Text Generation Web UI Files

Changes

• Merge the Default and Notebook tabs into a single Notebook tab (#7078), with an option in the Session tab to switch between one and two columns.
• Autosave text in the Notebook tab (both generated and manually typed), and add "New" and "Rename" buttons for management.
• Add a new Character tab for character settings.
• Remember the last selected chat for each chat mode and character.
• Truncate web search results to at most 8192 tokens to handle edge cases like pages with infinite scrolling.
• Remove images and links from web search results to reduce noise and focus on the relevant text content.
• Add an option to exclude attachments from previous messages in the chat prompt. It can be found in the Session tab.
• Improve the wpp chat style.
• Increase the size of the enlarged character profile picture that appears when clicking the profile picture.
• Move 'Custom system message' to the Parameters > Generation tab.
• Hide the navigation bar on Ctrl+S / Show controls click.
• Always close/open the two sidebars at the same time when clicking their close buttons on desktop.
• Only save active extensions and extensions settings on manual settings save.
• More informative log message when the user input gets truncated.
• Small style improvements to the chat tab.
• Optimize scrolling in the chat tab.
• Optimize syntax highlighting on long conversations.
• Optimize the token count at the end of generation with llama.cpp.
• Disable message action icons during streaming for better performance.
• Expose real model list via /v1/models endpoint (#7088). Thanks @NoxWorld2660
• Improved API examples in the documentation.

Bug fixes

• Ensure estimated vram is updated when switching between different models (#7071). Thanks @miriameng.
• Fix an edge case where the gpu-layers slider maximum is incorrectly limited.
• Add error handling for non-llama.cpp models in portable mode.
• Fix the character profile picture sometimes not appearing when switching from instruct to chat modes.
• Fix jittering while typing in the Chat tab on Firefox.

Backend updates

• Update llama.cpp to https://github.com/ggml-org/llama.cpp/commit/e434e69183fd9e1031f4445002083178c331a28b
• Update exllamav3 to 0.0.4

Portable builds

Below you can find portable builds: self-contained packages that work with GGUF models (llama.cpp) and require no installation! Just download the right version for your system, unzip, and run.

Which version to download:

• Windows/Linux:
• NVIDIA GPU: Use cuda12.4 for newer GPUs or cuda11.7 for older GPUs and systems with older drivers.
• AMD/Intel GPU: Use vulkan builds.

• CPU only: Use cpu builds.

• Mac:

• Apple Silicon: Use macos-arm64.
• Intel CPU: Use macos-x86_64.

Updating a portable install:

1. Download and unzip the latest version.
2. Replace the user_data folder with the one in your existing install. All your settings and models will be moved.