docs: gensim attribution and pretrain script in README

fix: robust text8 loading, gensim attribution in UI, training error handling

fix: LR schedule, analogy examples, remove old pretrain.py

fix: LR schedule + analogy tab defaults

feat: gensim-trained embeddings on full text8 (17M words)

cleanup: remove .venv, __pycache__, text8

Upload folder using huggingface_hub

initial commit