Increase GPU duration for score_and_create_matrix_all_singles to enhance performance; implement fallback to CPU mode on failure

Enhance model loading in transformers with improved error handling and fallback mechanisms; implement compatibility checks for trust_remote_code parameter

Enhance URL handling in transformers by patching multiple functions for comprehensive URL fixing; improve requests.get handling for malformed URLs

Enhance URL handling in transformers with comprehensive validation and retry mechanism; implement cache file validation and cleaning process

Refactor URL handling in transformers and remove app_wrapper script for Zero GPU management

Implement direct model loading with manual config handling and enhance cache management for model loading

Add patch for transformers URL handling and enhance model loading with manual config download

Enhance model loading process with improved error handling and alternative loading method; configure Hugging Face endpoint and disable offline mode

Implement runtime mode tracking and retry mechanism for Zero GPU initialization, falling back to CPU mode if necessary

Add error handling for Zero GPU initialization and fallback to CPU mode; update Gradio version in requirements

Remove model preloading at startup to prevent Zero GPU initialization issues; models will now load and cache on first use.

Refactor scoring matrix visualization to simplify font and figure size adjustments, ensuring consistent label formatting and improved clarity.

Enhance scoring matrix visualization with dynamic font and figure sizing based on mutation range

Implement model caching and preload functionality to optimize loading in Zero GPU spaces

Add external keep-alive script for Tranception Space and update settings

Add keep-alive scripts and environment configuration for Tranception Space

Update README and app.py to enhance project description and technical details for BASIS-China iGEM 2025 deployment of Tranception

Implement mock spaces module for Zero GPU fallback and improve error handling for GPU decorator application

Enhance Zero GPU support with environment variable control and improved error handling

Add debugging info and Zero GPU hardware suggestion

Fix Zero GPU compatibility and add conditional support

Add Zero GPU support for H200 GPU on Hugging Face Spaces

Fix resource management and memory leaks

Update model selection for HF Spaces to use Large model

Update model selection for HF Spaces to use Large model and add local settings configuration

Add resource management and cancellation support

Fix concurrency issues by adding unique identifiers to all generated files

Fix extract_sequence function signature for Gradio compatibility

Change default model from Small to Large

Create app.py