Merge dev branch

[oobabooga]

No description provided.

[Copilot]

Pull Request Overview

This PR merges the dev branch, bringing updates to dependency versions and implementing changes for Gradio image handling. The key updates include upgrading various Python packages to newer versions and modifying image-related functionality to use file paths instead of PIL objects.

• Updates multiple Python package versions including transformers, bitsandbytes, and exllamav3
• Changes Gradio image components from PIL to filepath handling for better compatibility
• Adds flash-linear-attention dependency and improves ExLlamav3 model handling

Reviewed Changes

Copilot reviewed 25 out of 25 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
requirements files	Updates llama-cpp-binaries from v0.46.0 to v0.49.0, transformers to 4.57., bitsandbytes to 0.48., and adds flash-linear-attention
modules/ui_chat.py	Changes Gradio Image components from type='pil' to type='filepath'
modules/chat.py	Refactors image handling to work with file paths instead of PIL objects
modules/exllamav3.py	Improves model loading and adds logits processing functionality
modules/exllamav3_hf.py	Streamlines forward pass implementation and fixes sequence handling
modules/models.py	Updates ExLlamav3 loader to return tuple format
modules/torch_utils.py	Adds device detection from model attribute
modules/logits.py	Extends ExLlamav3 support in logits processing
download-model.py	Handles HTTP 416 status code for completed downloads

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}