Datasets:
Tasks:
Summarization
Formats:
csv
Languages:
English
Size:
10K - 100K
Tags:
stacked summaries
License:
File size: 2,008 Bytes
d25808a d3f5b3c 77a79c2 d25808a b210341 b186aa6 04026e0 b186aa6 b210341 3ec4558 b210341 3ec4558 b210341 77a79c2 ad31112 b210341 ad31112 4ecee94 b210341 4ecee94 b210341 |
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 |
---
license: apache-2.0
source_datasets:
- samsum
task_categories:
- summarization
language:
- en
tags:
- stacked summaries
pretty_name: Stacked Samsum - 1024
size_categories:
- 10K<n<100K
---
# stacked samsum 1024
Created with the `stacked-booksum` repo version v0.25. It contains:
1. Original Dataset: copy of the base dataset
2. Stacked Rows: The original dataset is processed by stacking rows based on certain criteria:
- Maximum Input Length: The maximum length for input sequences is 1024 tokens in the longt5 model tokenizer.
- Maximum Output Length: The maximum length for output sequences is also 1024 tokens in the longt5 model tokenizer.
3. Special Token: The dataset utilizes the `[NEXT_CONCEPT]` token to indicate a new topic **within** the same summary. It is recommended to explicitly add this special token to your model's tokenizer before training, ensuring that it is recognized and processed correctly during downstream usage.
## stats
## dataset details
Default (train):
```python
[2022-12-04 13:19:32] INFO:root:{'num_columns': 4,
'num_rows': 14732,
'num_unique_target': 14730,
'num_unique_text': 14265,
'summary - average chars': 110.13,
'summary - average tokens': 28.693727939180015,
'text input - average chars': 511.22,
'text input - average tokens': 148.88759163725223}
```
stacked (train)
```python
[2022-12-05 00:49:04] INFO:root:stacked 14730 rows, 2 rows were ineligible
[2022-12-05 00:49:04] INFO:root:dropped 20 duplicate rows, 29442 rows remain
[2022-12-05 00:49:04] INFO:root:shuffling output with seed 182
[2022-12-05 00:49:04] INFO:root:STACKED - basic stats - train
[2022-12-05 00:49:04] INFO:root:{'num_columns': 5,
'num_rows': 29442,
'num_unique_chapters': 28975,
'num_unique_summaries': 29441,
'summary - average chars': 452.8,
'summary - average tokens': 106.46820868147545,
'text input - average chars': 1814.09,
'text input - average tokens': 528.665579783982}
``` |