dt-flux2-klein-4b
FLUX.2 [klein] 4B weights in the Draw Things (s4nnc .ckpt)
format, mirrored unmodified from the Draw Things model CDN for use by JoyFusion's on-device inference. The conversion and
quantization were done by the Draw Things project; the files here are byte-identical to the upstream ones.
| File | Variant | Size | SHA-256 |
|---|---|---|---|
flux_2_klein_4b_q6p.ckpt |
6-bit | 3,117,088,768 | ba42005a6d4e6c48430787db98ae8913aced4654682b59d0d0f1d9ccf820fe30 |
The shared components live in JoyFusionAI/dt-components:
| File | Role | Size | SHA-256 |
|---|---|---|---|
qwen_3_4b_q8p.ckpt |
Text encoder (Qwen3-4B, 8-bit) | 4,535,328,768 | a24e3f832917aafe7f6186e6dfb96ab19a4a0729cdf60a875b5da20024333350 |
flux_2_vae_f16.ckpt |
FLUX.2 VAE | 168,534,016 | 48369d4b1495aec3df0579591f4bf9c199a73cc288130df2359c4a173c5ae864 |
Upstream's recommended settings: 4 steps, text guidance 1, DDIM Trailing, shift 3. The model supports text-to-image, image editing and multi-reference composition.
License
FLUX.2 [klein] 4B, including its text encoder and VAE, is released by Black Forest Labs under the Apache License 2.0
(see LICENSE.md). These files are unmodified copies of the Draw Things conversions.
Model tree for JoyFusionAI/dt-flux2-klein-4b
Base model
black-forest-labs/FLUX.2-klein-4B