cleanup: remove pytorch_model.bin (superseded by model.safetensors at step 44k)

fix: replace step-23k model.safetensors with step-44k (final ckpt, eval_loss 3.12)

Add working list-of-tokens inference example

Add model card

Upload rna_celltype BERT medium fine-tune

docs: prominent input-format warning (this is a gene-ID model, not nucleotide)

fix: add [UNK] (and [PAD]) to WordLevel vocab to enable text-mode encoding

Add working WordLevel-tokenizer inference example

Add model card

Upload rna_celltype BERT medium fine-tune (early-stopped at 23k steps)

initial commit