Upload README.md

README.md

@@ -1,36 +1,33 @@
 ---
 license: apache-2.0
 inference: false 
+tags: [green, p1, llmware-encoder, ov]
 ---
 
-# bling-tiny-llama-ov
+# unitary-toxic-roberta-ov  
 
-<!-- Provide a quick summary of what the model is/does. -->
-
-**bling-tiny-llama-ov** is an OpenVino int4 quantized version of BLING Tiny-Llama 1B, providing a very fast, very small inference implementation, optimized for AI PCs using Intel GPU, CPU and NPU.    
-
-[**bling-tiny-llama**](https://huggingface.co/llmware/bling-tiny-llama-v0) is a fact-based question-answering model, optimized for complex business documents.  
-
-Get started right away with [OpenVino](https://github.com/openvinotoolkit/openvino) 
-
-Looking for AI PC solutions and demos, contact us at [llmware](https://www.llmware.ai)    
+**unitary-toxic-roberta-ov** is a toxicity classifier from [unitary/unbiased-toxic-roberta](https://www.huggingface.com/unitary/unbiased-toxic-roberta), packaged in OpenVino format. 
 
+The classifier can be used to evaluate toxic content in a prompt or in model output.  
 
 ### Model Description
 
-- **Developed by:** llmware  
-- **Model type:** tinyllama  
-- **Parameters:** 1.1 billion
-- **Model Parent:** llmware/bling-tiny-llama-v0  
+- **Developed by:** unitary
+- **Quantized by:** llmware
+- **Model type:** roberta  
+- **Parameters:** 125 million
+- **Model Parent:** unitary/unbiased-toxic-roberta
 - **Language(s) (NLP):** English  
 - **License:** Apache 2.0  
-- **Uses:** Fact-based question-answering  
-- **RAG Benchmark Accuracy Score:** 86.5  
+- **Uses:** Prompt safety    
+- **RAG Benchmark Accuracy Score:** NA  
 - **Quantization:** int4  
   
 
 ## Model Card Contact
 
+[llmware on github](https://www.github.com/llmware-ai/llmware)  
+
 [llmware on hf](https://www.huggingface.co/llmware)  
 
 [llmware website](https://www.llmware.ai)