066 |
100
+ | 1.7226 | 49.0 | 1225 | 1.9616 | 0.065 |
101
+ | 1.7013 | 50.0 | 1250 | 1.9516 | 0.067 |
102
+ | 1.68 | 51.0 | 1275 | 1.9408 | 0.065 |
103
+ | 1.6557 | 52.0 | 1300 | 1.9493 | 0.073 |
104
+ | 1.6396 | 53.0 | 1325 | 1.9012 | 0.099 |
105
+ | 1.6183 | 54.0 | 1350 | 1.9144 | 0.083 |
106
+ | 1.5918 | 55.0 | 1375 | 1.9150 | 0.085 |
107
+ | 1.5788 | 56.0 | 1400 | 1.9278 | 0.098 |
108
+ | 1.5601 | 57.0 | 1425 | 1.9072 | 0.088 |
109
+ | 1.5464 | 58.0 | 1450 | 1.8896 | 0.084 |
110
+ | 1.5333 | 59.0 | 1475 | 1.9001 | 0.111 |
111
+ | 1.5153 | 60.0 | 1500 | 1.8746 | 0.089 |
112
+ | 1.5034 | 61.0 | 1525 | 1.8869 | 0.089 |
113
+ | 1.4876 | 62.0 | 1550 | 1.8744 | 0.105 |
114
+ | 1.4779 | 63.0 | 1575 | 1.8866 | 0.089 |
115
+ | 1.459 | 64.0 | 1600 | 1.8615 | 0.128 |
116
+ | 1.4447 | 65.0 | 1625 | 1.8565 | 0.111 |
117
+ | 1.4321 | 66.0 | 1650 | 1.8659 | 0.129 |
118
+ | 1.4175 | 67.0 | 1675 | 1.8571 | 0.121 |
119
+ | 1.4071 | 68.0 | 1700 | 1.8831 | 0.076 |
120
+ | 1.3987 | 69.0 | 1725 | 1.8492 | 0.077 |
121
+ | 1.3837 | 70.0 | 1750 | 1.8430 | 0.101 |
122
+ | 1.374 | 71.0 | 1775 | 1.8455 | 0.082 |
123
+ | 1.3645 | 72.0 | 1800 | 1.8506 | 0.064 |
124
+ | 1.3537 | 73.0 | 1825 | 1.8345 | 0.07 |
125
+ | 1.3441 | 74.0 | 1850 | 1.8267 | 0.115 |
126
+ | 1.3348 | 75.0 | 1875 | 1.8504 | 0.084 |
127
+ | 1.3205 | 76.0 | 1900 | 1.8470 | 0.08 |
128
+ | 1.3108 | 77.0 | 1925 | 1.8397 | 0.089 |
129
+ | 1.3028 | 78.0 | 1950 | 1.8657 | 0.073 |
130
+ | 1.2978 | 79.0 | 1975 | 1.8595 | 0.067 |
131
+ | 1.2875 | 80.0 | 2000 | 1.8322 | 0.073 |
132
+ | 1.2753 | 81.0 | 2025 | 1.8697 | 0.04 |
133
+ | 1.2707 | 82.0 | 2050 | 1.8426 | 0.085 |
134
+ | 1.262 | 83.0 | 2075 | 1.8229 | 0.093 |
135
+ | 1.2532 | 84.0 | 2100 | 1.8420 | 0.054 |
136
+ | 1.2488 | 85.0 | 2125 | 1.8465 | 0.057 |
137
+ | 1.2373 | 86.0 | 2150 | 1.8701 | 0.051 |
138
+ | 1.229 | 87.0 | 2175 | 1.8474 | 0.054 |
139
+ | 1.2242 | 88.0 | 2200 | 1.8287 | 0.072 |
140
+ | 1.2195 | 89.0 | 2225 | 1.8424 | 0.057 |
141
+ | 1.212 | 90.0 | 2250 | 1.8663 | 0.062 |
142
+ | 1.2033 | 91.0 | 2275 | 1.8447 | 0.044 |
143
+ | 1.1994 | 92.0 | 2300 | 1.8287 | 0.058 |
144
+ | 1.1907 | 93.0 | 2325 | 1.8425 | 0.05 |
145
+ | 1.1847 | 94.0 | 2350 | 1.8993 | 0.004 |
146
+ | 1.1805 | 95.0 | 2375 | 1.9007 | 0.014 |
147
+ | 1.1739 | 96.0 | 2400 | 1.8792 | 0.015 |
148
+ | 1.1697 | 97.0 | 2425 | 1.8315 | 0.02 |
149
+ | 1.162 | 98.0 | 2450 | 1.8024 | 0.043 |
150
+ | 1.1585 | 99.0 | 2475 | 1.8437 | 0.017 |
151
+ | 1.1505 | 100.0 | 2500 | 1.8173 | 0.024 |
152
 
153
 
154
  ### Framework versions
generation_config.json CHANGED
@@ -2,6 +2,7 @@
2
  "decoder_start_token_id": 259,
3
  "eos_token_id": 1,
4
  "max_new_tokens": 20,
 
5
  "pad_token_id": 0,
6
  "transformers_version": "4.35.2"
7
  }
 
2
  "decoder_start_token_id": 259,
3
  "eos_token_id": 1,
4
  "max_new_tokens": 20,
5
+ "num_beams": 5,
6
  "pad_token_id": 0,
7
  "transformers_version": "4.35.2"
8
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e443f5f70227c9d7e85031e8221253e22bc5ea8ad04b76abeebdca320a220ae6
3
  size 20350456
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:af5a70ba296b3dc067cca4f7ad8ebddd6b199e562bc55b375df682295a0b1bb4
3
  size 20350456