Hugging Face's logo

cocorooxinnn
/
eu_adapter01

PEFT
Safetensors
Generated from Trainer
Model card Files Community
cocorooxinnn commited on
Commit
ffc9930
•
1 Parent(s): 530cccc

End of training

Browse files
Files changed (4) hide show
  1. README.md +21 -27
  2. adapter_config.json +2 -2
  3. adapter_model.safetensors +2 -2
  4. training_args.bin +1 -1
README.md CHANGED
@@ -19,13 +19,13 @@ should probably proofread and complete it, then remove this comment. -->
19
 
20
  # eu_adapter01
21
 
22
- This model is a fine-tuned lora adapter of [facebook/bart-large-mnli](https://huggingface.co/facebook/bart-large-mnli) with a single neuron classification head for binary classification on the EU taxonomie dataset.
23
  It achieves the following results on the evaluation set:
24
- - Loss: 0.1792
25
- - F1: 0.9346
26
- - Precision: 0.9199
27
- - Recall: 0.9499
28
- - Accuracy: 0.9336
29
 
30
  ## Model description
31
 
@@ -57,27 +57,21 @@ The following hyperparameters were used during training:
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | F1 | Precision | Recall | Accuracy |
59
  |:-------------:|:------:|:----:|:---------------:|:------:|:---------:|:------:|:--------:|
60
- | 0.4659 | 0.0933 | 50 | 0.3040 | 0.8600 | 0.8937 | 0.8288 | 0.8652 |
61
- | 0.3204 | 0.1866 | 100 | 0.2666 | 0.8818 | 0.9196 | 0.8470 | 0.8865 |
62
- | 0.3079 | 0.2799 | 150 | 0.2509 | 0.9094 | 0.8806 | 0.9401 | 0.9064 |
63
- | 0.2854 | 0.3731 | 200 | 0.2419 | 0.9133 | 0.8813 | 0.9477 | 0.9101 |
64
- | 0.2801 | 0.4664 | 250 | 0.2457 | 0.8902 | 0.9251 | 0.8579 | 0.8943 |
65
- | 0.2722 | 0.5597 | 300 | 0.2344 | 0.9072 | 0.9219 | 0.8930 | 0.9087 |
66
- | 0.2668 | 0.6530 | 350 | 0.2156 | 0.9221 | 0.9027 | 0.9423 | 0.9204 |
67
- | 0.265 | 0.7463 | 400 | 0.2160 | 0.9117 | 0.9286 | 0.8955 | 0.9133 |
68
- | 0.2439 | 0.8396 | 450 | 0.2017 | 0.9240 | 0.9144 | 0.9338 | 0.9233 |
69
- | 0.2253 | 0.9328 | 500 | 0.2043 | 0.9305 | 0.9059 | 0.9564 | 0.9286 |
70
- | 0.2411 | 1.0261 | 550 | 0.2170 | 0.9217 | 0.9254 | 0.9181 | 0.9220 |
71
- | 0.2236 | 1.1194 | 600 | 0.1978 | 0.9308 | 0.9104 | 0.9521 | 0.9292 |
72
- | 0.2095 | 1.2127 | 650 | 0.1884 | 0.9277 | 0.9213 | 0.9341 | 0.9272 |
73
- | 0.2149 | 1.3060 | 700 | 0.1881 | 0.9323 | 0.9197 | 0.9453 | 0.9314 |
74
- | 0.1823 | 1.3993 | 750 | 0.1931 | 0.9297 | 0.9253 | 0.9341 | 0.9294 |
75
- | 0.2052 | 1.4925 | 800 | 0.1838 | 0.9327 | 0.9193 | 0.9464 | 0.9317 |
76
- | 0.199 | 1.5858 | 850 | 0.1836 | 0.9313 | 0.9269 | 0.9357 | 0.9310 |
77
- | 0.1978 | 1.6791 | 900 | 0.1861 | 0.9346 | 0.9132 | 0.9570 | 0.9331 |
78
- | 0.2024 | 1.7724 | 950 | 0.1832 | 0.9349 | 0.9203 | 0.9499 | 0.9339 |
79
- | 0.1861 | 1.8657 | 1000 | 0.1818 | 0.9353 | 0.9204 | 0.9507 | 0.9343 |
80
- | 0.2032 | 1.9590 | 1050 | 0.1792 | 0.9346 | 0.9199 | 0.9499 | 0.9336 |
81
 
82
 
83
  ### Framework versions
 
19
 
20
  # eu_adapter01
21
 
22
+ This model is a fine-tuned version of [facebook/bart-large-mnli](https://huggingface.co/facebook/bart-large-mnli) on an unknown dataset.
23
  It achieves the following results on the evaluation set:
24
+ - Loss: 0.5797
25
+ - F1: 0.8211
26
+ - Precision: 0.8667
27
+ - Recall: 0.78
28
+ - Accuracy: 0.83
29
 
30
  ## Model description
31
 
 
57
 
58
  | Training Loss | Epoch | Step | Validation Loss | F1 | Precision | Recall | Accuracy |
59
  |:-------------:|:------:|:----:|:---------------:|:------:|:---------:|:------:|:--------:|
60
+ | 0.5858 | 0.1266 | 10 | 0.5755 | 0.8627 | 0.8462 | 0.88 | 0.86 |
61
+ | 0.583 | 0.2532 | 20 | 0.5953 | 0.7294 | 0.8857 | 0.62 | 0.77 |
62
+ | 0.5986 | 0.3797 | 30 | 0.5831 | 0.8085 | 0.8636 | 0.76 | 0.82 |
63
+ | 0.5829 | 0.5063 | 40 | 0.5860 | 0.8571 | 0.8182 | 0.9 | 0.85 |
64
+ | 0.5975 | 0.6329 | 50 | 0.5814 | 0.7865 | 0.8974 | 0.7 | 0.81 |
65
+ | 0.5881 | 0.7595 | 60 | 0.5745 | 0.8654 | 0.8333 | 0.9 | 0.86 |
66
+ | 0.5712 | 0.8861 | 70 | 0.5878 | 0.7957 | 0.8605 | 0.74 | 0.81 |
67
+ | 0.5536 | 1.0127 | 80 | 0.5853 | 0.8085 | 0.8636 | 0.76 | 0.82 |
68
+ | 0.5726 | 1.1392 | 90 | 0.5766 | 0.8485 | 0.8571 | 0.84 | 0.85 |
69
+ | 0.5843 | 1.2658 | 100 | 0.5762 | 0.8485 | 0.8571 | 0.84 | 0.85 |
70
+ | 0.5704 | 1.3924 | 110 | 0.5761 | 0.8298 | 0.8864 | 0.78 | 0.84 |
71
+ | 0.5714 | 1.5190 | 120 | 0.5860 | 0.7727 | 0.8947 | 0.68 | 0.8 |
72
+ | 0.5688 | 1.6456 | 130 | 0.5815 | 0.8132 | 0.9024 | 0.74 | 0.83 |
73
+ | 0.5723 | 1.7722 | 140 | 0.5760 | 0.8298 | 0.8864 | 0.78 | 0.84 |
74
+ | 0.5716 | 1.8987 | 150 | 0.5797 | 0.8211 | 0.8667 | 0.78 | 0.83 |
 
 
 
 
 
 
75
 
76
 
77
  ### Framework versions
adapter_config.json CHANGED
@@ -25,9 +25,9 @@
25
  "rank_pattern": {},
26
  "revision": null,
27
  "target_modules": [
28
- "v_proj",
29
  "q_proj",
30
- "k_proj"
 
31
  ],
32
  "task_type": null,
33
  "use_dora": false,
 
25
  "rank_pattern": {},
26
  "revision": null,
27
  "target_modules": [
 
28
  "q_proj",
29
+ "k_proj",
30
+ "v_proj"
31
  ],
32
  "task_type": null,
33
  "use_dora": false,
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:5ea6b28c658a6a3b2526e6e0e17a2645565559106e631f883959e63d84c1d08f
3
- size 29715788
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7239a9513b1fde269ea07d925ba3df87e6d54b13cbd8c1d42f516117d1fa9fce
3
+ size 29720112
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:553cba4ef04a076021946167468fa502af9e78fbfde5f65eee9dbdc66997d486
3
  size 5048
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c2dd5944482d6aefe3819d3a9d0fc3c9812102c6a276cf680f1d98132479429
3
  size 5048