Klein 9B lcs consistency lora 20260415RunningHub:Lora:<a target="_blank" rel="ugc" href="https://www.runninghub.ai/model/public/2044762492517228546?inviteCode=rh-v1279">https://www.runninghub.ai/model/public/2044762492517228546?inviteCode=rh-v1279</a>Online Workflow：<a target="_blank" rel="ugc" href="https://www.runninghub.ai/post/2016109710943981570/?inviteCode=rh-v1279">https://www.runninghub.ai/post/2016109710943981570/?inviteCode=rh-v1279</a>Klein 9B consist lcs preivew 20260328preview version of lcs training method.Further version is under developmentKlein 4B Consistency LoraThis lora inspired by ByteDance <a target="_blank" rel="ugc" href="https://github.com/PKU-YuanGroup/Helios">Heilos</a> Easy Anti-Drifting method.I implemented Frame-Aware Corrupt (Reference Latent Corrupt) latent and Color Statistics Loss.In additionally, I use frequence split dataset which created by converting an image into frequence domain and split it into high freq image and low freq image.Reference: low freq image, high freq image (0.8 dropout rate on high freq image)Target: clean imagelr: 2e-4scheduler: cosineoptimizer: adamw8bitsteps: 6000Suggested Prompt:{edit prompt}. Transform the image to realistic photograph. add realistic details to the corrupted image. restore high frequence details from the corrupted image.Suggested Strength:0.5-0.7Trained on 4B based. This lora aims to preserve image consistence across multiple input images. Control the lora via lora strength.More strength means more consistence but it would also hurt edit abilities.Different image might has different optimal lora strength.Running HubFlux2 Klein 4B一致性lora对比 runninghub.ai/post/2032812180667633666/?inviteCode=rh-v1279Flux2 Klein 4B一致性lora 不偏色 不偏移 runninghub.ai/post/2032817113190113281/?inviteCode=rh-v1279Flux2 Klein 9B Consist LoraHuggingFace:<a target="_blank" rel="ugc" href="https://huggingface.co/lrzjason/Consistance_Edit_Lora">https://huggingface.co/lrzjason/Consistance_Edit_Lora</a>Klein 9B Consis LoraSuggested Prompt:restore image detailsFlux2 Klein 9B Consis LoraTrained on 9B based. This lora aims to preserve image consistence across multiple input images. Control the lora via lora strength.Suggested strength: 0.2~1.5Use with any prompts.More strength means more consistence but it would also hurt edit abilities.Different image might has different optimal lora strength.RH:<a target="_blank" rel="ugc" href="https://www.runninghub.ai/post/2016109710943981570/?inviteCode=rh-v1279">https://www.runninghub.ai/post/2016109710943981570/?inviteCode=rh-v1279</a>======================QE2511================================RH:<a target="_blank" rel="ugc" href="https://www.runninghub.ai/model/public/2010085010870640641?inviteCode=rh-v1279">https://www.runninghub.ai/model/public/2010085010870640641?inviteCode=rh-v1279</a>Custom_node:<a target="_blank" rel="ugc" href="https://github.com/lrzjason/Comfyui-QwenEditUtils">https://github.com/lrzjason/Comfyui-QwenEditUtils</a>QE2511 Alpha is trained on QwenEdit 2511.As same as previous lora, this lora helps to maintain image consistence.The lora is trained using nf4 qlora.V2 lora is trained for QwenEditPlus aka QwenImage Edit (2509)It is recommanded to use with strength between 0.5 ~1.0 for multiple images0~0.5 for single imageIt needs to use with workflow v4.0 <a target="_blank" rel="ugc" href="https://civarchive.com/models/1939540?modelVersionId=2256750">https://civarchive.com/models/1939540</a>RH workflow:QwenEdit Consistence Edit workflow v4.0Experience link: <a target="_blank" rel="ugc" href="https://www.runninghub.ai/post/1965033238311915522">https://www.runninghub.ai/post/1965033238311915522</a><a target="_blank" rel="ugc" href="https://www.runninghub.ai/post/1965033238311915522/?inviteCode=rh-v1279">/?inviteCode=rh-v1279</a>QwenEdit Consistence Edit workflow v2.0Solved pixels shift issue and keep qwenvl encode image ability.RH workflow:QwenEdit Consistence Edit workflow v2.0Experience link: <a target="_blank" rel="ugc" href="https://www.runninghub.ai/post/1965033238311915522/?inviteCode=rh-v1279">https://www.runninghub.ai/post/1965033238311915522/?inviteCode=rh-v1279</a>Trained with no ref image into qwenvl.This lora typical to test the consistency with kontext like workflow.It is suggested to use 0.45~0.5 lora stregthWorkflow:<a target="_blank" rel="ugc" href="https://civarchive.com/models/1939540/qwenedit-consistance-edit-workflow">https://civarchive.com/models/1939540/qwenedit-consistance-edit-workflow</a>RH workflow:It's extremely fun and I highly recommend it to you!Workflow: QwenEdit Consistence Edit workflowExperience link: <a target="_blank" rel="ugc" href="https://www.runninghub.ai/post/1965033238311915522/?inviteCode=rh-v1279">https://www.runninghub.ai/post/1965033238311915522/?inviteCode=rh-v1279</a>

Consistence Edit Lora

f2k_consist_20260225.safetensors

f2k_4B_consist_20260314.safetensors

Flux2 Klein 4B 20260314

Klein 9B lcs 20260416

Klein 9B lcs preview

Flux2 Klein 9B 20260225

F2K Workflow

Flux2 Klein 9B

qe2511 alpha

v2.0

v1.0

Files with SHA256: b76181b7c07093e287a959ae1cfcd5e2e666dc567bfb98c3178674d9d7e2d79f