File size: 2,435 Bytes
e83364d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
# 📖 Wan2.1 VACE + Phantom Causvid (Finetune)

**Author / Creator:** [Inner_Reflections_AI](https://civitai.com/articles/17908/guide-wan-vace-phantom-merge-an-inner-reflections-guide)  
**Packaged for WanGP by:** [hdtrnk](https://huggingface.co/hdtrnk)

---

## 🔹 About This Finetune
This is a **custom merge of Wan 2.1 VACE and Phantom**, created by **Inner_Reflections_AI** and purified to **FP16** for compatibility with **WanGP**.  

- **Architecture:** `vace_14B`  
- **Mode:** Image/Video conditioning with multi-reference support  
- **Multi-Image:** Optimized for 2–4 image references in WanGP custom UI  
- **Variants:** Both **FP16** and **INT8** versions provided  

---

## 🔹 Files
- [Wan2.1_VACE_Phantom_Causvid_fp16_pure.safetensors](https://huggingface.co/hdtrnk/VACEPhantomCausvid/blob/main/Wan2.1_VACE_Phantom_Causvid_fp16_pure.safetensors)  
- [Wan2.1_VACE_Phantom_Causvid_quanto_bf16_int8_pure.safetensors](https://huggingface.co/hdtrnk/VACEPhantomCausvid/blob/main/Wan2.1_VACE_Phantom_Causvid_quanto_bf16_int8_pure.safetensors)  

---

## 🔹 Usage in WanGP
Place the JSON definition file into:

```
app/finetunes/vace_phantom_causvid.json
```

Sample finetune entry:

```json
{
  "model": {
    "name": "VACE Phantom Causvid 14B",
    "architecture": "vace_14B",
    "description": "VACE-first UI with Phantom multi-image references enabled. Causvid merge, purified FP16 for clean loading & optional INT8 quantization.",
    "URLs": [
      "https://huggingface.co/hdtrnk/VACEPhantomCausvid/blob/main/Wan2.1_VACE_Phantom_Causvid_fp16_pure.safetensors",
      "https://huggingface.co/hdtrnk/VACEPhantomCausvid/blob/main/Wan2.1_VACE_Phantom_Causvid_quanto_bf16_int8_pure.safetensors"
    ],
    "modules": [],
    "auto_quantize": false
  }
}
```

---

## 🔹 Notes
- This is an **experimental finetune**. You may need to tweak **steps**, **guidance scale**, and **reference image setup** for best results.  
- Works best with **multi-image prompts** where identities are clearly separated (e.g., “Character A on left, Character B on right”).  
- If you encounter **dropdown errors** (`Value: on is not in the list...`), refresh the UI and reselect options.  

---

## 🔹 Credits
- **Merge & Guide:** [Inner_Reflections_AI](https://civitai.com/articles/17908/guide-wan-vace-phantom-merge-an-inner-reflections-guide)  
- **Light JSON Packaging & Uploads:** [hdtrnk](https://huggingface.co/hdtrnk)