Today, @Dattarajguttedr and myself are excited to unveil our AI-backed Mental Health Chat Bot built on the robust #MistralModel. This project harnesses cutting-edge open source tech to enhance mental health support. #MentalHealthAI By leveraging @langchain , we parse and extract insights from a wide range of mental health research papers and content. #Research #AI Our journey includes a self-hosted model deployment: A standout feature? We’re introducing a vernacular, voice-based mental health AI that supports all major Indian languages—making support accessible for everyone across India soon. We’ve fine-tuned our model with PDFs loaded with context-rich mental health research. This gives our Chat Bot a deep knowledge base to back every interaction. All PDFs are dynamically downloaded and integrated into the system. Our code uses semantic search to ensure every prompt is matched with the right context and insights. This project marks a pivotal step in merging cutting-edge AI and mental healthcare. I’m proud to bring together open source tech, innovative deployments, and inclusive design. #AIForGood Feel Free to use this - itsokay.life

Expand Your AI Library with Llama, Mistral, and Other Popular LLM Models ... myscrappydays.com/news.php?s… #LLMmodels #Oobaboogaintegration #LlamaAI #Mistralmodel #GPTNeoX #OpensourceAI #AItextgeneration

Anyone had success torch.compile MistralModel from transformers? Only being able to torch.compile vision_tower and enable flash_attn so far (for LlavaMistralModel).

e5-mistral-7b-instructがMistralForCausalLMではなくMistralModelで作られていてマージ前に準備が必要だったので、手順をmodel cardに書いておきました。 huggingface.co/oshizo/japane…