update model card README.md
Browse files
README.md
CHANGED
@@ -1,9 +1,85 @@
|
|
1 |
---
|
2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
## Training procedure
|
5 |
|
6 |
-
###
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
|
9 |
-
-
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
license: apache-2.0
|
3 |
+
base_model: google/flan-t5-small
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
datasets:
|
7 |
+
- billsum
|
8 |
+
metrics:
|
9 |
+
- rouge
|
10 |
+
model-index:
|
11 |
+
- name: 3_loa
|
12 |
+
results: []
|
13 |
---
|
14 |
+
|
15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
16 |
+
should probably proofread and complete it, then remove this comment. -->
|
17 |
+
|
18 |
+
# 3_loa
|
19 |
+
|
20 |
+
This model is a fine-tuned version of [google/flan-t5-small](https://huggingface.co/google/flan-t5-small) on the billsum dataset.
|
21 |
+
It achieves the following results on the evaluation set:
|
22 |
+
- Loss: 2.1919
|
23 |
+
- Rouge1: 0.1973
|
24 |
+
- Rouge2: 0.1007
|
25 |
+
- Rougel: 0.1708
|
26 |
+
- Rougelsum: 0.1711
|
27 |
+
- Gen Len: 19.0
|
28 |
+
|
29 |
+
## Model description
|
30 |
+
|
31 |
+
More information needed
|
32 |
+
|
33 |
+
## Intended uses & limitations
|
34 |
+
|
35 |
+
More information needed
|
36 |
+
|
37 |
+
## Training and evaluation data
|
38 |
+
|
39 |
+
More information needed
|
40 |
+
|
41 |
## Training procedure
|
42 |
|
43 |
+
### Training hyperparameters
|
44 |
+
|
45 |
+
The following hyperparameters were used during training:
|
46 |
+
- learning_rate: 2e-05
|
47 |
+
- train_batch_size: 8
|
48 |
+
- eval_batch_size: 8
|
49 |
+
- seed: 42
|
50 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
51 |
+
- lr_scheduler_type: linear
|
52 |
+
- num_epochs: 20
|
53 |
|
54 |
+
### Training results
|
55 |
+
|
56 |
+
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|
57 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:|
|
58 |
+
| No log | 1.0 | 124 | 2.2622 | 0.1909 | 0.0921 | 0.1656 | 0.1659 | 19.0 |
|
59 |
+
| No log | 2.0 | 248 | 2.2534 | 0.1931 | 0.0956 | 0.1679 | 0.1681 | 19.0 |
|
60 |
+
| No log | 3.0 | 372 | 2.2433 | 0.1952 | 0.0967 | 0.1697 | 0.1699 | 19.0 |
|
61 |
+
| No log | 4.0 | 496 | 2.2358 | 0.1953 | 0.0978 | 0.1701 | 0.1702 | 19.0 |
|
62 |
+
| 2.4755 | 5.0 | 620 | 2.2323 | 0.1951 | 0.0981 | 0.1705 | 0.1706 | 19.0 |
|
63 |
+
| 2.4755 | 6.0 | 744 | 2.2253 | 0.1962 | 0.0996 | 0.1712 | 0.1714 | 19.0 |
|
64 |
+
| 2.4755 | 7.0 | 868 | 2.2199 | 0.1968 | 0.1003 | 0.1719 | 0.172 | 19.0 |
|
65 |
+
| 2.4755 | 8.0 | 992 | 2.2170 | 0.1963 | 0.0999 | 0.1717 | 0.1717 | 19.0 |
|
66 |
+
| 2.4416 | 9.0 | 1116 | 2.2134 | 0.1971 | 0.1002 | 0.1723 | 0.1724 | 19.0 |
|
67 |
+
| 2.4416 | 10.0 | 1240 | 2.2069 | 0.1967 | 0.0995 | 0.1715 | 0.1716 | 19.0 |
|
68 |
+
| 2.4416 | 11.0 | 1364 | 2.2053 | 0.1983 | 0.102 | 0.1729 | 0.1732 | 19.0 |
|
69 |
+
| 2.4416 | 12.0 | 1488 | 2.2034 | 0.1976 | 0.1018 | 0.1722 | 0.1725 | 19.0 |
|
70 |
+
| 2.4153 | 13.0 | 1612 | 2.1995 | 0.1985 | 0.1019 | 0.1725 | 0.1727 | 19.0 |
|
71 |
+
| 2.4153 | 14.0 | 1736 | 2.1980 | 0.198 | 0.1016 | 0.1721 | 0.1722 | 19.0 |
|
72 |
+
| 2.4153 | 15.0 | 1860 | 2.1961 | 0.1983 | 0.1017 | 0.172 | 0.1721 | 19.0 |
|
73 |
+
| 2.4153 | 16.0 | 1984 | 2.1947 | 0.1977 | 0.1013 | 0.1715 | 0.1717 | 19.0 |
|
74 |
+
| 2.4069 | 17.0 | 2108 | 2.1936 | 0.1976 | 0.101 | 0.1714 | 0.1716 | 19.0 |
|
75 |
+
| 2.4069 | 18.0 | 2232 | 2.1925 | 0.1977 | 0.1013 | 0.1713 | 0.1715 | 19.0 |
|
76 |
+
| 2.4069 | 19.0 | 2356 | 2.1918 | 0.1973 | 0.1007 | 0.1709 | 0.1711 | 19.0 |
|
77 |
+
| 2.4069 | 20.0 | 2480 | 2.1919 | 0.1973 | 0.1007 | 0.1708 | 0.1711 | 19.0 |
|
78 |
+
|
79 |
+
|
80 |
+
### Framework versions
|
81 |
|
82 |
+
- Transformers 4.31.0
|
83 |
+
- Pytorch 1.13.1.post200
|
84 |
+
- Datasets 2.10.0
|
85 |
+
- Tokenizers 0.13.2
|