Hugging Face's logo

Francois2511
/
cuad_qa_model

Question Answering
Transformers
TensorBoard
Safetensors
modernbert
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
Francois2511 commited on
Commit
2093a0b
路
verified 路
1 Parent(s): 38d5ed8

End of training

Browse files
Files changed (1) hide show
  1. README.md +16 -44
README.md CHANGED
@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
18
 
19
  This model is a fine-tuned version of [answerdotai/ModernBERT-base](https://huggingface.co/answerdotai/ModernBERT-base) on the cuad-qa dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 56.3253
22
- - Jaccard: 0.1325
23
 
24
  ## Model description
25
 
@@ -38,62 +38,34 @@ More information needed
38
  ### Training hyperparameters
39
 
40
  The following hyperparameters were used during training:
41
- - learning_rate: 2e-05
42
  - train_batch_size: 3
43
  - eval_batch_size: 4
44
  - seed: 42
45
  - gradient_accumulation_steps: 4
46
  - total_train_batch_size: 12
47
- - optimizer: Use adamw_torch with betas=(0.9,0.98) and epsilon=1e-06 and optimizer_args=No additional optimizer arguments
48
  - lr_scheduler_type: linear
49
- - num_epochs: 4
50
 
51
  ### Training results
52
 
53
  | Training Loss | Epoch | Step | Validation Loss | Jaccard |
54
  |:-------------:|:------:|:----:|:---------------:|:-------:|
55
- | 2931.2947 | 0.1075 | 100 | 125.3868 | 0.0261 |
56
- | 114.0476 | 0.2149 | 200 | 98.0385 | 0.0225 |
57
- | 92.3046 | 0.3224 | 300 | 86.1094 | 0.0279 |
58
- | 83.4547 | 0.4299 | 400 | 80.0709 | 0.0403 |
59
- | 80.4591 | 0.5373 | 500 | 75.2658 | 0.0433 |
60
- | 76.238 | 0.6448 | 600 | 71.9617 | 0.0445 |
61
- | 73.2576 | 0.7523 | 700 | 68.1718 | 0.0463 |
62
- | 70.5061 | 0.8598 | 800 | 64.2118 | 0.0536 |
63
- | 72.0594 | 0.9672 | 900 | 82.5902 | 0.0243 |
64
- | 65.2249 | 1.0742 | 1000 | 59.8434 | 0.0647 |
65
- | 63.2437 | 1.1816 | 1100 | 60.3719 | 0.0932 |
66
- | 67.1502 | 1.2891 | 1200 | 63.5264 | 0.1114 |
67
- | 65.1003 | 1.3966 | 1300 | 60.7845 | 0.1243 |
68
- | 64.7538 | 1.5040 | 1400 | 66.3558 | 0.1200 |
69
- | 66.7688 | 1.6115 | 1500 | 69.2212 | 0.1149 |
70
- | 76.4721 | 1.7190 | 1600 | 69.5449 | 0.1458 |
71
- | 82.2733 | 1.8264 | 1700 | 82.1182 | 0.0449 |
72
- | 78.7475 | 1.9339 | 1800 | 62.4942 | 0.1581 |
73
- | 69.5967 | 2.0408 | 1900 | 63.3104 | 0.1507 |
74
- | 67.6753 | 2.1483 | 2000 | 56.4553 | 0.2238 |
75
- | 64.0365 | 2.2558 | 2100 | 60.3552 | 0.1978 |
76
- | 62.561 | 2.3632 | 2200 | 55.5222 | 0.2238 |
77
- | 62.0848 | 2.4707 | 2300 | 51.5148 | 0.2239 |
78
- | 59.3192 | 2.5782 | 2400 | 56.1338 | 0.1939 |
79
- | 63.3072 | 2.6857 | 2500 | 55.3624 | 0.2385 |
80
- | 63.0132 | 2.7931 | 2600 | 48.8478 | 0.2614 |
81
- | 61.0742 | 2.9006 | 2700 | 57.2687 | 0.2574 |
82
- | 63.7064 | 3.0075 | 2800 | 58.7552 | 0.2569 |
83
- | 61.3371 | 3.1150 | 2900 | 62.7214 | 0.2473 |
84
- | 66.2795 | 3.2225 | 3000 | 60.0179 | 0.2640 |
85
- | 65.9729 | 3.3299 | 3100 | 59.7260 | 0.2879 |
86
- | 67.5846 | 3.4374 | 3200 | 63.1864 | 0.2627 |
87
- | 65.6924 | 3.5449 | 3300 | 58.8332 | 0.2743 |
88
- | 64.2456 | 3.6523 | 3400 | 59.7355 | 0.1667 |
89
- | 64.9793 | 3.7598 | 3500 | 57.0126 | 0.1622 |
90
- | 63.8452 | 3.8673 | 3600 | 56.8423 | 0.1332 |
91
- | 65.2058 | 3.9747 | 3700 | 56.3253 | 0.1325 |
92
 
93
 
94
  ### Framework versions
95
 
96
- - Transformers 4.48.2
97
- - Pytorch 2.2.1+cu121
98
  - Datasets 3.2.0
99
  - Tokenizers 0.21.0
 
18
 
19
  This model is a fine-tuned version of [answerdotai/ModernBERT-base](https://huggingface.co/answerdotai/ModernBERT-base) on the cuad-qa dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 1.2765
22
+ - Jaccard: 0.3298
23
 
24
  ## Model description
25
 
 
38
  ### Training hyperparameters
39
 
40
  The following hyperparameters were used during training:
41
+ - learning_rate: 8e-05
42
  - train_batch_size: 3
43
  - eval_batch_size: 4
44
  - seed: 42
45
  - gradient_accumulation_steps: 4
46
  - total_train_batch_size: 12
47
+ - optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.98) and epsilon=1e-06 and optimizer_args=No additional optimizer arguments
48
  - lr_scheduler_type: linear
49
+ - num_epochs: 3
50
 
51
  ### Training results
52
 
53
  | Training Loss | Epoch | Step | Validation Loss | Jaccard |
54
  |:-------------:|:------:|:----:|:---------------:|:-------:|
55
+ | 4.0374 | 0.3224 | 300 | 2.0316 | 0.3097 |
56
+ | 2.2477 | 0.6448 | 600 | 1.4555 | 0.3400 |
57
+ | 2.0917 | 0.9672 | 900 | 1.3932 | 0.3600 |
58
+ | 1.7293 | 1.2891 | 1200 | 1.2932 | 0.3288 |
59
+ | 1.709 | 1.6115 | 1500 | 1.2642 | 0.3280 |
60
+ | 1.7021 | 1.9339 | 1800 | 1.2142 | 0.3456 |
61
+ | 1.3928 | 2.2558 | 2100 | 1.2952 | 0.3295 |
62
+ | 1.3582 | 2.5782 | 2400 | 1.2791 | 0.3298 |
63
+ | 1.3509 | 2.9006 | 2700 | 1.2765 | 0.3298 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
64
 
65
 
66
  ### Framework versions
67
 
68
+ - Transformers 4.49.0.dev0
69
+ - Pytorch 2.4.1+cu124
70
  - Datasets 3.2.0
71
  - Tokenizers 0.21.0