File size: 3,526 Bytes
6647bd5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
---
license: apache-2.0
base_model: t5-3b
tags:
- generated_from_trainer
datasets:
- glue
metrics:
- accuracy
model-index:
- name: t5-3b_rte_sp0_ar0
  results:
  - task:
      name: Text Classification
      type: text-classification
    dataset:
      name: glue
      type: glue
      config: rte
      split: validation
      args: rte
    metrics:
    - name: Accuracy
      type: accuracy
      value: 0.8995983935742972
---

<!-- 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. -->

# t5-3b_rte_sp0_ar0

This model is a fine-tuned version of [t5-3b](https://huggingface.co/t5-3b) on the glue dataset.
It achieves the following results on the evaluation set:
- Loss: 2.4886
- Accuracy: 0.8996

## 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: 5e-05
- train_batch_size: 8
- eval_batch_size: 16
- seed: 1
- distributed_type: multi-GPU
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 20
- training_steps: 750

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.7116        | 0.18  | 25   | 0.7078          | 0.4838   |
| 0.615         | 0.36  | 50   | 0.4571          | 0.8484   |
| 0.501         | 0.53  | 75   | 0.4176          | 0.8484   |
| 0.337         | 0.71  | 100  | 0.4462          | 0.8809   |
| 0.4199        | 0.89  | 125  | 0.4309          | 0.7942   |
| 0.2477        | 1.07  | 150  | 0.4325          | 0.8881   |
| 0.2253        | 1.25  | 175  | 0.4596          | 0.8845   |
| 0.214         | 1.42  | 200  | 0.9106          | 0.8736   |
| 0.2577        | 1.6   | 225  | 0.3652          | 0.9025   |
| 0.1966        | 1.78  | 250  | 0.3645          | 0.9097   |
| 0.2278        | 1.96  | 275  | 0.3337          | 0.9206   |
| 0.1221        | 2.14  | 300  | 1.2373          | 0.8881   |
| 0.3576        | 2.31  | 325  | 2.0514          | 0.8809   |
| 0.6186        | 2.49  | 350  | 3.9886          | 0.8809   |
| 1.0915        | 2.67  | 375  | 3.3403          | 0.8845   |
| 0.321         | 2.85  | 400  | 4.6906          | 0.8989   |
| 0.3969        | 3.02  | 425  | 1.2608          | 0.8736   |
| 0.026         | 3.2   | 450  | 4.4563          | 0.8809   |
| 0.0695        | 3.38  | 475  | 4.6858          | 0.8917   |
| 0.7626        | 3.56  | 500  | 4.7502          | 0.8917   |
| 0.2675        | 3.74  | 525  | 5.0576          | 0.9025   |
| 0.82          | 3.91  | 550  | 3.6297          | 0.9025   |
| 0.0011        | 4.09  | 575  | 5.7629          | 0.8989   |
| 0.3           | 4.27  | 600  | 2.3117          | 0.9097   |
| 0.0544        | 4.45  | 625  | 1.5657          | 0.9097   |
| 0.2907        | 4.63  | 650  | 2.6475          | 0.9025   |
| 0.2868        | 4.8   | 675  | 2.6720          | 0.8989   |
| 0.3191        | 4.98  | 700  | 2.8372          | 0.8773   |
| 0.0382        | 5.16  | 725  | 6.0565          | 0.9025   |
| 0.0297        | 5.34  | 750  | 4.1639          | 0.8736   |


### Framework versions

- Transformers 4.35.2
- Pytorch 2.1.1+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0