christinacdl
commited on
Commit
•
b97c019
1
Parent(s):
2fc9414
Update README.md
Browse files
README.md
CHANGED
@@ -5,9 +5,23 @@ tags:
|
|
5 |
- generated_from_trainer
|
6 |
metrics:
|
7 |
- accuracy
|
|
|
|
|
|
|
8 |
model-index:
|
9 |
- name: XLM_RoBERTa-Clickbait-Detection-new
|
10 |
results: []
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
---
|
12 |
|
13 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
@@ -15,24 +29,27 @@ should probably proofread and complete it, then remove this comment. -->
|
|
15 |
|
16 |
# XLM_RoBERTa-Clickbait-Detection-new
|
17 |
|
18 |
-
This model is a fine-tuned version of [xlm-roberta-large](https://huggingface.co/xlm-roberta-large) on the
|
19 |
It achieves the following results on the evaluation set:
|
20 |
- Loss: 0.1071
|
21 |
- Micro F1: 0.9834
|
22 |
- Macro F1: 0.9833
|
23 |
- Accuracy: 0.9834
|
24 |
|
25 |
-
|
|
|
|
|
|
|
|
|
26 |
|
27 |
-
|
|
|
|
|
28 |
|
29 |
## Intended uses & limitations
|
30 |
|
31 |
More information needed
|
32 |
|
33 |
-
## Training and evaluation data
|
34 |
-
|
35 |
-
More information needed
|
36 |
|
37 |
## Training procedure
|
38 |
|
@@ -48,9 +65,47 @@ The following hyperparameters were used during training:
|
|
48 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
49 |
- lr_scheduler_type: linear
|
50 |
- num_epochs: 4
|
51 |
-
|
52 |
-
|
53 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
|
55 |
|
56 |
### Framework versions
|
@@ -58,4 +113,4 @@ The following hyperparameters were used during training:
|
|
58 |
- Transformers 4.36.1
|
59 |
- Pytorch 2.1.0+cu121
|
60 |
- Datasets 2.13.1
|
61 |
-
- Tokenizers 0.15.0
|
|
|
5 |
- generated_from_trainer
|
6 |
metrics:
|
7 |
- accuracy
|
8 |
+
- f1
|
9 |
+
- recall
|
10 |
+
- precision
|
11 |
model-index:
|
12 |
- name: XLM_RoBERTa-Clickbait-Detection-new
|
13 |
results: []
|
14 |
+
datasets:
|
15 |
+
- christinacdl/clickbait_detection_dataset
|
16 |
+
language:
|
17 |
+
- en
|
18 |
+
- el
|
19 |
+
- ru
|
20 |
+
- ro
|
21 |
+
- de
|
22 |
+
- it
|
23 |
+
- es
|
24 |
+
pipeline_tag: text-classification
|
25 |
---
|
26 |
|
27 |
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
|
|
29 |
|
30 |
# XLM_RoBERTa-Clickbait-Detection-new
|
31 |
|
32 |
+
This model is a fine-tuned version of [xlm-roberta-large](https://huggingface.co/xlm-roberta-large) on the christinacdl/clickbait_detection_dataset dataset.
|
33 |
It achieves the following results on the evaluation set:
|
34 |
- Loss: 0.1071
|
35 |
- Micro F1: 0.9834
|
36 |
- Macro F1: 0.9833
|
37 |
- Accuracy: 0.9834
|
38 |
|
39 |
+
It achieves the following results on the test set:
|
40 |
+
- Accuracy: 0.9838922630050172
|
41 |
+
- Micro-F1 Score: 0.9838922630050172
|
42 |
+
- Macro-F1 Score: 0.9838416247418498
|
43 |
+
- Matthews Correlation Coefficient: 0.9676867009951606
|
44 |
|
45 |
+
- Precision of each class: [0.98156425 0.98597897]
|
46 |
+
- Recall of each class: [0.98431373 0.98351648]
|
47 |
+
- F1 score of each class: [0.98293706 0.98474619]
|
48 |
|
49 |
## Intended uses & limitations
|
50 |
|
51 |
More information needed
|
52 |
|
|
|
|
|
|
|
53 |
|
54 |
## Training procedure
|
55 |
|
|
|
65 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
66 |
- lr_scheduler_type: linear
|
67 |
- num_epochs: 4
|
68 |
+
- early stopping patience: 2
|
69 |
+
- adam epsilon: 1e-8
|
70 |
+
- gradient_checkpointing: True
|
71 |
+
- max_grad_norm: 1.0
|
72 |
+
- seed: 42
|
73 |
+
- optimizer: adamw_torch_fused
|
74 |
+
- max_steps': -1
|
75 |
+
- warmup_ratio: 0
|
76 |
+
- group_by_length: True
|
77 |
+
- max_seq_length': 512
|
78 |
+
- save_steps: 1000
|
79 |
+
- logging_steps': 500
|
80 |
+
- evaluation_strategy: epoch
|
81 |
+
- save_strategy:epoch
|
82 |
+
- eval_steps: 1000
|
83 |
+
- save_total_limit: 2
|
84 |
+
|
85 |
+
|
86 |
+
### All results from Training and Evaluation
|
87 |
+
- "epoch": 4.0,
|
88 |
+
- "eval_accuracy": 0.9844203855294428,
|
89 |
+
- "eval_loss": 0.08027808368206024,
|
90 |
+
- "eval_macro_f1": 0.9843695357857132,
|
91 |
+
- "eval_micro_f1": 0.9844203855294428,
|
92 |
+
- "eval_runtime": 124.9733,
|
93 |
+
- "eval_samples": 3787,
|
94 |
+
- "eval_samples_per_second": 30.302,
|
95 |
+
- "eval_steps_per_second": 1.896,
|
96 |
+
- "predict_accuracy": 0.9838922630050172,
|
97 |
+
- "predict_loss": 0.07716809958219528,
|
98 |
+
- "predict_macro_f1": 0.9838416247418498,
|
99 |
+
- "predict_micro_f1": 0.9838922630050172,
|
100 |
+
- "predict_runtime": 127.7861,
|
101 |
+
- "predict_samples": 3787,
|
102 |
+
- "predict_samples_per_second": 29.635,
|
103 |
+
- "predict_steps_per_second": 1.855,
|
104 |
+
- "train_loss": 0.057462599486458765,
|
105 |
+
- "train_runtime": 25253.576,
|
106 |
+
- "train_samples": 30296,
|
107 |
+
- "train_samples_per_second": 4.799,
|
108 |
+
- "train_steps_per_second": 0.15
|
109 |
|
110 |
|
111 |
### Framework versions
|
|
|
113 |
- Transformers 4.36.1
|
114 |
- Pytorch 2.1.0+cu121
|
115 |
- Datasets 2.13.1
|
116 |
+
- Tokenizers 0.15.0
|