Update README.md
Browse files
README.md
CHANGED
@@ -16,24 +16,25 @@ should probably proofread and complete it, then remove this comment. -->
|
|
16 |
|
17 |
# ai-chat-censor6
|
18 |
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
- Accuracy: 0.9903
|
23 |
|
24 |
-
|
25 |
|
26 |
-
|
27 |
|
28 |
-
## Intended uses & limitations
|
29 |
|
30 |
-
More information needed
|
31 |
|
32 |
## Training and evaluation data
|
33 |
|
34 |
-
|
|
|
|
|
|
|
|
|
|
|
35 |
|
36 |
-
## Training procedure
|
37 |
|
38 |
### Training hyperparameters
|
39 |
|
|
|
16 |
|
17 |
# ai-chat-censor6
|
18 |
|
19 |
+
The primary focus of the model is detecting sexual/minors category of messages.
|
20 |
+
|
21 |
+
Possible flags are: regular, racist, underage, sexual
|
|
|
22 |
|
23 |
+
# BEWARE
|
24 |
|
25 |
+
The model might categorize any talk about race as racism, for example: "Black people suffer so much in America" will be flagged as "racist".
|
26 |
|
|
|
27 |
|
|
|
28 |
|
29 |
## Training and evaluation data
|
30 |
|
31 |
+
Model was trained on a fully synthetic dataset generated by uncensored 72b models based on qwen2.
|
32 |
+
|
33 |
+
This model is a fine-tuned version of [distilbert/distilbert-base-uncased](https://huggingface.co/distilbert/distilbert-base-uncased) on an unknown dataset.
|
34 |
+
It achieves the following results on the evaluation set:
|
35 |
+
- Loss: 0.0637
|
36 |
+
- Accuracy: 0.9903
|
37 |
|
|
|
38 |
|
39 |
### Training hyperparameters
|
40 |
|