Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -77,7 +77,9 @@ source_datasets: graelo/wikipedia
|
|
| 77 |
|
| 78 |
This is really more of a "high quality diverse sample" rather than _"we are trying to remove literal duplicate documents"_. Source dataset: [graelo/wikipedia](https://huggingface.co/datasets/graelo/wikipedia).
|
| 79 |
|
| 80 |
-
##
|
|
|
|
|
|
|
| 81 |
|
| 82 |
command:
|
| 83 |
|
|
@@ -137,4 +139,23 @@ config_name = "text-only"
|
|
| 137 |
dataset = load_dataset("BEE-spoke-data/wikipedia-deduped", config_name)
|
| 138 |
```
|
| 139 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 140 |
---
|
|
|
|
| 77 |
|
| 78 |
This is really more of a "high quality diverse sample" rather than _"we are trying to remove literal duplicate documents"_. Source dataset: [graelo/wikipedia](https://huggingface.co/datasets/graelo/wikipedia).
|
| 79 |
|
| 80 |
+
## configs
|
| 81 |
+
|
| 82 |
+
### `default`
|
| 83 |
|
| 84 |
command:
|
| 85 |
|
|
|
|
| 139 |
dataset = load_dataset("BEE-spoke-data/wikipedia-deduped", config_name)
|
| 140 |
```
|
| 141 |
|
| 142 |
+
## token counts
|
| 143 |
+
|
| 144 |
+
### train
|
| 145 |
+
|
| 146 |
+
Using `tiktoken` GPT-4 tokenizer, `train` split, and `text` column:
|
| 147 |
+
|
| 148 |
+
| | num_tokens |
|
| 149 |
+
|:------|----------------:|
|
| 150 |
+
| count | 5.67337e+06 |
|
| 151 |
+
| mean | 612.413 |
|
| 152 |
+
| std | 739.331 |
|
| 153 |
+
| min | 3 |
|
| 154 |
+
| 25% | 163 |
|
| 155 |
+
| 50% | 359 |
|
| 156 |
+
| 75% | 761 |
|
| 157 |
+
| max | 34298 |
|
| 158 |
+
|
| 159 |
+
total: 3,474,446,396
|
| 160 |
+
|
| 161 |
---
|