Drop attention_mask at inference: the checkpoint was trained without one, and generate()'s auto-mask degenerates the output

Serve with greedy decoding: sampling hallucinates on this lightly-tuned checkpoint

Upload README.md with huggingface_hub

count wording: attention types, not mechanisms

list 7Attn and 11Attn in Space models metadata

add Aether family section (7Attn, 11Attn, checkpoints, blog)

sync Aether family cross-links

Upload README.md with huggingface_hub

clarify: the live demo answers with the -it model, not the base

fix: unwrap BatchEncoding from apply_chat_template before generate()

diagnostics: full traceback in gen_error

diagnostics: surface gen_error + vram in /api/health

serving: manual load + GenerationMixin + use_cache=False + streaming

add blog article badge

link intermediate-checkpoints dataset

serving: manual load + GenerationMixin + use_cache=False + streaming

serving: manual load + GenerationMixin + use_cache=False + streaming

serving: manual load + GenerationMixin + use_cache=False + streaming

fix README short_description

add index.html

add app.py

add requirements.txt

add Dockerfile

initial commit