| { | |
| "format_version": 1, | |
| "model": "flux-3-image-klein", | |
| "quantization": { | |
| "group_size": 64, | |
| "bits": 6, | |
| "mode": "affine", | |
| "bf16_layers": [ | |
| "*modulation*", | |
| "time_in", | |
| "ref_time_in", | |
| "emb_in", | |
| "txt_in", | |
| "final_layer" | |
| ] | |
| }, | |
| "torch_dtype": "bfloat16", | |
| "dit": { | |
| "hidden": 3072, | |
| "heads": 24, | |
| "depth": 5, | |
| "depth_single": 28, | |
| "mlp_ratio": 3.0, | |
| "axes_dim": [ | |
| 32, | |
| 32, | |
| 32, | |
| 32 | |
| ], | |
| "theta": 10000, | |
| "in_channels": { | |
| "image": 128, | |
| "image_cond": 128, | |
| "text": 20480 | |
| }, | |
| "out_modes": [ | |
| "image" | |
| ], | |
| "vec_in_dim": null, | |
| "image_rope_scale": 0.5, | |
| "flow_map": true | |
| }, | |
| "text_encoder": { | |
| "path": "text_encoder", | |
| "tokenizer": "text_encoder", | |
| "layers": [ | |
| 4, | |
| 8, | |
| 12, | |
| 16, | |
| 20, | |
| 24, | |
| 28, | |
| 32 | |
| ], | |
| "pad_multiple": 80, | |
| "max_tokens": 8192, | |
| "dtype": "bfloat16" | |
| }, | |
| "vae": { | |
| "path": "ae.safetensors", | |
| "dtype": "float32" | |
| }, | |
| "sampler": { | |
| "steps": 16, | |
| "guidance": null, | |
| "alpha": 6.93, | |
| "size": [ | |
| 1024, | |
| 1024 | |
| ] | |
| } | |
| } | |