Files

70 lines
1.5 KiB
Markdown

# Example Workflow Notes
## Split-Sigma Sampling
```text
Scheduler
-> SplitSigmas
SplitSigmas high_sigmas
-> Sigmas Resample
-> ZIB EmberFrame SamplerCustomAdvanced Capture sigmas
ZIB output
-> ZIT EmberFrame SamplerCustomAdvanced Capture latent_image
SplitSigmas low_sigmas
-> ZIT EmberFrame SamplerCustomAdvanced Capture sigmas
```
## PiD Final-Latent Decode
```text
ZIT output or denoised_output
-> EmberFrame Normalize ZImage/Flux Latent
-> PiD Decode latent
```
Use:
```text
PiD Decode backbone = zimage
PiD Decode pid_ckpt_type = 2k or 2kto4k
PiD Decode scale = 4
PiD Decode pid_steps = 4
PiD Decode cfg_scale = 1
PiD Decode sigma = 0
```
## PiD Staged Intermediate Decode
```text
ZIT pid_latent
-> PiD Prepare latent
ZIT pid_sigma
-> PiD Prepare sigma
PiD Prepare
-> PiD Sample
-> PiD Finalize
```
Start diagnostics with:
```text
capture_mode = from_end
capture_step = 1
capture_tensor_source = x
sigma_source = next
capture_debug = true
```
For Z-Image / Flux PiD, keep PiD scale at `4`.
## Detailer File Notes
`z-image_5.1_emberframe_pid.json` uses the Impact detailer pass. Its workflow note lists the detector files and links for `person_yolov8m-seg.pt`, `face_yolov8m.pt`, `Eyes.pt`, and `sam_vit_l_0b3195.pth`.
`z-image_5.2_emberframe_pid.json` uses the SAM3 detailer pass. Its workflow note links the required `models/sam3/sam3.pt` file to the gated `facebook/sam3` Hugging Face repo and notes that ComfyUI needs authenticated access if auto-download is used.