README.md

metadata

title: DeepHat V1 7B Chat
emoji: 💬
colorFrom: yellow
colorTo: purple
sdk: gradio
sdk_version: 5.49.1
app_file: app.py
pinned: false
hf_oauth: true
hf_oauth_scopes:
  - inference-api
license: apache-2.0
models:
  - DeepHat/DeepHat-V1-7B
  - meta-llama/Llama-2-7b-chat-hf
  - mistralai/Mistral-7B-Instruct-v0.3
datasets:
  - lmsys/lmsys-chat-1m
  - HuggingFaceH4/ultrachat_200k
  - RUCAIBox/Open-Dialogue
  - bitext/Bitext-customer-support-llm-chatbot-training-dataset

💬 DeepHat V1 7B Chat

DeepHat V1 7B Chat is an open-source conversational interface built with Gradio 5 and the Hugging Face Inference API.
It serves as a lightweight front-end for the DeepHat V1 7B model — a cybersecurity-aware 7-billion-parameter chat model fine-tuned for threat-analysis, red/blue-team simulation, and secure-code review contexts.

---

🚀 Features

• Streaming chat interface with token-by-token updates
• Hugging Face OAuth integration (secure inference token handling)
• Adjustable generation parameters (temperature, top-p, max tokens)
• Compatible with any model exposing the chat_completion API

---

🧠 Core Model

Model	Description	License
DeepHat/DeepHat-V1-7B	Domain-tuned 7B parameter model for cybersecurity and DevSecOps reasoning	Apache-2.0 + usage restrictions
meta-llama/Llama-2-7b-chat-hf	General-purpose 7B chat baseline model	LLaMA 2 Community License
mistralai/Mistral-7B-Instruct-v0.3	Efficient instruction-tuned 7B chat model for general dialogue	Apache-2.0

---

📚 Datasets for Fine-Tuning and Evaluation

Dataset	Purpose	Link
lmsys/lmsys-chat-1m	Large multi-turn conversation dataset for chat alignment	HF Datasets
HuggingFaceH4/ultrachat_200k	Compact chat corpus used for instruction-tuned 7B models	HF Datasets
RUCAIBox/Open-Dialogue	Generic dialogue meta-collection (PersonaChat, DailyDialog, etc.)	HF Datasets
bitext/Bitext-customer-support-llm-chatbot-training-dataset	Example domain dataset — can be adapted for incident response dialogs	HF Datasets

---

🧩 Project Metadata