# Chat Model

CrystalChat is a chat model fine-tuned based on LLM360/CrystalCoder. It is the latest and highest-performing chat model released under the LLM360 project. CrystalChat is trained using a curated dataset of public language and code. The model's training data, training code, and benchmark data are all publicly available.

LLaMa2lang is a convenient script for fine-tuning LLaMa2-7b into a chat model for a specific language. It uses the Huggingface Open Assistant dataset as the base data and uses the OPUS translation model to fully translate it into the target language. Then, the translated dataset is converted into the input format for the chat model using LLaMa2's prompt format. Finally, LLaMa2-chat is fine-tuned using QLoRA and PEFT. By using LLaMa2lang, you can create a LLaMa2 model suitable for chatting in any non-English language.

SummyMonkey is an intelligent information processing tool. It provides key insights through speech-to-text, email summarization, and chat modes, significantly improving work efficiency. This product utilizes speech-to-text technology to automatically convert audio from meetings, lectures, etc., into text and key points, eliminating the tedious manual note-taking work. It can also summarize a large number of emails, generating key points to avoid being overwhelmed by email information. Moreover, by chatting with emails, you can gain a deeper understanding of the key points and obtain more targeted information. This product is suitable for busy workers who handle a large amount of information.

HuggingChat is an open-source AI chat model designed to provide the best AI chat experience for everyone. Based on the latest GPT model, it has been carefully fine-tuned and can be used in a variety of scenarios, such as writing, programming assistance, and task assistance. HuggingChat has excellent dialogue generation capabilities and can intelligently answer questions, provide help and suggestions.