Fix: Robust range-resumable downloaders and FastAPI-level stop filters to eliminate prompt template leakage

Fix: Optimize configs, relax system prompts, and improve StarCoder2 prompt template stopping

Update display title in README.md frontmatter to show correct name

feat: configure StarCoder2 3B model deployment configurations and ports

fix: resolve javascript scoping exceptions, adjust button accents, and integrate marked.js with responsive layouts

feat: Add premium chat features including interactive copy buttons, timestamps and full bubble copy controls

Fix UI to permanently dock input row at the bottom and grow scrollable chat list above it

Perfect playground layout positioning welcome message inside chat box and input below it

Enable detailed background server logs for startup debugging

Reposition user input box exactly below the default LLM welcome message

Upgrade backend to support high-performance scaling & concurrency caching

Position input row exact below LLM message using premium chat-messages flex design

Implement Stop button, AbortController, OpenAI-compatible chat backend, and zero-scrollbar layout

Enable real-time interactive chat streaming in dashboard UI

Fix binary detection and setup logic to support cross-platform Linux/Docker environments

perf: limit LLAMA_THREADS to 2 on Linux to prevent cgroup throttling, and override max_tokens < 100 to a minimum of 100 tokens

feat: increase context size to 8192, add robust input validations, and enable cross-platform telemetry matching both llama-server.exe and llama-server

feat: add OpenAI compatibility proxy and beautiful glassmorphism dashboard UI

Fix Dockerfile: copy all compiled libraries and define LD_LIBRARY_PATH

Add Hugging Face Space YAML frontmatter to README.md

Deploy production-grade ultra-fast Qwen 7B Space and GitLab repo