File size: 1,980 Bytes
2050ee9
 
15612e6
2050ee9
 
7ea645a
 
2050ee9
 
 
 
 
 
 
 
 
 
15612e6
7ea645a
ce15d4b
 
 
 
 
2050ee9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ce15d4b
2050ee9
 
 
 
7ea645a
 
ce15d4b
 
 
 
2050ee9
 
 
 
 
 
 
 
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
---
license: mit
base_model: li-jay-cs/gptj-supervised-summarize-checkpoint
tags:
- generated_from_trainer
metrics:
- rouge
model-index:
- name: gptj-supervised-summarize-checkpoint
  results: []
---

<!-- 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. -->

# gptj-supervised-summarize-checkpoint

This model is a fine-tuned version of [li-jay-cs/gptj-supervised-summarize-checkpoint](https://huggingface.co/li-jay-cs/gptj-supervised-summarize-checkpoint) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.8506
- Rouge1: 0.5938
- Rouge2: 0.1912
- Rougel: 0.3937
- Rougelsum: 0.5184

## 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: 1e-05
- train_batch_size: 50
- eval_batch_size: 50
- seed: 42
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 2
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|
| 1.904         | 0.43  | 1000 | 1.8633          | 0.5912 | 0.1888 | 0.3913 | 0.5149    |
| 1.8931        | 0.86  | 2000 | 1.8584          | 0.5907 | 0.1890 | 0.3920 | 0.5153    |
| 1.8758        | 1.28  | 3000 | 1.8545          | 0.5929 | 0.1906 | 0.3929 | 0.5168    |
| 1.8699        | 1.71  | 4000 | 1.8506          | 0.5938 | 0.1912 | 0.3937 | 0.5184    |


### Framework versions

- Transformers 4.35.0
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.14.1