File size: 10,612 Bytes
bd5c065
dc10fea
bd5c065
 
 
 
 
dc10fea
5bf5666
 
bd5c065
 
5bf5666
 
 
 
 
dc10fea
 
5bf5666
1293db3
5bf5666
 
 
 
dc10fea
bd5c065
 
 
 
 
 
 
dc10fea
5bf5666
dc10fea
 
 
 
 
 
 
 
 
 
 
bd5c065
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
dc10fea
bd5c065
 
 
dc10fea
 
 
 
 
 
 
bd5c065
 
 
 
dc10fea
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
---
library_name: transformers
license: apache-2.0
base_model: allenai/longformer-base-4096
tags:
- generated_from_trainer
datasets:
- stab-gurevych-essays
metrics:
- accuracy
model-index:
- name: longformer-full_labels
  results:
  - task:
      name: Token Classification
      type: token-classification
    dataset:
      name: stab-gurevych-essays
      type: stab-gurevych-essays
      config: full_labels
      split: train[0%:20%]
      args: full_labels
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.8572502648576603
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# longformer-full_labels

This model is a fine-tuned version of [allenai/longformer-base-4096](https://huggingface.co/allenai/longformer-base-4096) on the stab-gurevych-essays dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3818
- B-claim: {'precision': 0.5588235294117647, 'recall': 0.46830985915492956, 'f1-score': 0.5095785440613027, 'support': 284.0}
- B-majorclaim: {'precision': 0.8787878787878788, 'recall': 0.20567375886524822, 'f1-score': 0.3333333333333333, 'support': 141.0}
- B-premise: {'precision': 0.7287735849056604, 'recall': 0.8728813559322034, 'f1-score': 0.794344473007712, 'support': 708.0}
- I-claim: {'precision': 0.6021926389976507, 'recall': 0.5673880964092474, 'f1-score': 0.5842725085475498, 'support': 4066.0}
- I-majorclaim: {'precision': 0.7885196374622356, 'recall': 0.7767857142857143, 'f1-score': 0.782608695652174, 'support': 2016.0}
- I-premise: {'precision': 0.8760707709550877, 'recall': 0.8973349733497334, 'f1-score': 0.8865753868589484, 'support': 12195.0}
- O: {'precision': 0.9648159446817165, 'recall': 0.9631509491422191, 'f1-score': 0.9639827279654559, 'support': 9851.0}
- Accuracy: 0.8573
- Macro avg: {'precision': 0.7711405693145706, 'recall': 0.6787892438770422, 'f1-score': 0.693527952775211, 'support': 29261.0}
- Weighted avg: {'precision': 0.8552285449410628, 'recall': 0.8572502648576603, 'f1-score': 0.8549088561404111, 'support': 29261.0}

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5

### Training results

| Training Loss | Epoch | Step | Validation Loss | B-claim                                                                                                             | B-majorclaim                                                                                                       | B-premise                                                                                                           | I-claim                                                                                                               | I-majorclaim                                                                                                       | I-premise                                                                                                           | O                                                                                                                  | Accuracy | Macro avg                                                                                                           | Weighted avg                                                                                                        |
|:-------------:|:-----:|:----:|:---------------:|:-------------------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------------------:|:---------------------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------------------:|:--------:|:-------------------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------------------:|
| No log        | 1.0   | 41   | 0.7363          | {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 284.0}                                                | {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 141.0}                                               | {'precision': 0.7931034482758621, 'recall': 0.06497175141242938, 'f1-score': 0.12010443864229765, 'support': 708.0} | {'precision': 0.35688405797101447, 'recall': 0.09690113133300542, 'f1-score': 0.15241779497098645, 'support': 4066.0} | {'precision': 0.4854771784232365, 'recall': 0.3482142857142857, 'f1-score': 0.4055459272097054, 'support': 2016.0} | {'precision': 0.7254034519284691, 'recall': 0.9546535465354653, 'f1-score': 0.8243874805268375, 'support': 12195.0} | {'precision': 0.8224254998113919, 'recall': 0.8852908334179271, 'f1-score': 0.8527010510877536, 'support': 9851.0} | 0.7349   | {'precision': 0.4547562337728534, 'recall': 0.3357187926304447, 'f1-score': 0.3364509560625115, 'support': 29261.0} | {'precision': 0.6814305221181916, 'recall': 0.7349372885410614, 'f1-score': 0.6826734788782265, 'support': 29261.0} |
| No log        | 2.0   | 82   | 0.4757          | {'precision': 1.0, 'recall': 0.01056338028169014, 'f1-score': 0.020905923344947737, 'support': 284.0}               | {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 141.0}                                               | {'precision': 0.6255364806866953, 'recall': 0.8234463276836158, 'f1-score': 0.7109756097560975, 'support': 708.0}   | {'precision': 0.5658734764944864, 'recall': 0.4795868175110674, 'f1-score': 0.5191693290734825, 'support': 4066.0}    | {'precision': 0.745417515274949, 'recall': 0.5446428571428571, 'f1-score': 0.6294067067927773, 'support': 2016.0}  | {'precision': 0.8514935768456895, 'recall': 0.9022550225502255, 'f1-score': 0.8761396663614285, 'support': 12195.0} | {'precision': 0.9034811635670005, 'recall': 0.9616282610902447, 'f1-score': 0.931648308418568, 'support': 9851.0}  | 0.8240   | {'precision': 0.6702574589812601, 'recall': 0.5317318094656714, 'f1-score': 0.5268922205353288, 'support': 29261.0} | {'precision': 0.8138696629123667, 'recall': 0.8239636376063703, 'f1-score': 0.8117058591419718, 'support': 29261.0} |
| No log        | 3.0   | 123  | 0.4101          | {'precision': 0.49624060150375937, 'recall': 0.2323943661971831, 'f1-score': 0.31654676258992803, 'support': 284.0} | {'precision': 1.0, 'recall': 0.014184397163120567, 'f1-score': 0.027972027972027972, 'support': 141.0}             | {'precision': 0.6877777777777778, 'recall': 0.8742937853107344, 'f1-score': 0.7699004975124378, 'support': 708.0}   | {'precision': 0.6374125874125874, 'recall': 0.4483521888834235, 'f1-score': 0.5264221773029165, 'support': 4066.0}    | {'precision': 0.7599795291709315, 'recall': 0.7366071428571429, 'f1-score': 0.7481108312342569, 'support': 2016.0} | {'precision': 0.843370836090889, 'recall': 0.9404674046740468, 'f1-score': 0.8892765759478949, 'support': 12195.0}  | {'precision': 0.9602568022011617, 'recall': 0.9565526342503299, 'f1-score': 0.9584011391375101, 'support': 9851.0} | 0.8505   | {'precision': 0.7692911620224437, 'recall': 0.6004074170479973, 'f1-score': 0.6052328588138531, 'support': 29261.0} | {'precision': 0.841977868607838, 'recall': 0.8505177540070401, 'f1-score': 0.8398036418020065, 'support': 29261.0}  |
| No log        | 4.0   | 164  | 0.3859          | {'precision': 0.538135593220339, 'recall': 0.4471830985915493, 'f1-score': 0.48846153846153845, 'support': 284.0}   | {'precision': 1.0, 'recall': 0.10638297872340426, 'f1-score': 0.19230769230769232, 'support': 141.0}               | {'precision': 0.7128146453089245, 'recall': 0.8799435028248588, 'f1-score': 0.7876106194690266, 'support': 708.0}   | {'precision': 0.6014307613694431, 'recall': 0.5789473684210527, 'f1-score': 0.5899749373433584, 'support': 4066.0}    | {'precision': 0.7848036715961244, 'recall': 0.7633928571428571, 'f1-score': 0.7739502137289415, 'support': 2016.0} | {'precision': 0.8792672100718263, 'recall': 0.8933989339893399, 'f1-score': 0.8862767428617913, 'support': 12195.0} | {'precision': 0.9612968591691996, 'recall': 0.9631509491422191, 'f1-score': 0.9622230110034988, 'support': 9851.0} | 0.8558   | {'precision': 0.7825355343908367, 'recall': 0.6617713841193258, 'f1-score': 0.6686863935965496, 'support': 29261.0} | {'precision': 0.8550112416363399, 'recall': 0.855780732032398, 'f1-score': 0.8533403597564397, 'support': 29261.0}  |
| No log        | 5.0   | 205  | 0.3818          | {'precision': 0.5588235294117647, 'recall': 0.46830985915492956, 'f1-score': 0.5095785440613027, 'support': 284.0}  | {'precision': 0.8787878787878788, 'recall': 0.20567375886524822, 'f1-score': 0.3333333333333333, 'support': 141.0} | {'precision': 0.7287735849056604, 'recall': 0.8728813559322034, 'f1-score': 0.794344473007712, 'support': 708.0}    | {'precision': 0.6021926389976507, 'recall': 0.5673880964092474, 'f1-score': 0.5842725085475498, 'support': 4066.0}    | {'precision': 0.7885196374622356, 'recall': 0.7767857142857143, 'f1-score': 0.782608695652174, 'support': 2016.0}  | {'precision': 0.8760707709550877, 'recall': 0.8973349733497334, 'f1-score': 0.8865753868589484, 'support': 12195.0} | {'precision': 0.9648159446817165, 'recall': 0.9631509491422191, 'f1-score': 0.9639827279654559, 'support': 9851.0} | 0.8573   | {'precision': 0.7711405693145706, 'recall': 0.6787892438770422, 'f1-score': 0.693527952775211, 'support': 29261.0}  | {'precision': 0.8552285449410628, 'recall': 0.8572502648576603, 'f1-score': 0.8549088561404111, 'support': 29261.0} |


### Framework versions

- Transformers 4.45.2
- Pytorch 2.5.0+cu124
- Datasets 2.19.1
- Tokenizers 0.20.1