Update README.md

README.md

@@ -21,63 +21,42 @@ library_name: transformers
 pipeline_tag: text-generation
 ---
 
-# Arete OSS 3B - Emotional Intelligence AI 🧠❤️
+# Arete OSS 3B 🧠❤️
 
-**Arete OSS** is a 3-billion parameter emotional AI assistant fine-tuned for empathetic conversations in **English and Bengali**. 
+An emotionally intelligent AI that actually listens.
 
-Built by **[Iseer & Co.](https://iseer.co)** - A cognitive research organization focused on synthetic intelligence.
+Built by **[Iseer & Co.](https://iseer.co)** - Making AI that understands humans, not just language.
 
-## 🎯 Model Overview
+## What makes Arete different?
 
-- **Base Model:** Qwen2.5-3B-Instruct
-- **Fine-tuning:** QLoRA (4-bit quantization) using Unsloth
-- **Training Duration:** ~7 hours
-- **Dataset:** ~30,000 examples across 4 domains
-- **Languages:** English, Bengali (বাংলা), with code-switching support
+Most AI assistants optimize for being "smart." Arete optimizes for being **empathetic**.
 
-## ✨ Key Features
+- 🇧🇩 Speaks **English and Bengali** fluently
+- 💬 Trained on real emotional support conversations
+- 🎯 Recognizes when you're struggling and responds with care
+- 🌍 Understands cultural context (especially South Asian family dynamics)
 
-✅ **Emotional Intelligence**: Trained on mental health counseling conversations  
-✅ **Bilingual Support**: Natural conversation in English and Bengali  
-✅ **Crisis Awareness**: Recognizes distress signals and responds supportively  
-✅ **Cultural Sensitivity**: Understands Bengali cultural context  
-✅ **Empathetic Responses**: Non-judgmental, validating communication style  
+**3 billion parameters. Infinite empathy.**
 
-## 🚀 Quick Start
+## Try it
 
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
-import torch
 
-# Load model and tokenizer
-model = AutoModelForCausalLM.from_pretrained(
-    "inanxr/Arete-OSS-3B",
-    torch_dtype=torch.float16,
-    device_map="auto"
-)
+model = AutoModelForCausalLM.from_pretrained("inanxr/Arete-OSS-3B")
 tokenizer = AutoTokenizer.from_pretrained("inanxr/Arete-OSS-3B")
 
-# Format prompt (Alpaca style)
 prompt = """Below is an instruction that describes a task. Write a response that appropriately completes the request.
 
 ### Instruction:
 Respond with empathy and understanding to the following message.
 
 ### Input:
-I'm feeling really stressed about work and don't know what to do.
+I'm feeling overwhelmed and don't know what to do.
 
 ### Response:
 """
 
-# Generate response
-inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
-outputs = model.generate(
-    **inputs,
-    max_new_tokens=256,
-    temperature=0.7,
-    top_p=0.9,
-    do_sample=True
-)
-
-response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-print(response.split("### Response:")[-1].strip())
+inputs = tokenizer(prompt, return_tensors="pt")
+outputs = model.generate(**inputs, max_new_tokens=200, temperature=0.7)
+print(tokenizer.decode(outputs[0], skip_special_tokens=True))