What's Changed

• Add 12 new model architectures for CPU and Metal inference (https://github.com/nomic-ai/gpt4all/pull/1914) These are Baichuan, BLOOM, CodeShell, GPT-2, Orion, Persimmon, Phi and Phi-2, Plamo, Qwen, Qwen2, Refact, and StableLM. We don't have official downloads for these yet, but TheBloke offers plenty of compatible GGUF quantizations.
• Restore minimum window size of 720x480 (1b524c46179ce94cfdba676cebd373d25189e538)
• Use ChatML for Mistral OpenOrca to make its output formatting more reliable (https://github.com/nomic-ai/gpt4all/pull/1935)

Bug Fixes

• Fix VRAM not being freed when CPU fallback occurs - this makes switching models more reliable (https://github.com/nomic-ai/gpt4all/pull/1901)
• Disable offloading of Mixtral to GPU because we crash otherwise (https://github.com/nomic-ai/gpt4all/pull/1931)
• Limit extensions scanned by LocalDocs to txt, pdf, md, rst - other formats were inserting useless binary data (https://github.com/nomic-ai/gpt4all/pull/1912)
• Fix missing scrollbar for chat history (490404dbb285b89d28964e5389b730ffedfcd7fd)
• Accessibility improvements (4258bb1f8a767fca706008e2296cb9d59661bcb2)

New Contributors

• @boshk0 made their first contribution in https://github.com/nomic-ai/gpt4all/pull/1924

Full Changelog: https://github.com/nomic-ai/gpt4all/compare/v2.6.2...v2.7.0