vae/ is an unmodified, byte-identical copy of black-forest-labs/FLUX.2-klein-base-4B vae/diffusion_pytorch_model.safetensors (sha256 ca70d220...) and its config.json, redistributed under Apache-2.0 with the licence text as LICENSE-FLUX2-VAE.md and attribution in the model card. Users no longer need a separate, gated download to encode or decode latents. The DiT weights are unchanged.
Require the Apache-2.0 Flux.2 VAE from FLUX.2-klein-base-4B
The model card pointed users at ae.safetensors from FLUX.2-dev, whose FLUX Non-Commercial License v2.1 4(a)(iii) forbids all research and development related to surveillance. Black Forest Labs ships the same autoencoder inside FLUX.2-klein-base-4B under Apache-2.0, ungated. The two were verified identical by pairing every tensor on value (250/251 pair, worst deviation 7.802e-03 = bf16 rounding; the odd one out is a BatchNorm step counter), and in bf16 they give bit-identical latents, reconstructions and samples. GeoCore-9B is now Apache-2.0 end to end. No weight file in this repository changed.
Fix released weights: publish trained final state (issue #2)
The previous shards contained an EMA state that was never updated during training (initialization weights) due to a key-prefix mismatch in update_ema under torch.compile + DeepSpeed. This upload replaces them with the final 300K-step training weights, verified to load strictly into Flux2 and to denoise (delta_rms > 0). Adds SHA256SUMS.
