jandres/GenAIExamples
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Fix typos and add definitions for toxicity detection microservice (#553)

Browse Source 
* fix typos

Signed-off-by: Tyler Wilbers <tyler.wilbers@intel.com>

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Signed-off-by: Tyler Wilbers <tyler.wilbers@intel.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
This commit is contained in:
Tyler W
2024-08-22 17:33:25 -07:00
committed by GitHub
parent 068527df31
commit 9b8798a3eb
2 changed files with 14 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
comps/guardrails/README.md
View File

@@ -2,9 +2,10 @@
The Guardrails service enhances the security of LLM-based applications by offering a suite of microservices designed to ensure trustworthiness, safety, and security.
| MicroService | Description |
| ------------------------------------------ | ------------------------------------------------------------------------------------------ |
| [Llama Guard](./llama_guard/README.md) | Provides guardrails for inputs and outputs to ensure safe interactions |
| [PII Detection](./pii_detection/README.md) | Detects Personally Identifiable Information (PII) and Business Sensitive Information (BSI) |
| MicroService | Description |
| ---------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------ |
| [Llama Guard](./llama_guard/README.md) | Provides guardrails for inputs and outputs to ensure safe interactions |
| [PII Detection](./pii_detection/README.md) | Detects Personally Identifiable Information (PII) and Business Sensitive Information (BSI) |
| [Toxicity Detection](./toxicity_detection/README.md) | Detects Toxic language (rude, disrespectful, or unreasonable language that is likely to make someone leave a discussion) |
Additional safety-related microservices will be available soon.

18
comps/guardrails/toxicity_detection/README.md
View File

@@ -1,14 +1,14 @@
# Toxicity Detection Microservice
# ☣️💥🛡️<span style="color:royalblue"> Intel Toxicity Detection Model </span>
# ☣️💥🛡️Toxicity Detection Microservice
## Introduction
Intel also provides toxicity detection model, which is lightweight, runs efficiently on a CPU, and performs well on toxic_chat and jigsaws datasets. More datasets are being fine-tuned. If you're interested, please contact abolfazl.shahbazi@intel.com.
Toxicity Detection Microservice allows AI Application developers to safeguard user input and LLM output from harmful language in a RAG environment. By leveraging a smaller fine-tuned Transformer model for toxicity classification (e.g. DistilledBERT, RoBERTa, etc.), we maintain a lightweight guardrails microservice without significantly sacrificing performance making it readily deployable on both Intel Gaudi and Xeon.
## Training Customerizable Toxicity Model on Gaudi2
Toxicity is defined as rude, disrespectful, or unreasonable language likely to make someone leave a conversation. This can include instances of aggression, bullying, targeted hate speech, or offensive language. For more information on labels see [Jigsaw Toxic Comment Classification Challenge](http://kaggle.com/c/jigsaw-toxic-comment-classification-challenge).
Additionally, we offer a fine-tuning workflow on Intel Gaudi2, allowing you to customerize your toxicity detecction model to suit your unique needs.
## Future Development
- Add a RoBERTa (125M params) toxicity model fine-tuned on Gaudi2 with ToxicChat and Jigsaw dataset in an optimized serving framework.
# 🚀1. Start Microservice with Python（Option 1）
@@ -24,7 +24,7 @@ pip install -r requirements.txt
python toxicity_detection.py
```
# 🚀2. Start Microservie with Docker (Option 2)
# 🚀2. Start Microservice with Docker (Option 2)
## 2.1 Prepare toxicity detection model
@@ -58,7 +58,7 @@ Once microservice starts, users can use examples (bash or python) below to apply
```bash
curl localhost:9091/v1/toxicity
-X POST
-d '{"text":"How to poison your neighbor'\''s dog secretly"}'
-d '{"text":"How to poison my neighbor'\''s dog without being caught?"}'
-H 'Content-Type: application/json'
```
@@ -76,7 +76,7 @@ import json
proxies = {"http": ""}
url = "http://localhost:9091/v1/toxicity"
data = {"text": "How to poison your neighbor'''s dog without being caught?"}
data = {"text": "How to poison my neighbor'''s dog without being caught?"}
try:
resp = requests.post(url=url, data=data, proxies=proxies)