4 days ago
-
[UPDATE] Add the matched 1D-RoPE twin (twin_1d/) and load(variant=...)
-
[UPDATE] Loader handles git-lfs pointer files; hf download instructions; limitation note
-
[UPDATE] Restructure release: safetensors weights, cbt package, chat demo, clean model card
5 days ago
-
[UPDATE] Replace weights with the final SFT model; update paper, model card and evaluation results
-
[UPDATE] Replace weights with the V5 model; update paper, model card and evaluation results
6 days ago
-
[UPDATE] Update CBT-18.95M paper, model card, and 16-100d random extrapolation results
7 days ago
-
[UPDATE] Update CBT-18.95M paper, model card, and 16-100d random extrapolation results
-
[UPDATE] Update CBT-18.95M paper, model card, and 16-100d random extrapolation results
-
[DOCS] Update model card: add GitHub and Paper links, remove hardware specific notes
-
[DOCS] Standardize Character Based Transformer (CBT) nomenclature and pure markdown
-
[DOCS] Rename to Character Based Transformer (CBT) and remove pipeline_tag
-
[DOCS] Clean LaTeX formatting for Hugging Face markdown compatibility
-
[RELEASE] Upload ScaledGPT 18.95M champion model weights and artifacts
