File size: 2,141 Bytes
9ceb86a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- cnn_dailymail
metrics:
- rouge
model-index:
- name: flan-t5-xl
  results:
  - task:
      name: Sequence-to-sequence Language Modeling
      type: text2text-generation
    dataset:
      name: cnn_dailymail
      type: cnn_dailymail
      config: 3.0.0
      split: validation
      args: 3.0.0
    metrics:
    - name: Rouge1
      type: rouge
      value: 45.1318
---

<!-- 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. -->

# flan-t5-xl

This model is a fine-tuned version of [google/flan-t5-xl](https://huggingface.co/google/flan-t5-xl) on the cnn_dailymail dataset.
It achieves the following results on the evaluation set:
- Loss: 1.2648
- Rouge1: 45.1318
- Rouge2: 22.2773
- Rougel: 31.9084
- Rougelsum: 42.0558
- Gen Len: 94.2332

## 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: 0.0001
- train_batch_size: 24
- eval_batch_size: 24
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- total_train_batch_size: 96
- total_eval_batch_size: 96
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rouge1  | Rouge2  | Rougel  | Rougelsum | Gen Len  |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:|
| 1.4352        | 1.0   | 2991 | 1.2645          | 43.8582 | 21.2227 | 30.7038 | 40.761    | 101.9968 |
| 1.3198        | 2.0   | 5982 | 1.2525          | 44.4594 | 21.8174 | 31.4304 | 41.4563   | 94.0733  |
| 1.2151        | 3.0   | 8973 | 1.2648          | 45.1318 | 22.2773 | 31.9084 | 42.0558   | 94.2332  |


### Framework versions

- Transformers 4.27.1
- Pytorch 2.0.0
- Datasets 2.10.1
- Tokenizers 0.13.2