File size: 4,958 Bytes
93cf0dd
 
 
 
8d56678
93cf0dd
 
 
 
430f725
 
93cf0dd
deab01e
3190039
 
 
 
93cf0dd
 
deab01e
93cf0dd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
base_model: google/flan-t5-large
datasets:
- samsum
library_name: peft
license: apache-2.0
metrics:
- rouge
tags:
- summarization
- peft
- generated_from_trainer
widget:
- text: Enter some text to summarize
pipeline_tag: summarization
language:
- en
---


<!-- 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. -->

[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/daigt_team/Summarization%20by%20Finetuning%20FlanT5-LoRA/runs/bzfwtjcj)
# FlanT5Summarization-samsum

This model is a fine-tuned version of [google/flan-t5-large](https://huggingface.co./google/flan-t5-large) on the samsum dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3001
- Rouge1: 0.2788
- Rouge2: 0.1310
- Rougel: 0.2363
- Rougelsum: 0.2369

## 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: 3e-05
- train_batch_size: 128
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 512
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 3

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum |
|:-------------:|:------:|:----:|:---------------:|:------:|:------:|:------:|:---------:|
| 1.1072        | 0.0866 | 5    | 0.9165          | 0.2705 | 0.1135 | 0.2226 | 0.2229    |
| 1.1039        | 0.1732 | 10   | 0.9080          | 0.2709 | 0.1138 | 0.2230 | 0.2234    |
| 1.0848        | 0.2597 | 15   | 0.8917          | 0.2706 | 0.1137 | 0.2228 | 0.2231    |
| 1.0706        | 0.3463 | 20   | 0.8654          | 0.2709 | 0.1142 | 0.2232 | 0.2234    |
| 1.0461        | 0.4329 | 25   | 0.8336          | 0.2706 | 0.1140 | 0.2228 | 0.2232    |
| 1.0187        | 0.5195 | 30   | 0.7960          | 0.2718 | 0.1145 | 0.2240 | 0.2243    |
| 0.9774        | 0.6061 | 35   | 0.7532          | 0.2723 | 0.1152 | 0.2250 | 0.2253    |
| 0.9326        | 0.6926 | 40   | 0.7064          | 0.2726 | 0.1153 | 0.2253 | 0.2257    |
| 0.8834        | 0.7792 | 45   | 0.6570          | 0.2728 | 0.1160 | 0.2259 | 0.2261    |
| 0.833         | 0.8658 | 50   | 0.6080          | 0.2734 | 0.1161 | 0.2262 | 0.2263    |
| 0.7871        | 0.9524 | 55   | 0.5614          | 0.2726 | 0.1156 | 0.2260 | 0.2260    |
| 0.735         | 1.0390 | 60   | 0.5180          | 0.2731 | 0.1169 | 0.2262 | 0.2264    |
| 0.6978        | 1.1255 | 65   | 0.4802          | 0.2736 | 0.1179 | 0.2275 | 0.2276    |
| 0.6464        | 1.2121 | 70   | 0.4482          | 0.2741 | 0.1188 | 0.2283 | 0.2286    |
| 0.6175        | 1.2987 | 75   | 0.4222          | 0.2742 | 0.1193 | 0.2291 | 0.2292    |
| 0.5722        | 1.3853 | 80   | 0.4007          | 0.2740 | 0.1187 | 0.2287 | 0.2287    |
| 0.5443        | 1.4719 | 85   | 0.3834          | 0.2730 | 0.1180 | 0.2282 | 0.2282    |
| 0.5203        | 1.5584 | 90   | 0.3692          | 0.2740 | 0.1192 | 0.2293 | 0.2293    |
| 0.4851        | 1.6450 | 95   | 0.3568          | 0.2744 | 0.1201 | 0.2300 | 0.2302    |
| 0.4619        | 1.7316 | 100  | 0.3466          | 0.2746 | 0.1201 | 0.2304 | 0.2305    |
| 0.4484        | 1.8182 | 105  | 0.3379          | 0.2754 | 0.1218 | 0.2314 | 0.2319    |
| 0.4357        | 1.9048 | 110  | 0.3305          | 0.2766 | 0.1241 | 0.2325 | 0.2330    |
| 0.4246        | 1.9913 | 115  | 0.3243          | 0.2772 | 0.1254 | 0.2338 | 0.2341    |
| 0.4074        | 2.0779 | 120  | 0.3190          | 0.2776 | 0.1263 | 0.2343 | 0.2347    |
| 0.3965        | 2.1645 | 125  | 0.3144          | 0.2775 | 0.1264 | 0.2342 | 0.2345    |
| 0.3922        | 2.2511 | 130  | 0.3105          | 0.2776 | 0.1266 | 0.2344 | 0.2347    |
| 0.3861        | 2.3377 | 135  | 0.3073          | 0.2786 | 0.1289 | 0.2357 | 0.2362    |
| 0.382         | 2.4242 | 140  | 0.3048          | 0.2782 | 0.1289 | 0.2354 | 0.2358    |
| 0.3807        | 2.5108 | 145  | 0.3029          | 0.2787 | 0.1297 | 0.2359 | 0.2364    |
| 0.3717        | 2.5974 | 150  | 0.3016          | 0.2787 | 0.1303 | 0.2363 | 0.2367    |
| 0.3708        | 2.6840 | 155  | 0.3008          | 0.2788 | 0.1305 | 0.2363 | 0.2368    |
| 0.372         | 2.7706 | 160  | 0.3003          | 0.2789 | 0.1310 | 0.2365 | 0.2370    |
| 0.3696        | 2.8571 | 165  | 0.3002          | 0.2788 | 0.1310 | 0.2363 | 0.2369    |
| 0.3646        | 2.9437 | 170  | 0.3001          | 0.2788 | 0.1310 | 0.2363 | 0.2369    |


### Framework versions

- PEFT 0.12.0
- Transformers 4.43.2
- Pytorch 2.1.2
- Datasets 2.20.0
- Tokenizers 0.19.1