Hugging Face's logo

thrunlab
/
t5-large_cola_dense_sp0_ar0

Text Classification
Transformers
PyTorch
t5
Generated from Trainer
Eval Results
text-generation-inference
Inference Endpoints
Model card Files Community
lukeleeai commited on
Commit
d32c360
·
1 Parent(s): 6d1c416

End of training

Browse files
Files changed (1) hide show
  1. README.md +31 -29
README.md CHANGED
@@ -32,7 +32,7 @@ should probably proofread and complete it, then remove this comment. -->
32
 
33
  This model is a fine-tuned version of [t5-large](https://huggingface.co/t5-large) on the glue dataset.
34
  It achieves the following results on the evaluation set:
35
- - Loss: 4.7159
36
  - Accuracy: 0.0
37
 
38
  ## Model description
@@ -65,34 +65,36 @@ The following hyperparameters were used during training:
65
 
66
  | Training Loss | Epoch | Step | Validation Loss | Accuracy |
67
  |:-------------:|:-----:|:----:|:---------------:|:--------:|
68
- | 0.6148 | 0.05 | 25 | 0.6325 | 0.6913 |
69
- | 0.4516 | 0.09 | 50 | 0.5597 | 0.6932 |
70
- | 0.3091 | 0.14 | 75 | 0.6187 | 0.7996 |
71
- | 0.6072 | 0.19 | 100 | 0.5791 | 0.7776 |
72
- | 0.4384 | 0.23 | 125 | 0.4396 | 0.8035 |
73
- | 0.3521 | 0.28 | 150 | 0.4528 | 0.8092 |
74
- | 0.2873 | 0.33 | 175 | 0.4780 | 0.8169 |
75
- | 0.2986 | 0.37 | 200 | 0.5905 | 0.7996 |
76
- | 0.3324 | 0.42 | 225 | 0.4971 | 0.8150 |
77
- | 0.4326 | 0.47 | 250 | 0.4197 | 0.8265 |
78
- | 0.3141 | 0.51 | 275 | 0.4582 | 0.8178 |
79
- | 0.5325 | 0.56 | 300 | 0.4662 | 0.8245 |
80
- | 0.3367 | 0.61 | 325 | 0.4346 | 0.8217 |
81
- | 0.3546 | 0.65 | 350 | 0.4180 | 0.8245 |
82
- | 0.3396 | 0.7 | 375 | 0.3809 | 0.8236 |
83
- | 0.3023 | 0.75 | 400 | 0.3908 | 0.8245 |
84
- | 0.3148 | 0.79 | 425 | 0.4312 | 0.8265 |
85
- | 0.3596 | 0.84 | 450 | 0.3919 | 0.8389 |
86
- | 0.2059 | 0.89 | 475 | 0.5070 | 0.8322 |
87
- | 0.4051 | 0.93 | 500 | 0.4253 | 0.8274 |
88
- | 0.2948 | 0.98 | 525 | 0.4947 | 0.8313 |
89
- | 0.2357 | 1.03 | 550 | 0.6221 | 0.8188 |
90
- | 0.3183 | 1.07 | 575 | 0.6693 | 0.8178 |
91
- | 0.4272 | 1.12 | 600 | 0.5816 | 0.8236 |
92
- | 0.1839 | 1.17 | 625 | 0.6645 | 0.8207 |
93
- | 0.3269 | 1.21 | 650 | 0.7249 | 0.8188 |
94
- | 0.2694 | 1.26 | 675 | 0.5396 | 0.8313 |
95
- | 0.2222 | 1.31 | 700 | 0.4576 | 0.8245 |
 
 
96
 
97
 
98
  ### Framework versions
 
32
 
33
  This model is a fine-tuned version of [t5-large](https://huggingface.co/t5-large) on the glue dataset.
34
  It achieves the following results on the evaluation set:
35
+ - Loss: 4.7611
36
  - Accuracy: 0.0
37
 
38
  ## Model description
 
65
 
66
  | Training Loss | Epoch | Step | Validation Loss | Accuracy |
67
  |:-------------:|:-----:|:----:|:---------------:|:--------:|
68
+ | 0.6121 | 0.05 | 25 | 0.6257 | 0.6913 |
69
+ | 0.4507 | 0.09 | 50 | 0.6018 | 0.6913 |
70
+ | 0.2862 | 0.14 | 75 | 0.5646 | 0.8006 |
71
+ | 0.5917 | 0.19 | 100 | 0.5203 | 0.7929 |
72
+ | 0.3317 | 0.23 | 125 | 0.4479 | 0.8236 |
73
+ | 0.3637 | 0.28 | 150 | 0.4355 | 0.8245 |
74
+ | 0.2844 | 0.33 | 175 | 0.5032 | 0.8245 |
75
+ | 0.3406 | 0.37 | 200 | 0.5102 | 0.8121 |
76
+ | 0.4321 | 0.42 | 225 | 0.4290 | 0.8150 |
77
+ | 0.5212 | 0.47 | 250 | 0.4134 | 0.8293 |
78
+ | 0.4152 | 0.51 | 275 | 0.5055 | 0.8207 |
79
+ | 0.453 | 0.56 | 300 | 0.3974 | 0.8265 |
80
+ | 0.3412 | 0.61 | 325 | 0.4409 | 0.8245 |
81
+ | 0.3251 | 0.65 | 350 | 0.4538 | 0.8255 |
82
+ | 0.3255 | 0.7 | 375 | 0.3817 | 0.8313 |
83
+ | 0.2671 | 0.75 | 400 | 0.4162 | 0.8255 |
84
+ | 0.3995 | 0.79 | 425 | 0.4150 | 0.8303 |
85
+ | 0.4005 | 0.84 | 450 | 0.4125 | 0.8303 |
86
+ | 0.2897 | 0.89 | 475 | 0.4895 | 0.8226 |
87
+ | 0.4079 | 0.93 | 500 | 0.4064 | 0.8351 |
88
+ | 0.2597 | 0.98 | 525 | 0.6631 | 0.8447 |
89
+ | 0.2189 | 1.03 | 550 | 0.5056 | 0.8236 |
90
+ | 0.329 | 1.07 | 575 | 6.1282 | 0.8284 |
91
+ | 0.44 | 1.12 | 600 | 0.5057 | 0.8380 |
92
+ | 0.164 | 1.17 | 625 | 0.5032 | 0.8313 |
93
+ | 0.2996 | 1.21 | 650 | 0.9884 | 0.8341 |
94
+ | 0.2425 | 1.26 | 675 | 0.5208 | 0.8418 |
95
+ | 0.1987 | 1.31 | 700 | 0.4573 | 0.8389 |
96
+ | 0.1581 | 1.36 | 725 | 1.1812 | 0.8150 |
97
+ | 0.4067 | 1.4 | 750 | 0.6437 | 0.8293 |
98
 
99
 
100
  ### Framework versions