Views
No views yet
Efficient-Large-Model/Sana_Sprint_1.6B_1024px_diffusers
for the SceneWorks native MLX worker (mlx-gen-sana).
SANA-Sprint is the few-step, CFG-free distillation of SANA (continuous-time
consistency / SCM sampler + a guidance-embedding trunk) — ~2-step 1024×1024 gen.SanaPipeline::new_sprint / from_snapshot expects:transformer/diffusion_pytorch_model.safetensors # SANA-Sprint 1.6B Linear-DiT trunk (BF16)
vae/diffusion_pytorch_model.safetensors # 32x DC-AE f32c32 decoder (F32)
text_encoder/gemma-2-2b-it.safetensors # gemma-2-2b-it CHI caption encoder (BF16)
text_encoder/tokenizer.jsonff sub-module
keys (inverted_conv.conv/depth_conv.conv/point_conv.conv →
conv_inverted/conv_depth/conv_point); tensor bytes are unchanged. The
gemma-2-2b-it text encoder is sourced from the un-gated
SceneWorks/gemma-2-2b-it
mirror (merged from its two shards) so the snapshot is self-contained.LICENSE and NOTICE). This is a non-commercial license:
the model and its outputs are for research and evaluation use only. By
downloading these weights you agree to the NVIDIA Open Model License terms. The
bundled gemma-2-2b-it encoder is under the Google Gemma Terms of Use.