Hugging Face's logo

juancavallotti
/
roberta-base-culinary-finetuned

Text Classification
Transformers
PyTorch
TensorBoard
bert
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
1
juancavallotti commited on
Commit
b2986f0
·
1 Parent(s): de4c93b

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +91 -7
README.md CHANGED
@@ -1,10 +1,94 @@
1
- # Culinary Roberta
 
 
 
 
 
 
 
 
2
 
3
- I call this `roberta` as I only did language adaptation using MLM.
 
4
 
5
- This model starts from BERT-BASE. I ran a MLM language adaptation stage based on:
6
- * The steps and ingredients of the `recipe_nlg` dataset.
7
- * Food-related books found on project gutenberg.
8
- * Food blogs scraped from the internet.
9
 
10
- After this, I fine-tune it as a sentence-pair classification, based on ingredient synonyms using WordNet.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ tags:
3
+ - generated_from_trainer
4
+ metrics:
5
+ - f1
6
+ model-index:
7
+ - name: roberta-base-culinary-finetuned
8
+ results: []
9
+ ---
10
 
11
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
12
+ should probably proofread and complete it, then remove this comment. -->
13
 
14
+ # roberta-base-culinary-finetuned
 
 
 
15
 
16
+ This model was trained from scratch on the None dataset.
17
+ It achieves the following results on the evaluation set:
18
+ - Loss: 0.0657
19
+ - F1: 0.9929
20
+
21
+ ## Model description
22
+
23
+ More information needed
24
+
25
+ ## Intended uses & limitations
26
+
27
+ More information needed
28
+
29
+ ## Training and evaluation data
30
+
31
+ More information needed
32
+
33
+ ## Training procedure
34
+
35
+ ### Training hyperparameters
36
+
37
+ The following hyperparameters were used during training:
38
+ - learning_rate: 2e-05
39
+ - train_batch_size: 8
40
+ - eval_batch_size: 8
41
+ - seed: 42
42
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
43
+ - lr_scheduler_type: linear
44
+ - num_epochs: 4
45
+
46
+ ### Training results
47
+
48
+ | Training Loss | Epoch | Step | Validation Loss | F1 |
49
+ |:-------------:|:-----:|:-----:|:---------------:|:------:|
50
+ | 0.1803 | 0.11 | 500 | 0.1939 | 0.9611 |
51
+ | 0.1543 | 0.22 | 1000 | 0.1364 | 0.9669 |
52
+ | 0.1213 | 0.32 | 1500 | 0.1487 | 0.9728 |
53
+ | 0.1079 | 0.43 | 2000 | 0.0855 | 0.9773 |
54
+ | 0.0975 | 0.54 | 2500 | 0.0844 | 0.9831 |
55
+ | 0.0855 | 0.65 | 3000 | 0.0785 | 0.9831 |
56
+ | 0.0844 | 0.76 | 3500 | 0.0679 | 0.9857 |
57
+ | 0.0793 | 0.86 | 4000 | 0.0489 | 0.9890 |
58
+ | 0.0864 | 0.97 | 4500 | 0.0399 | 0.9903 |
59
+ | 0.049 | 1.08 | 5000 | 0.0528 | 0.9890 |
60
+ | 0.0353 | 1.19 | 5500 | 0.0635 | 0.9877 |
61
+ | 0.0321 | 1.3 | 6000 | 0.0542 | 0.9903 |
62
+ | 0.0311 | 1.41 | 6500 | 0.0559 | 0.9896 |
63
+ | 0.0315 | 1.51 | 7000 | 0.0736 | 0.9857 |
64
+ | 0.04 | 1.62 | 7500 | 0.0648 | 0.9909 |
65
+ | 0.0265 | 1.73 | 8000 | 0.0608 | 0.9909 |
66
+ | 0.0443 | 1.84 | 8500 | 0.0617 | 0.9883 |
67
+ | 0.0443 | 1.95 | 9000 | 0.0555 | 0.9896 |
68
+ | 0.0235 | 2.05 | 9500 | 0.0608 | 0.9903 |
69
+ | 0.0139 | 2.16 | 10000 | 0.0613 | 0.9922 |
70
+ | 0.0126 | 2.27 | 10500 | 0.0739 | 0.9903 |
71
+ | 0.0164 | 2.38 | 11000 | 0.0679 | 0.9903 |
72
+ | 0.0172 | 2.49 | 11500 | 0.0606 | 0.9922 |
73
+ | 0.0175 | 2.59 | 12000 | 0.0442 | 0.9942 |
74
+ | 0.01 | 2.7 | 12500 | 0.0661 | 0.9916 |
75
+ | 0.0059 | 2.81 | 13000 | 0.0659 | 0.9929 |
76
+ | 0.0216 | 2.92 | 13500 | 0.0504 | 0.9929 |
77
+ | 0.0123 | 3.03 | 14000 | 0.0584 | 0.9929 |
78
+ | 0.0047 | 3.14 | 14500 | 0.0573 | 0.9929 |
79
+ | 0.0123 | 3.24 | 15000 | 0.0511 | 0.9935 |
80
+ | 0.0027 | 3.35 | 15500 | 0.0579 | 0.9942 |
81
+ | 0.0025 | 3.46 | 16000 | 0.0602 | 0.9935 |
82
+ | 0.0051 | 3.57 | 16500 | 0.0598 | 0.9935 |
83
+ | 0.0044 | 3.68 | 17000 | 0.0617 | 0.9929 |
84
+ | 0.0061 | 3.78 | 17500 | 0.0634 | 0.9935 |
85
+ | 0.0048 | 3.89 | 18000 | 0.0672 | 0.9929 |
86
+ | 0.0078 | 4.0 | 18500 | 0.0657 | 0.9929 |
87
+
88
+
89
+ ### Framework versions
90
+
91
+ - Transformers 4.18.0
92
+ - Pytorch 1.11.0+cu113
93
+ - Datasets 2.1.0
94
+ - Tokenizers 0.12.1