File size: 1,833 Bytes
f13902e be421bf f8159f8 f13902e 9ac2944 b03a128 9ac2944 7d55635 119c75f 7d55635 be421bf |
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 64 65 66 |
---
license: mit
datasets:
- ifmain/text-moderation
pipeline_tag: text-classification
---
# moderation by embeddings
This is a simple multilingual model for text moderation using embeddings.
PS: Although this model itself is MIT, it uses sentence `sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2` under `license: apache-2.0`.
exaple usage:
```python
from moderation import * #From files this project
# Load model
moderation = ModerationModel()
moderation.load_state_dict(torch.load('moderation_model.pth'))
# Test text
text = "I want to kill them."
embeddings_for_prediction = getEmb(text)
prediction = predict(moderation, embeddings_for_prediction)
print(json.dumps(prediction,indent=4))
```
Output:
```json
{
"category_scores": {
"harassment": 0.039179909974336624,
"harassment-threatening": 0.5689294338226318,
"hate": 0.0096114631742239,
"hate-threatening": 0.00895680021494627,
"self-harm": 0.0008832099265418947,
"self-harm-instructions": 2.1136918803676963e-05,
"self-harm-intent": 0.00033596932189539075,
"sexual": 5.425313793239184e-05,
"sexual-minors": 5.160131422599079e-06,
"violence": 0.9684166312217712,
"violence-graphic": 0.0015151903498917818
},
"detect": {
"harassment": false,
"harassment-threatening": true,
"hate": false,
"hate-threatening": false,
"self-harm": false,
"self-harm-instructions": false,
"self-harm-intent": false,
"sexual": false,
"sexual-minors": false,
"violence": true,
"violence-graphic": false
},
"detected": true
}
```
This model covert embedings to moderaton score
The dataset helped with normalizing the model output, but the model does not include rows from the dataset |