Model training and deployment

MNN-LLM is an efficient inference framework designed to optimize and accelerate the deployment of large language models on mobile devices and local PCs. It solves the problem of high memory consumption and computational cost through model quantization, hybrid storage and hardware-specific optimization. MNN-LLM performs well in CPU benchmarks and is significantly faster, making it suitable for users who require privacy protection and efficient inference.

bRAG AI is an innovative AI platform that allows users to create and train their own AI models to provide accurate, real-time answers. Its main advantage is that it can conduct personalized training based on data provided by users to ensure the accuracy and reliability of answers. This product is suitable for enterprises and individuals who need customized AI solutions. The price has not yet been determined, but it is expected to provide flexible pricing strategies to meet the needs of different users.

Confident AI is an open source evaluation infrastructure that provides confidence in LLM (Language Model). Users can evaluate their LLM applications by writing and executing test cases and measure their performance using a rich set of open source metrics. By defining expected output and comparing it to actual output, users can determine whether the LLM is performing as expected and identify areas for improvement. Confident AI also provides advanced difference tracking capabilities to help users optimize LLM configurations. In addition, users can take advantage of comprehensive analytics to identify focus use cases and move LLM into production with confidence. Confident AI also provides powerful capabilities to help users confidently take LLM into production, including A/B testing, evaluation, output classification, reporting dashboards, dataset generation and detailed monitoring.

Starling-7B is an open large language model (LLM) trained by reinforcement learning from AI feedback (RLAIF). It comes to full use with our new GPT-4 labeled ranking dataset Nectar and new reward training and policy tuning pipelines. Starling-7B scored 8.09 on MT Bench using GPT-4 as a judge, surpassing all current models on MT-Bench except OpenAI’s GPT-4 and GPT-4 Turbo. We released the ranking dataset Nectar, reward model Starling-RM-7B-alpha, and language model Starling-LM-7B-alpha on HuggingFace, as well as online demos in LMSYS Chatbot Arena. Please look forward to our upcoming code and paper releases, which will provide more details on the entire process.

LLMonitor is a platform that provides observation, analysis and testing for LLM (Language Model) applications. It can record LLM call logs, indicators and tracking, support dialogue evaluation and chat record playback, and help optimize the performance and cost control of AI applications. LLMonitor provides log monitoring, performance analysis, error tracking, user conversation recording, user feedback collection and other functions. It is suitable for various AI development scenarios, including agents, chatbots, etc.

Make Custom GPT is a customized GPT model building platform. Users can customize and develop GPT models according to their own needs. The platform offers a wealth of features and tools that enable users to easily create their own chatbots, language generation models, and more. Product pricing is based on user demand.

Train and tune an OpenAI ChatGPT model in three easy steps using your custom data. Let AI trigger internal workflows based on conversations to make things more efficient.

Stanford CRFM is a model based on the LLaMA 7B model fine-tuned on the 52K instruction demonstration. Through preliminary evaluation with OpenAI's text-davinci-003, it was found that Alpaca's performance is similar to text-davinci-003 in terms of single-round instruction following, but the model is small and easy to reproduce. Specific pricing and release date are unknown. Please visit the official website for more information.

Botdocs are a series of high-quality data sets used to train artificial intelligence to handle common customer service interactions. It can be used to train large language models, intent classifiers, and natural language understanding engines to help enterprises automate common customer service interactions and provide an understanding of customer intent and deliver superior customer experiences. Botdocs are provided in CSV, JSONL and Dialogflow (ES) formats to meet the different needs of AI developers and systems for large language models, intent classifiers and natural language understanding engines.

ChatNode is a platform that allows you to train your own chatbot. By providing various data sources (such as web pages, documents, PDFs, etc.), you can train your chatbot to answer user questions and customize it to your needs. ChatNode also offers the ability to embed into websites, share links, or connect to tools like Slack so that you can use your chatbot in a variety of scenarios.