⚡Join my FREE PATREON for experiments 🧪 My Social Links!
About this:
This Workflow is a 2nd Pass workflows that works for 2 pairs of different models. By "2 pairs" of models I mean "Kroma Base and Turbo" or "Krea 2 Raw and Turbo". How it works: It begins to generate using Kroma Base (Or Krea Raw) and continues with Turbo, so at the beginning you can use negative prompts. Then it will upscale it 2x. It uses a nice UltraReal (Lenovo) LoRA to make it more realistic.
Structure
📂 models/
└── 📂 diffusion_models/
└── kroma-v0.3.1-turbo-opd.safe...
└── kroma-v0.3-base-int8...
📂 text-encoder/
└── Huihui-Qwen3-VL-4B-Instruct...
📂 vae/
└── wan_2.1_vae.safetensors
📂 lora/
└── lenovo_krea2.safeten...
📂 custom_nodes/
└── 📂 comfyui-krea2-conditioning/
Usando Krea 2?
📂 models/
└── 📂 diffusion_models/
└── krea2_turbo_fp8_scaled
└── krea2_raw_fp8_scaled.safe...
Download links
🔗 kroma-v0.3.1-turbo-opd
https://huggingface.co/lodestones/Kroma/blob/main/kroma-v0.3.1-turbo-opd.safetensors
🔗 kroma-v0.3-base-INT8...
https://huggingface.co/Stabhappy/kroma-v0.3-base-INT8-convrot/tree/main
🔗 krea2_raw_fp8_scaled
https://huggingface.co/Comfy-Org/Krea-2/blob/main/diffusion_models/krea2_raw_fp8_scaled.safetensors
🔗 krea2_turbo_fp8_scaled
https://huggingface.co/Comfy-Org/Krea-2/blob/main/diffusion_models/krea2_turbo_fp8_scaled.safetensors
🔗 Huihui-Qwen3-VL-4B-Instruct-ablit...
https://huggingface.co/ahmed22xa/Huihui-Qwen3-VL-4B-Instruct-abliterated-comfy/blob/main/Huihui-Qwen3-VL-4B-Instruct-abliterated-fp8_scaled.safetensors
🔗 wan_2.1_vae
https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/blob/main/split_files/vae/wan_2.1_vae.safetensors
🔗 lenovo_krea2
https://civarchive.com/models/1662740/lenovo-ultrareal?modelVersionId=3075606
About the conditioner custom node:
You only need this if you're going to use Krea 2 but not if you're using Kroma!
🔗 ComfyUI-Krea2-Conditioning
https://github.com/huwhitememes/comfyui-krea2-conditioning
⚡Join my FREE PATREON for experiments 🧪 My Social Links!