Commit
•
abcb186
1
Parent(s):
9d506e7
Update overview.py
Browse files- overview.py +0 -1
overview.py
CHANGED
|
@@ -319,7 +319,6 @@ overview_div = Div(
|
|
| 319 |
Li("removing duplicataive data can lead to better accuracy", style = "margin-bottom: 5px"),
|
| 320 |
Li("prevent train-test overlap", style = "margin-bottom: 5px"),
|
| 321 |
Li("minimizes the risk of memorization leading to test loss", style = "margin-bottom: 5px"),
|
| 322 |
-
Li("", style = "margin-bottom: 5px"),
|
| 323 |
),
|
| 324 |
P("Furthermore, by controlling the pretraining data distribution through deduplication and selective upsampling avoida relying on the often inconsistent distribution of internet-sourced data."),
|
| 325 |
H3("Highlights of the TxT360 Deduplication Process"),
|
|
|
|
| 319 |
Li("removing duplicataive data can lead to better accuracy", style = "margin-bottom: 5px"),
|
| 320 |
Li("prevent train-test overlap", style = "margin-bottom: 5px"),
|
| 321 |
Li("minimizes the risk of memorization leading to test loss", style = "margin-bottom: 5px"),
|
|
|
|
| 322 |
),
|
| 323 |
P("Furthermore, by controlling the pretraining data distribution through deduplication and selective upsampling avoida relying on the often inconsistent distribution of internet-sourced data."),
|
| 324 |
H3("Highlights of the TxT360 Deduplication Process"),
|