# Tools
- [ComfyUI](https://github.com/comfyanonymous/ComfyUI)
- [sd-mecha](https://github.com/ljleb/sd-mecha)
- [sd-mecha for Comfyui](https://github.com/ljleb/comfy-mecha)

# Checkpoints
- [NoobAI-XL (NAI-XL) V-Pred-1.0-Version](https://civitai.com/models/833294?modelVersionId=1190596)
- [NoobAI-XL (NAI-XL) V-Pred-0.75S-Version](https://civitai.com/models/833294?modelVersionId=1140829)
- [NoobAI-XL (NAI-XL) V-Pred-0.65S-Version](https://civitai.com/models/833294?modelVersionId=1093948)
- [NoobAI-XL (NAI-XL) V-Pred-0.6-Version](https://civitai.com/models/833294?modelVersionId=1070239)
- [NoobAI-XL (NAI-XL) Epsilon-pred 1.1-version](https://civitai.com/models/833294?modelVersionId=1116447)
- [NoobAI-XL (NAI-XL) Epsilon-pred 1.0-version](https://civitai.com/models/833294?modelVersionId=1022833)
- [Hikari Noob v-pred 1.0.0](https://civitai.com/models/938672?modelVersionId=1201414)
- [Rillusm - Realistic Illustrious v1.0](https://civitai.com/models/959736?modelVersionId=1074511)
- [Hassaku XL (Illustrious) V1.2 - Style](https://civitai.com/models/140272?modelVersionId=1200105)
- [Raehoshi illust XL v3.0](https://civitai.com/models/846917?modelVersionId=1151906)
- [NTR MIX | illustrious-XL | Noob-XL XIII](https://civitai.com/models/926443?modelVersionId=1166878)
- [LHC v-Pred-alpha0.4](https://civitai.com/models/940235?modelVersionId=1146192)
- [Cat Tower (NoobAI XL checkpoint) v1.5(vPred)](https://civitai.com/models/920709?modelVersionId=1200764)
- [Obsession (Illustrious-XL) v-pred_v1.0](https://civitai.com/models/820208?modelVersionId=1198119)
- [Illustrious XL v0.1](https://huggingface.co/OnomaAIResearch/Illustrious-xl-early-release-v0)
- [Shiitake-Mix v1.0](https://civitai.com/models/1066229?modelVersionId=1196701)
- [illustrious_pencil-XL-v2.0.0](https://huggingface.co/bluepen5805/illustrious_pencil-XL)

# Recipe

1. Merge using `sd-mecha for ComfyUI`. ( below sd-mecha format recipe )
```
model "sdxl\noob_v_pencil-XL-1.0.safetensors" "sdxl" "base"
model "sdxl\illustrious_pencil-XL-v2.0.0.safetensors" "sdxl" "base"
model "sdxl\rillusmRealistic_v10.safetensors" "sdxl" "base"
model "sdxl\noobaiXLNAIXL_epsilonPred10Version.safetensors" "sdxl" "base"
merge "train_difference" &1 &2 &3 alpha=1.1
model "sdxl\noobaiXLNAIXL_vPred10Version.safetensors" "sdxl" "base"
merge "train_difference" &4 &5 &3 alpha=1.1500000000000001
merge "add_cosine_a" &0 &6 alpha=0.65
merge "add_cosine_a" &7 &5 alpha=0.0
model "sdxl\noob_v_pencil-XL-v0.5.1.safetensors" "sdxl" "base"
model "sdxl\noobaiXLNAIXL_vPred075SVersion.safetensors" "sdxl" "base"
merge "clamped_add_opposite" &8 &9 &10 alpha=1.5
merge "slerp" &8 &11 alpha=0.6
model "sdxl\hassakuXLIllustrious_v12Style.safetensors" "sdxl" "base"
model "sdxl\Illustrious-XL-v0.1.safetensors" "sdxl" "base"
merge "clamped_add_opposite" &12 &13 &14 alpha=1.0
model "sdxl\raehoshiIllustXL_v30.safetensors" "sdxl" "base"
model "sdxl\noobaiXLNAIXL_epsilonPred11Version.safetensors" "sdxl" "base"
merge "clamped_add_opposite" &15 &16 &17 alpha=0.75
model "sdxl\ntrMIXIllustriousXL_xiii.safetensors" "sdxl" "base"
merge "clamped_add_opposite" &18 &19 &17 alpha=1.0
model "sdxl\shiitakeMix_v10.safetensors" "sdxl" "base"
merge "train_difference" &21 &5 &14 alpha=1.0
merge "add_cosine_a" &20 &22 alpha=0.25
merge "train_difference" &23 &0 &3 alpha=1.0
merge "slerp" &24 &23 alpha=0.7000000000000001
model "sdxl\catTowerNoobaiXL_v15Vpred.safetensors" "sdxl" "base"
merge "add_cosine_a" &25 &26 alpha=0.15
model "sdxl\obsessionIllustrious_vPredV10.safetensors" "sdxl" "base"
merge "add_cosine_a" &27 &28 alpha=0.2
merge "subtract" &25 &5
model "sdxl\lhc_vPredAlpha04.safetensors" "sdxl" "base"
model "sdxl\noobaiXLNAIXL_vPred06Version.safetensors" "sdxl" "base"
merge "clamped_add_opposite" &25 &31 &32 alpha=0.75
model "sdxl\noob_v_pencil-XL-v0.3.0.safetensors" "sdxl" "base"
model "sdxl\noobaiXLNAIXL_vPred065SVersion.safetensors" "sdxl" "base"
merge "clamped_add_opposite" &33 &34 &35 alpha=1.25
merge "subtract" &36 &5
merge "perpendicular_component" &30 &37
merge "add_difference" &29 &38 alpha=1.0
model "concept_sliders_1.safetensors" "sdxl" "lora"
merge "add_difference" &0 &40 alpha=1.0
merge "train_difference" &39 &41 &0 alpha=1.0
model "concept_sliders_2.safetensors" "sdxl" "lora"
merge "add_difference" &0 &43 alpha=4.0
merge "train_difference" &42 &44 &0 alpha=1.0
model "concept_sliders_3.safetensors" "sdxl" "lora"
merge "add_difference" &0 &46 alpha=16.0
merge "train_difference" &45 &47 &0 alpha=1.0
merge "add_opposite" &48 &25 &5 alpha=1.5
model "sdxl\hikariNoobVPred_100.safetensors" "sdxl" "base"
merge "add_opposite" &49 &50 &5 alpha=0.5
merge "subtract" &5 &3
merge "add_difference" &51 &52 alpha=0.1
merge "add_cosine_a" &53 &0 alpha=0.025
merge "add_cosine_b" &54 &12 alpha=0.45
merge "add_cosine_b" &55 &9 alpha=0.1
merge "add_cosine_a" &56 &5 alpha=0.225
dict sdxl_txt2_default=0.0 sdxl_unet_default=0.0 sdxl_txt_default=1.0
merge "weighted_sum" &57 &0 alpha=&58
```

2. Add `v_pred` and `ztsnr` keys.

```python
BASE_MODEL_NAME = "noob_v_pencil-XL-v1.2.0"

tensors = {}
with safe_open(BASE_MODEL_NAME + "-base.safetensors", framework="pt", device="cpu") as f:
    for key in f.keys():
        tensors[key] = f.get_tensor(key)

tensors["v_pred"] = torch.tensor([0.0])
tensors["ztsnr"] = torch.tensor([0.0])

save_file(tensors, BASE_MODEL_NAME + ".safetensors")
```