fix VAM: funasr+pyarrow deadlock, librosa+numpy compat, reduce iter_batches batch_size

feat: add VAM T2A config variant using omni-v initialization

feat: rename & add VAM mini pipeline configs

feat: add VAM SFT mini training (T2A Stage 1)

feat: add VLM SFT mini configs, resume training, dataset streaming, and eval fixes

refactor: rename VLM configs, add VLM SFT configs, and improve model loading

refactor: reorganize dataset into lm/vlm/vam subdirectories

reorganize configs into subdirs lm/vlm/vam, remove runs/, add interview docs for pretrain & full_sft

refactor(configs): flatten configs/model/ into configs/

Set base model hidden_size to 512 across all configs for 8G VRAM

Restructure repo: rename classes, split datasets, organize trainers by modality, add YAML config + root trainer entry points

Fuse minimind/minimind-v/minimind-o into omni with layered architecture