Hugging Face's logo

Michelvh
/
t5-end2end-questions-generation-dutch

Text2Text Generation
Transformers
PyTorch
t5
Generated from Trainer
text-generation-inference
Inference Endpoints
Model card Files Community
3
Michelvh commited on
Commit
f425c1c
•
1 Parent(s): fc30114

Revert to 23ca00c

Browse files
Files changed (4) hide show
  1. README.md +34 -20
  2. config.json +1 -1
  3. pytorch_model.bin +1 -1
  4. training_args.bin +1 -1
README.md CHANGED
@@ -15,18 +15,22 @@ inference:
15
  early_stopping: True
16
  ---
17
 
18
- <!-- This model card has been generated automatically according to the information the Trainer had access to. You
19
- should probably proofread and complete it, then remove this comment. -->
20
 
21
  # t5-end2end-questions-generation-dutch
22
 
23
- This model is a fine-tuned version of [Michelvh/t5-end2end-questions-generation-dutch](https://huggingface.co/Michelvh/t5-end2end-questions-generation-dutch) on an unknown dataset.
 
 
 
 
 
24
  It achieves the following results on the evaluation set:
25
- - Loss: 1.6715
26
 
27
  ## Model description
28
 
29
- More information needed
30
 
31
  ## Intended uses & limitations
32
 
@@ -41,12 +45,12 @@ More information needed
41
  ### Training hyperparameters
42
 
43
  The following hyperparameters were used during training:
44
- - learning_rate: 5e-05
45
- - train_batch_size: 8
46
- - eval_batch_size: 8
47
  - seed: 42
48
  - gradient_accumulation_steps: 16
49
- - total_train_batch_size: 128
50
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
51
  - lr_scheduler_type: linear
52
  - num_epochs: 7
@@ -55,16 +59,26 @@ The following hyperparameters were used during training:
55
 
56
  | Training Loss | Epoch | Step | Validation Loss |
57
  |:-------------:|:-----:|:----:|:---------------:|
58
- | 1.7309 | 0.68 | 100 | 1.7055 |
59
- | 1.6965 | 1.35 | 200 | 1.6985 |
60
- | 1.6758 | 2.03 | 300 | 1.6900 |
61
- | 1.6514 | 2.71 | 400 | 1.6842 |
62
- | 1.6412 | 3.39 | 500 | 1.6812 |
63
- | 1.631 | 4.06 | 600 | 1.6765 |
64
- | 1.6228 | 4.74 | 700 | 1.6742 |
65
- | 1.6145 | 5.42 | 800 | 1.6737 |
66
- | 1.612 | 6.1 | 900 | 1.6710 |
67
- | 1.6082 | 6.77 | 1000 | 1.6715 |
 
 
 
 
 
 
 
 
 
 
68
 
69
 
70
  ### Framework versions
@@ -72,4 +86,4 @@ The following hyperparameters were used during training:
72
  - Transformers 4.27.4
73
  - Pytorch 1.13.0
74
  - Datasets 2.1.0
75
- - Tokenizers 0.13.2
 
15
  early_stopping: True
16
  ---
17
 
18
+
 
19
 
20
  # t5-end2end-questions-generation-dutch
21
 
22
+ This model is a fine-tuned version of [yhavinga/t5-base-dutch](https://huggingface.co/yhavinga/t5-base-dutch) on a
23
+ Google translated version of SQUAD 1.1 found here: https://www.kaggle.com/datasets/michelvanheijningen/squad1-dutch.
24
+
25
+ The code used to finetune the model is largely based on the work by Thomas Simonini. You can find his English model
26
+ [here](https://huggingface.co/ThomasSimonini/t5-end2end-question-generation) and his Google colab tutorial [here](https://colab.research.google.com/drive/1z-Zl2hftMrFXabYfmz8o9YZpgYx6sGeW?usp=sharing)
27
+
28
  It achieves the following results on the evaluation set:
29
+ - Loss: 1.6546
30
 
31
  ## Model description
32
 
33
+ This is my first model ever and still a work in progress ;)
34
 
35
  ## Intended uses & limitations
36
 
 
45
  ### Training hyperparameters
46
 
47
  The following hyperparameters were used during training:
48
+ - learning_rate: 0.0001
49
+ - train_batch_size: 4
50
+ - eval_batch_size: 4
51
  - seed: 42
52
  - gradient_accumulation_steps: 16
53
+ - total_train_batch_size: 64
54
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
55
  - lr_scheduler_type: linear
56
  - num_epochs: 7
 
59
 
60
  | Training Loss | Epoch | Step | Validation Loss |
61
  |:-------------:|:-----:|:----:|:---------------:|
62
+ | 2.6528 | 0.34 | 100 | 1.9249 |
63
+ | 1.964 | 0.68 | 200 | 1.7897 |
64
+ | 1.8695 | 1.02 | 300 | 1.7554 |
65
+ | 1.7922 | 1.35 | 400 | 1.7270 |
66
+ | 1.7747 | 1.69 | 500 | 1.7054 |
67
+ | 1.7473 | 2.03 | 600 | 1.7019 |
68
+ | 1.697 | 2.37 | 700 | 1.6868 |
69
+ | 1.6848 | 2.71 | 800 | 1.6810 |
70
+ | 1.6756 | 3.05 | 900 | 1.6779 |
71
+ | 1.6282 | 3.39 | 1000 | 1.6712 |
72
+ | 1.6285 | 3.73 | 1100 | 1.6626 |
73
+ | 1.6161 | 4.06 | 1200 | 1.6616 |
74
+ | 1.5887 | 4.4 | 1300 | 1.6588 |
75
+ | 1.5877 | 4.74 | 1400 | 1.6583 |
76
+ | 1.5723 | 5.08 | 1500 | 1.6560 |
77
+ | 1.5545 | 5.42 | 1600 | 1.6550 |
78
+ | 1.5415 | 5.76 | 1700 | 1.6540 |
79
+ | 1.5509 | 6.1 | 1800 | 1.6541 |
80
+ | 1.5326 | 6.44 | 1900 | 1.6539 |
81
+ | 1.5268 | 6.77 | 2000 | 1.6546 |
82
 
83
 
84
  ### Framework versions
 
86
  - Transformers 4.27.4
87
  - Pytorch 1.13.0
88
  - Datasets 2.1.0
89
+ - Tokenizers 0.13.2
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "Michelvh/t5-end2end-questions-generation-dutch",
3
  "architectures": [
4
  "T5ForConditionalGeneration"
5
  ],
 
1
  {
2
+ "_name_or_path": "yhavinga/t5-base-dutch",
3
  "architectures": [
4
  "T5ForConditionalGeneration"
5
  ],
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b6b86565dfa5355980a84386c4c3db84937e6c13882f0f2978119ae29e90d70a
3
  size 891629201
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0caa777b800a6cfa9fe3d1e43de59426d5f38c088cf0d6503def33109a95e88f
3
  size 891629201
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:44f512b3252a1abb43cc1dfd90a29b055052f4244e0433bf6780884c4c2708a3
3
  size 3643
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad7f07b8448b73ae722318b05b7e3e6d093fea002144900b7d782e152599968c
3
  size 3643