amalsp/mistral-7b-ai-tools-finetuned

🛠️ Mistral-7B AI Tools Explainer (LoRA Fine-tuned)

This model is a fine-tuned version of Mistral-7B using LoRA (Low-Rank Adaptation). It is trained specifically to explain the use, purpose, and functionality of AI tools when given their name, task, and description.

---

📌 Model Details

• Developed by: amalsp
• Model type: Causal Language Model (Decoder-only Transformer)
• Base model: mistralai/Mistral-7B-v0.1
• Fine-tuning technique: PEFT with LoRA
• Language(s): English
• License: Apache 2.0
• Trained on: 2000 entries about AI tools
• GPU Used: T4 (Colab Pro)
• Training Time: ~2.7 hours
• Loss: 1.83
• Sequence Length: 256 tokens

---

🚀 Intended Use

✅ Direct Use

• Generate accurate and informative descriptions of AI tools
• Help end-users, students, or professionals understand what a given AI tool does

🚫 Out-of-Scope Use

• Not intended for general Q&A about geography, politics, history, or unrelated domains
• Will reply with: "I am not trained for this." to unrelated prompts

---

💡 Usage

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel
import torch

# Load model
base = AutoModelForCausalLM.from_pretrained("mistralai/Mistral-7B-v0.1", torch_dtype=torch.float16, device_map="auto")
model = PeftModel.from_pretrained(base, "amalsp/mistral-7b-ai-tools-finetuned")
tokenizer = AutoTokenizer.from_pretrained("amalsp/mistral-7b-ai-tools-finetuned")

# Generate
prompt = """Tool: ChatGPT
Task: Conversational AI
Short Description: AI chatbot for natural conversations
Explain the tool:"""

inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_length=300)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))