Changes (#2726)

• Release/0.4.12 to main @Van-QA (#2808)
• Feat: Bump version @hahuyhoang411 (#2812)
• Feat: Change featured models @hahuyhoang411 (#2805)
• Release cut 0.4.12 @Van-QA (#2787)
• Feat: Add new Llama 3 and models to Groq Extension @metaspartan (#2786)
• Chore: temporary disable cache for server migration @hiento09 (#2742)
• Bump nitro to 0.3.22 @Van-QA (#2740)
• Refactor/deprecate hugging face ext @namchuai (#2620)
• Sync main to dev (0.4.11) @louis-jan (#2725)
• Fix typos @deining (#2714)

🚀 Features

• feat: add LaTeX extension support via Marked @louis-jan (#2766)
• Feature windows nsis @hiento09 (#2705)

🐛 Fixes

• Fix macos ci @hiento09 (#2814)
• fix: infinite talking in auto thread titles generators @louis-jan (#2810)
• fix: stop server if failed to load model @namchuai (#2807)
• fix: new thread with overridden settings @louis-jan (#2798)
• fix: override cpu_threads setting from model.json @louis-jan (#2789)
• fix: wrong condition for GPU acceleration @louis-jan (#2770)
• fix: thread title not persisted @namchuai (#2765)
• fix: broken stop word input - support multiple stop words @louis-jan (#2762)
• fix: cannot download phin34 model @namchuai (#2745)
• fix: user cannot click on copy icon of model id @namchuai (#2744)
• fix: the log server does not write to the log file @louis-jan (#2738)
• fix: log is not generated after clear log from setting @namchuai (#2732)
• fix: broken several ui when model name too long @urmauur (#2728)
• fix: cannot cancel model which is progress at 0 percent @namchuai (#2727)

🧰 Maintenance

• chore: detailed message when fetch invalid url @namchuai (#2780)
• chore: add model.json for Llama3 and other outdated model version @Van-QA (#2773)
• chore: set default context length to 2048 @namchuai (#2776)
• refactor: good bye react-scroll-to-bottom @namchuai (#2769)
• chore: correct copy for Jan data folder to highlight original Jan data folder intact upon relocation @Van-QA (#2751)
• chore: default context length to 2048 @namchuai (#2746)

Contributor

@Van-QA, @deining, @hahuyhoang411, @hiento09, @louis-jan, @metaspartan, @namchuai and @urmauur