restore: transformers requirements

restore: transformers app (vLLM pilot done; vLLM=SM80+ only)

fix: render-to-text + explicit encode (transformers v5 tokenize=True returns dict)

fix: render-to-text + explicit encode (transformers v5 tokenize=True returns dict)

fix: render-to-text + explicit encode (transformers v5 tokenize=True returns dict)

pilot: vllm requirements

pilot: vLLM engine behind the same /run contract (prefix caching, paged KV)

fix: fallback tool injection for templates that ignore tools kwarg (Phi-4-mini)

fix: fallback tool injection for templates that ignore tools kwarg (Phi-4-mini)

fix: fallback tool injection for templates that ignore tools kwarg (Phi-4-mini)

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload app.py with huggingface_hub

Upload folder using huggingface_hub

initial commit