---
title: Qwen-Image2.1-ZZZ-chibi
canonical_url: "https://www.modelscope.cn/models/wesjos/Qwen-Image2.1-ZZZ-chibi"
md_url: "https://www.modelscope.cn/models/wesjos/Qwen-Image2.1-ZZZ-chibi.md"
repository: wesjos/Qwen-Image2.1-ZZZ-chibi
chinese_name: "Qwen Image 2.1 绝区零Chibi风格LoRA"
last_updated: 2026-10-02
license: apache-2.0
pipeline_tag: text-to-image-synthesis
tasks:
  - text-to-image-synthesis
base_model:
  - Qwen/Qwen-Image-2.1
base_model_relation: adapter
parameters: 39.8M
tensor_type:
  - F16
library_name:
  - lora
  - safetensors
language:
  - en
  - zh
supports_inference: txt2img
downloads: 5
stars: 1
tags:
  - text-to-image
  - lora
  - qwen-image
  - style-lora
  - chibi
  - q-version
  - zenless-zone-zero
  - game-style
  - comfyui
  - ai-toolkit
---

# Qwen-Image2.1-ZZZ-chibi

> Qwen-Image2.1-ZZZ-chibi - wesjos 在 ModelScope 开源的模型。Zenless Zone Zero Chibi LoRA — Qwen-Image-2.1

wesjos/Qwen-Image2.1-ZZZ-chibi 是 ModelScope 魔搭社区上的 39.8M 参数text-to-image-synthesis模型，采用 apache-2.0 许可，基于 Qwen/Qwen-Image-2.1 构建，并支持在线推理（txt2img）。

- **Repository**: wesjos/Qwen-Image2.1-ZZZ-chibi
- **License**: apache-2.0
- **Tasks**: text-to-image-synthesis
- **Parameters**: 39.8M
- **Base model**: Qwen/Qwen-Image-2.1
- **Online inference**: txt2img
- **Tags**: text-to-image, lora, qwen-image, style-lora, chibi, q-version, zenless-zone-zero, game-style, comfyui, ai-toolkit
- **Downloads**: 5
- **Stars**: 1
- **Last updated**: 2026-10-02

Source: https://www.modelscope.cn/models/wesjos/Qwen-Image2.1-ZZZ-chibi

---

# Zenless Zone Zero Chibi LoRA — Qwen-Image-2.1

> note: base model is actually from ComfyUI-org int8 version.
> trigger `zenlesszonezero_chibi`.


[![License: Apache-2.0](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](https://huggingface.co/wesjos/Qwen-Image2.1-ZenlessZoneZeroChibi-Lora/blob/main/LICENSE)
[![Base model: Qwen-Image-2.1](https://img.shields.io/badge/Base%20model-Qwen--Image--2.1-ff69b4)](https://huggingface.co/Qwen/Qwen-Image-2.1)
[![Type: LoRA adapter](https://img.shields.io/badge/Type-LoRA%20adapter-ff69b4)](https://huggingface.co/docs/diffusers/using-diffusers/loading_lora)

---

## 📋 Model Details

| Property | Value |
|---|---|
| **Base model** | [Qwen-Image-2.1](https://huggingface.co/Qwen/Qwen-Image-2.1) (`qwen_image_2`) |
| **Type** | Style LoRA — chibi / Q-version proportion compression |
| **Trigger word** | `zenlesszonezero_chibi` — optional; the style can also transfer naturally |
| **Rank** | 16 |
| **Tensors** | 384 (192 pairs of `lora_A`/`lora_B`) |
| **Target modules** | Per layer: `attn.to_q` · `to_k` · `to_v` · `to_out.0` · `img_mlp.gate_up` · `img_mlp.out` |
| **Blocks covered** | All 32 layers `transformer_blocks.0–31` (no missing, no extra layers) |
| **Dtype** | F16 |
| **Alpha** | Not written → ComfyUI / diffusers treat as `alpha = rank = 16` (i.e., scale 1.0) |
| **Training framework** | [ai-toolkit](https://github.com/ostris/ai-toolkit) v0.13.21 · flowmatch |
| **Trained on** | int8 convrot quantized base model (Comfy-Org) |
| **Size per file** | 76 MiB |

---

## 📦 Checkpoint Guide

| File | Step / Epoch | Recommendation |
|---|---|---|
| `zzz_chibi_v2_st2000.safetensors` | 2000 / 8 | ✅ Safe bet; style is formed, least underfitting |
| `zzz_chibi_v2_st2500.safetensors` | 2500 / 11 | ⭐ **Recommended default** — most example images come from this checkpoint |
| `zzz_chibi_v2_st3000.safetensors` | 3000 / 13 | ✅ Strongest style; chibi proportions are flatter and more "sticker-like" |

---

## 🚀 Usage

### Prompt

```
zenlesszonezero_chibi, nahida, shy
zenlesszonezero_chibi, a girl with fox tail, shy
```

Both English and Chinese prompts work; character names can be written directly (`nahida` / `furina` / `不知火舞` have all been tested).

### LoRA strength

| Scale | Effect |
|---|---|
| 0.55 – 0.65 | Mild chibi; retains more original detail and compositional freedom |
| **0.70** | **Empirical sweet spot** — chibi features are clear, image is not overly compressed |
| 0.80 – 0.90 | Stronger chibi compression; head-to-body ratio becomes more extreme; above 1.0 composition starts to suffer |

> `07` / `08` are strictly controlled strength comparison references: same prompt, same checkpoint (st2500), same strength 0.70, **seeds differ by only 1**, allowing visual comparison of randomness.
> `04` / `05` share a prompt but differ in seed, checkpoint, and strength — **not** a strict strength comparison experiment.

### ComfyUI

**Base model files** (`diffusion_models` + `text_encoder` + `vae`):

| Slot | File |
|---|---|
| diffusion model | `qwen_image_2.1_int8_convrot.safetensors` |
| text encoder | `qwen3vl_8b_int8_convrot.safetensors` |
| VAE | `qwen_image_2.1_vae_bf16.safetensors` |


**KSampler parameters** (measured values from this repo's example images):

| Parameter | Value |
|---|---|
| Steps | 25 |
| CFG | 5.0 |
| Sampler | `euler` |
| Scheduler | `simple` |
| Shift (ModelSamplingAuraFlow) | 3.0 |
| Size | 1024 × 1024 |

### Negative prompt


```
deformed, wrong hands, mutated, fused fingers, extra hand, extra feet, ugly underwear
```

---

## ⚠️ Limitations

1. **Training set is mainly single-character half-body / portrait shots** — multi-character scenes, male characters, and full-body compositions generalize poorly.
2. **Trained on an int8 quantized base model** — inherits some of convrot8's texture loss; detail sharpness is slightly lower than fp16 training.
3. **Chibi style actively simplifies** — complex accessories and multi-layered clothing patterns are easily flattened; lower strength below 0.6 when detail matters.
4. **Generated content may not match official lore** — character details and clothing patterns should defer to official source material.

---