Detail
Rekomendasi
v1.1
SDXLMixSampler (Workflow)
SDXLMixSampler - ComfyUI JNode (Workflow Included)

SDXLMixSampler - ComfyUI JNode (Workflow Included)

0
4
0
#ComfyUI
#tool
#SDXL

To support my work, you could buy me a coffee

https://www.buymeacoffee.com/JasonAICreator

SDXLMixSampler 1.1

The node updated to accept loop:number rather total_steps:number

The total_steps would be calculated as mix_steps*loop

add final_only input to determine output multiple images or only the final loop image

If fina_only set to no, it would produce each loop image in output

SDXLMixSampler

The concept is to make one node to use both base+refiner like simple sampler.

Let me explains some parameter.

total_steps: total steps would run with this sampler. default is 20.

mixing_steps: total steps of one loop, loop = total_steps / mixing steps

base_steps_percentage: the percentage of base model's steps in each loop, the refiner would take the leftover

The whole pocess would only add_noise at the first base model sampling, other would disable_noise. The return_with_leftover_noise would be false at the last refiner node.

Why would I create this node? I was testing around generating text and I found it helps to generating text in multiple loop of the base+refiner process and I think it could done by 1 node.

All my custom_nodes would goto the flowing git page:

https://github.com/lrzjason/ComfyUIJasonNode/tree/main

Contact:

欢迎加群讨论AI相关的话题,遇到问题有能力的我会尽量解答

验证:小志Jason 

Lihat terjemahan

Rating dan Ulasan

-- /5
0 rating

Belum menerima penilaian atau komentar yang cukup

no-data
Tidak Ada Data
X
Pengumuman
2024-07-25
Memposting model
2023-07-30
Perbarui informasi model
Detail model
Jenis
Other
Waktu publikasi
2023-07-30
Model Dasar
SDXL 1.0
Perkenalan versi

Changelog:

new

final_only:

only output the final loop image. If it set to no, it would produce each loop output and use more time to generate each output.

update

Total Steps->loop

it detemine how many time of the base->refiner process would run.

mixing_steps:

default:20, min:10, steps:1

able to set other steps rather than 20*x

Cakupan Lisensi
Model Source: civitai

1. Hak untuk model yang diposting ulang adalah milik pembuat aslinya.

2. Pencipta asli yang ingin mengklaim model harap hubungi staf SeaArt AI melalui saluran resmi. Klik untuk mengklaim

Cakupan Lisensi Penciptaan
Gambar Online
Melakukan Penggabungan
Izinkan Unduhan
Lisensi Komersial
Gambar yang dihasilkan dapat dijual atau digunakan untuk tujuan komersial
Izinkan model dijual kembali atau dijual setelah penggabungan
QR Code
Unduh SeaArt App
Lanjutkan perjalanan kreasi AI Anda di perangkat mobile