Update README.md
Browse files
README.md
CHANGED
|
@@ -19,17 +19,39 @@ It achieves the following results on the evaluation set:
|
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
| 22 |
-
|
| 23 |
|
| 24 |
-
## Intended uses & limitations
|
| 25 |
|
| 26 |
-
More information needed
|
| 27 |
|
| 28 |
## Training and evaluation data
|
| 29 |
|
| 30 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
|
| 32 |
-
## Training procedure
|
| 33 |
|
| 34 |
### Training hyperparameters
|
| 35 |
|
|
|
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
| 22 |
+
We have fine-tuned the distilbert-base-uncased to classify news in 20 main topics based on the labeled dataset [20Newsgroups](http://qwone.com/~jason/20Newsgroups/)
|
| 23 |
|
|
|
|
| 24 |
|
|
|
|
| 25 |
|
| 26 |
## Training and evaluation data
|
| 27 |
|
| 28 |
+
The 20 newsgroups dataset comprises around 18000 newsgroups posts on 20 topics split in two subsets: one for training (or development) and
|
| 29 |
+
the other one for testing (or for performance evaluation).
|
| 30 |
+
The split between the train and test set is based upon a messages posted before and after a specific date.
|
| 31 |
+
|
| 32 |
+
These are the 20 topics we trained the model on:
|
| 33 |
+
|
| 34 |
+
'alt.atheism',
|
| 35 |
+
'comp.graphics',
|
| 36 |
+
'comp.os.ms-windows.misc',
|
| 37 |
+
'comp.sys.ibm.pc.hardware',
|
| 38 |
+
'comp.sys.mac.hardware',
|
| 39 |
+
'comp.windows.x',
|
| 40 |
+
'misc.forsale',
|
| 41 |
+
'rec.autos',
|
| 42 |
+
'rec.motorcycles',
|
| 43 |
+
'rec.sport.baseball',
|
| 44 |
+
'rec.sport.hockey',
|
| 45 |
+
'sci.crypt',
|
| 46 |
+
'sci.electronics',
|
| 47 |
+
'sci.med',
|
| 48 |
+
'sci.space',
|
| 49 |
+
'soc.religion.christian',
|
| 50 |
+
'talk.politics.guns',
|
| 51 |
+
'talk.politics.mideast',
|
| 52 |
+
'talk.politics.misc',
|
| 53 |
+
'talk.religion.misc'
|
| 54 |
|
|
|
|
| 55 |
|
| 56 |
### Training hyperparameters
|
| 57 |
|