Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer/tokenizer.json filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,118 @@
+---
+license: apache-2.0
+language:
+- en
+- zh
+base_model:
+- Qwen/Qwen-Image-Layered
+pipeline_tag: image-text-to-image
+library_name: mflux
+tags:
+- mflux
+- apple-silicon
+- mlx
+- quantized
+- image-decomposition
+- rgba
+---
+
+# Qwen-Image-Layered (6-bit Quantized for MFLUX)
+
+<p align="center">
+    <img src="https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-Image/layered/qwen-image-layered-logo.png" width="600"/>
+</p>
+
+6-bit quantized version of [Qwen/Qwen-Image-Layered](https://huggingface.co/Qwen/Qwen-Image-Layered) optimized for **Apple Silicon** via [MFLUX](https://github.com/filipstrand/mflux).
+
+| Version | Size | VRAM |
+|---------|------|------|
+| BF16 (original) | ~55 GB | ~55 GB |
+| **6-bit (this)** | **~31 GB** | **~29 GB** |
+
+## What is Qwen-Image-Layered?
+
+Qwen-Image-Layered decomposes any image into multiple **RGBA layers** with transparency. Each layer can be independently edited, moved, resized, or recolored—enabling high-fidelity, consistent image editing.
+
+<p align="center">
+    <img src="https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-Image/layered/layered.JPG" width="800"/>
+</p>
+
+## Quick Start
+
+### Installation
+
+Requires [MFLUX](https://github.com/filipstrand/mflux) with Qwen-Image-Layered support ([PR #302](https://github.com/filipstrand/mflux/pull/302)):
+
+```bash
+# Install from the PR
+git clone https://github.com/zimengxiong/mflux.git
+cd mflux
+uv sync
+```
+
+### Usage
+
+```bash
+uv run flux-generate-qwen-layered \
+  --image input.png \
+  --layers 4 \
+  --steps 50 \
+  -q 6 \
+  --output-dir ./layers
+```
+
+**Output:** 4 RGBA PNG files (`layer_0.png`, `layer_1.png`, etc.) with transparency.
+
+### Parameters
+
+| Parameter | Description | Default |
+|-----------|-------------|---------|
+| `--image` | Input image path | Required |
+| `--layers` | Number of layers to decompose | 4 |
+| `--steps` | Inference steps | 50 |
+| `-q` | Quantization (4, 6, or 8-bit) | None (BF16) |
+| `--resolution` | Resolution bucket (640 or 1024) | 640 |
+| `--output-dir` | Output directory for layers | `./` |
+
+## Example
+
+**Input image:**
+
+![Input](https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen-Image/layered/幻灯片1.JPG)
+
+## Model Architecture
+
+- **Transformer:** QwenImageTransformer2DModel with Layer3D RoPE (3D positional encoding for `[layer, height, width]`)
+- **VAE:** AutoencoderKLQwenImage (RGBA 4-channel with 3D temporal convolutions)
+- **Text Encoder:** Qwen2.5-VL for conditioning
+
+## Hardware Requirements
+
+- **Apple Silicon Mac** (M1/M2/M3/M4 series)
+- **Minimum 24GB unified memory** (32GB+ recommended)
+- macOS 13.0+
+
+## Links
+
+- 📦 [MFLUX](https://github.com/filipstrand/mflux) - Fast MLX inference for Apple Silicon
+- 🤗 [Original Model](https://huggingface.co/Qwen/Qwen-Image-Layered) - Full BF16 weights
+- 📑 [Paper](https://arxiv.org/abs/2512.15603) - Research paper
+- 📑 [Blog](https://qwenlm.github.io/blog/qwen-image-layered/) - Official blog post
+
+## License
+
+Apache 2.0 (same as original model)
+
+## Citation
+
+```bibtex
+@misc{yin2025qwenimagelayered,
+      title={Qwen-Image-Layered: Towards Inherent Editability via Layer Decomposition}, 
+      author={Shengming Yin, Zekai Zhang, Zecheng Tang, Kaiyuan Gao, Xiao Xu, Kun Yan, Jiahao Li, Yilei Chen, Yuxiang Chen, Heung-Yeung Shum, Lionel M. Ni, Jingren Zhou, Junyang Lin, Chenfei Wu},
+      year={2025},
+      eprint={2512.15603},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV},
+      url={https://arxiv.org/abs/2512.15603}, 
+}
+```

text_encoder/0.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5057d57ba51ccbe16b6cb72c2c50fa60790d557cbfce68d3a46fd31ad506a5ec
+size 2080974775

text_encoder/1.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25487b7ac743a1352ec19468fd8bca5887453ceb007b4e9a8f9560a81f229375
+size 2136050841

text_encoder/2.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e7a76ccb09e833bb67bcd2a5bdd6b48d6b61f238fe2ae2d8e4e98b477cbf6956
+size 2059005337

text_encoder/3.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bc59df07bc51cdcfd531128c0a4b806722af27bfac022561f8dfdf0d193d519
+size 2136050889

text_encoder/4.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc4a07b54f3347290cca0a3fc0ffa9f13d5098591b64a34136905ac1d3a35d4e
+size 2059005367

text_encoder/5.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:127fa46f1daf82ae0c24d7ebb1b4b4dd0aef147473283776fd07957d2a366a7c
+size 2136050889

text_encoder/6.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c720f15496d5a400923b5dcf2eab2ed9e04479e999e76ad33be1d26d5901068b
+size 1534150412

text_encoder/model.safetensors.index.json

@@ -0,0 +1,375 @@
+{
+  "metadata": {
+    "quantization_level": "6",
+    "mflux_version": "0.13.3"
+  },
+  "weight_map": {
+    "encoder.embed_tokens.weight": "0.safetensors",
+    "encoder.layers.0.input_layernorm.weight": "0.safetensors",
+    "encoder.layers.0.self_attn.q_proj.weight": "0.safetensors",
+    "encoder.layers.0.self_attn.q_proj.bias": "0.safetensors",
+    "encoder.layers.0.self_attn.k_proj.weight": "0.safetensors",
+    "encoder.layers.0.self_attn.k_proj.bias": "0.safetensors",
+    "encoder.layers.0.self_attn.v_proj.weight": "0.safetensors",
+    "encoder.layers.0.self_attn.v_proj.bias": "0.safetensors",
+    "encoder.layers.0.self_attn.o_proj.weight": "0.safetensors",
+    "encoder.layers.0.self_attn.rotary_emb.inv_freq": "0.safetensors",
+    "encoder.layers.0.post_attention_layernorm.weight": "0.safetensors",
+    "encoder.layers.0.mlp.gate_proj.weight": "0.safetensors",
+    "encoder.layers.0.mlp.up_proj.weight": "0.safetensors",
+    "encoder.layers.0.mlp.down_proj.weight": "0.safetensors",
+    "encoder.layers.1.input_layernorm.weight": "0.safetensors",
+    "encoder.layers.1.self_attn.q_proj.weight": "0.safetensors",
+    "encoder.layers.1.self_attn.q_proj.bias": "0.safetensors",
+    "encoder.layers.1.self_attn.k_proj.weight": "0.safetensors",
+    "encoder.layers.1.self_attn.k_proj.bias": "0.safetensors",
+    "encoder.layers.1.self_attn.v_proj.weight": "0.safetensors",
+    "encoder.layers.1.self_attn.v_proj.bias": "0.safetensors",
+    "encoder.layers.1.self_attn.o_proj.weight": "0.safetensors",
+    "encoder.layers.1.self_attn.rotary_emb.inv_freq": "0.safetensors",
+    "encoder.layers.1.post_attention_layernorm.weight": "0.safetensors",
+    "encoder.layers.1.mlp.gate_proj.weight": "0.safetensors",
+    "encoder.layers.1.mlp.up_proj.weight": "0.safetensors",
+    "encoder.layers.1.mlp.down_proj.weight": "0.safetensors",
+    "encoder.layers.2.input_layernorm.weight": "0.safetensors",
+    "encoder.layers.2.self_attn.q_proj.weight": "0.safetensors",
+    "encoder.layers.2.self_attn.q_proj.bias": "0.safetensors",
+    "encoder.layers.2.self_attn.k_proj.weight": "0.safetensors",
+    "encoder.layers.2.self_attn.k_proj.bias": "0.safetensors",
+    "encoder.layers.2.self_attn.v_proj.weight": "0.safetensors",
+    "encoder.layers.2.self_attn.v_proj.bias": "0.safetensors",
+    "encoder.layers.2.self_attn.o_proj.weight": "0.safetensors",
+    "encoder.layers.2.self_attn.rotary_emb.inv_freq": "0.safetensors",
+    "encoder.layers.2.post_attention_layernorm.weight": "0.safetensors",
+    "encoder.layers.2.mlp.gate_proj.weight": "1.safetensors",
+    "encoder.layers.2.mlp.up_proj.weight": "1.safetensors",
+    "encoder.layers.2.mlp.down_proj.weight": "1.safetensors",
+    "encoder.layers.3.input_layernorm.weight": "1.safetensors",
+    "encoder.layers.3.self_attn.q_proj.weight": "1.safetensors",
+    "encoder.layers.3.self_attn.q_proj.bias": "1.safetensors",
+    "encoder.layers.3.self_attn.k_proj.weight": "1.safetensors",
+    "encoder.layers.3.self_attn.k_proj.bias": "1.safetensors",
+    "encoder.layers.3.self_attn.v_proj.weight": "1.safetensors",
+    "encoder.layers.3.self_attn.v_proj.bias": "1.safetensors",
+    "encoder.layers.3.self_attn.o_proj.weight": "1.safetensors",
+    "encoder.layers.3.self_attn.rotary_emb.inv_freq": "1.safetensors",
+    "encoder.layers.3.post_attention_layernorm.weight": "1.safetensors",
+    "encoder.layers.3.mlp.gate_proj.weight": "1.safetensors",
+    "encoder.layers.3.mlp.up_proj.weight": "1.safetensors",
+    "encoder.layers.3.mlp.down_proj.weight": "1.safetensors",
+    "encoder.layers.4.input_layernorm.weight": "1.safetensors",
+    "encoder.layers.4.self_attn.q_proj.weight": "1.safetensors",
+    "encoder.layers.4.self_attn.q_proj.bias": "1.safetensors",
+    "encoder.layers.4.self_attn.k_proj.weight": "1.safetensors",
+    "encoder.layers.4.self_attn.k_proj.bias": "1.safetensors",
+    "encoder.layers.4.self_attn.v_proj.weight": "1.safetensors",
+    "encoder.layers.4.self_attn.v_proj.bias": "1.safetensors",
+    "encoder.layers.4.self_attn.o_proj.weight": "1.safetensors",
+    "encoder.layers.4.self_attn.rotary_emb.inv_freq": "1.safetensors",
+    "encoder.layers.4.post_attention_layernorm.weight": "1.safetensors",
+    "encoder.layers.4.mlp.gate_proj.weight": "1.safetensors",
+    "encoder.layers.4.mlp.up_proj.weight": "1.safetensors",
+    "encoder.layers.4.mlp.down_proj.weight": "1.safetensors",
+    "encoder.layers.5.input_layernorm.weight": "1.safetensors",
+    "encoder.layers.5.self_attn.q_proj.weight": "1.safetensors",
+    "encoder.layers.5.self_attn.q_proj.bias": "1.safetensors",
+    "encoder.layers.5.self_attn.k_proj.weight": "1.safetensors",
+    "encoder.layers.5.self_attn.k_proj.bias": "1.safetensors",
+    "encoder.layers.5.self_attn.v_proj.weight": "1.safetensors",
+    "encoder.layers.5.self_attn.v_proj.bias": "1.safetensors",
+    "encoder.layers.5.self_attn.o_proj.weight": "1.safetensors",
+    "encoder.layers.5.self_attn.rotary_emb.inv_freq": "1.safetensors",
+    "encoder.layers.5.post_attention_layernorm.weight": "1.safetensors",
+    "encoder.layers.5.mlp.gate_proj.weight": "1.safetensors",
+    "encoder.layers.5.mlp.up_proj.weight": "1.safetensors",
+    "encoder.layers.5.mlp.down_proj.weight": "1.safetensors",
+    "encoder.layers.6.input_layernorm.weight": "1.safetensors",
+    "encoder.layers.6.self_attn.q_proj.weight": "1.safetensors",
+    "encoder.layers.6.self_attn.q_proj.bias": "1.safetensors",
+    "encoder.layers.6.self_attn.k_proj.weight": "1.safetensors",
+    "encoder.layers.6.self_attn.k_proj.bias": "1.safetensors",
+    "encoder.layers.6.self_attn.v_proj.weight": "1.safetensors",
+    "encoder.layers.6.self_attn.v_proj.bias": "1.safetensors",
+    "encoder.layers.6.self_attn.o_proj.weight": "1.safetensors",
+    "encoder.layers.6.self_attn.rotary_emb.inv_freq": "1.safetensors",
+    "encoder.layers.6.post_attention_layernorm.weight": "1.safetensors",
+    "encoder.layers.6.mlp.gate_proj.weight": "1.safetensors",
+    "encoder.layers.6.mlp.up_proj.weight": "1.safetensors",
+    "encoder.layers.6.mlp.down_proj.weight": "2.safetensors",
+    "encoder.layers.7.input_layernorm.weight": "2.safetensors",
+    "encoder.layers.7.self_attn.q_proj.weight": "2.safetensors",
+    "encoder.layers.7.self_attn.q_proj.bias": "2.safetensors",
+    "encoder.layers.7.self_attn.k_proj.weight": "2.safetensors",
+    "encoder.layers.7.self_attn.k_proj.bias": "2.safetensors",
+    "encoder.layers.7.self_attn.v_proj.weight": "2.safetensors",
+    "encoder.layers.7.self_attn.v_proj.bias": "2.safetensors",
+    "encoder.layers.7.self_attn.o_proj.weight": "2.safetensors",
+    "encoder.layers.7.self_attn.rotary_emb.inv_freq": "2.safetensors",
+    "encoder.layers.7.post_attention_layernorm.weight": "2.safetensors",
+    "encoder.layers.7.mlp.gate_proj.weight": "2.safetensors",
+    "encoder.layers.7.mlp.up_proj.weight": "2.safetensors",
+    "encoder.layers.7.mlp.down_proj.weight": "2.safetensors",
+    "encoder.layers.8.input_layernorm.weight": "2.safetensors",
+    "encoder.layers.8.self_attn.q_proj.weight": "2.safetensors",
+    "encoder.layers.8.self_attn.q_proj.bias": "2.safetensors",
+    "encoder.layers.8.self_attn.k_proj.weight": "2.safetensors",
+    "encoder.layers.8.self_attn.k_proj.bias": "2.safetensors",
+    "encoder.layers.8.self_attn.v_proj.weight": "2.safetensors",
+    "encoder.layers.8.self_attn.v_proj.bias": "2.safetensors",
+    "encoder.layers.8.self_attn.o_proj.weight": "2.safetensors",
+    "encoder.layers.8.self_attn.rotary_emb.inv_freq": "2.safetensors",
+    "encoder.layers.8.post_attention_layernorm.weight": "2.safetensors",
+    "encoder.layers.8.mlp.gate_proj.weight": "2.safetensors",
+    "encoder.layers.8.mlp.up_proj.weight": "2.safetensors",
+    "encoder.layers.8.mlp.down_proj.weight": "2.safetensors",
+    "encoder.layers.9.input_layernorm.weight": "2.safetensors",
+    "encoder.layers.9.self_attn.q_proj.weight": "2.safetensors",
+    "encoder.layers.9.self_attn.q_proj.bias": "2.safetensors",
+    "encoder.layers.9.self_attn.k_proj.weight": "2.safetensors",
+    "encoder.layers.9.self_attn.k_proj.bias": "2.safetensors",
+    "encoder.layers.9.self_attn.v_proj.weight": "2.safetensors",
+    "encoder.layers.9.self_attn.v_proj.bias": "2.safetensors",
+    "encoder.layers.9.self_attn.o_proj.weight": "2.safetensors",
+    "encoder.layers.9.self_attn.rotary_emb.inv_freq": "2.safetensors",
+    "encoder.layers.9.post_attention_layernorm.weight": "2.safetensors",
+    "encoder.layers.9.mlp.gate_proj.weight": "2.safetensors",
+    "encoder.layers.9.mlp.up_proj.weight": "2.safetensors",
+    "encoder.layers.9.mlp.down_proj.weight": "2.safetensors",
+    "encoder.layers.10.input_layernorm.weight": "2.safetensors",
+    "encoder.layers.10.self_attn.q_proj.weight": "2.safetensors",
+    "encoder.layers.10.self_attn.q_proj.bias": "2.safetensors",
+    "encoder.layers.10.self_attn.k_proj.weight": "2.safetensors",
+    "encoder.layers.10.self_attn.k_proj.bias": "2.safetensors",
+    "encoder.layers.10.self_attn.v_proj.weight": "2.safetensors",
+    "encoder.layers.10.self_attn.v_proj.bias": "2.safetensors",
+    "encoder.layers.10.self_attn.o_proj.weight": "2.safetensors",
+    "encoder.layers.10.self_attn.rotary_emb.inv_freq": "2.safetensors",
+    "encoder.layers.10.post_attention_layernorm.weight": "2.safetensors",
+    "encoder.layers.10.mlp.gate_proj.weight": "2.safetensors",
+    "encoder.layers.10.mlp.up_proj.weight": "2.safetensors",
+    "encoder.layers.10.mlp.down_proj.weight": "2.safetensors",
+    "encoder.layers.11.input_layernorm.weight": "2.safetensors",
+    "encoder.layers.11.self_attn.q_proj.weight": "2.safetensors",
+    "encoder.layers.11.self_attn.q_proj.bias": "2.safetensors",
+    "encoder.layers.11.self_attn.k_proj.weight": "2.safetensors",
+    "encoder.layers.11.self_attn.k_proj.bias": "2.safetensors",
+    "encoder.layers.11.self_attn.v_proj.weight": "2.safetensors",
+    "encoder.layers.11.self_attn.v_proj.bias": "2.safetensors",
+    "encoder.layers.11.self_attn.o_proj.weight": "2.safetensors",
+    "encoder.layers.11.self_attn.rotary_emb.inv_freq": "2.safetensors",
+    "encoder.layers.11.post_attention_layernorm.weight": "2.safetensors",
+    "encoder.layers.11.mlp.gate_proj.weight": "3.safetensors",
+    "encoder.layers.11.mlp.up_proj.weight": "3.safetensors",
+    "encoder.layers.11.mlp.down_proj.weight": "3.safetensors",
+    "encoder.layers.12.input_layernorm.weight": "3.safetensors",
+    "encoder.layers.12.self_attn.q_proj.weight": "3.safetensors",
+    "encoder.layers.12.self_attn.q_proj.bias": "3.safetensors",
+    "encoder.layers.12.self_attn.k_proj.weight": "3.safetensors",
+    "encoder.layers.12.self_attn.k_proj.bias": "3.safetensors",
+    "encoder.layers.12.self_attn.v_proj.weight": "3.safetensors",
+    "encoder.layers.12.self_attn.v_proj.bias": "3.safetensors",
+    "encoder.layers.12.self_attn.o_proj.weight": "3.safetensors",
+    "encoder.layers.12.self_attn.rotary_emb.inv_freq": "3.safetensors",
+    "encoder.layers.12.post_attention_layernorm.weight": "3.safetensors",
+    "encoder.layers.12.mlp.gate_proj.weight": "3.safetensors",
+    "encoder.layers.12.mlp.up_proj.weight": "3.safetensors",
+    "encoder.layers.12.mlp.down_proj.weight": "3.safetensors",
+    "encoder.layers.13.input_layernorm.weight": "3.safetensors",
+    "encoder.layers.13.self_attn.q_proj.weight": "3.safetensors",
+    "encoder.layers.13.self_attn.q_proj.bias": "3.safetensors",
+    "encoder.layers.13.self_attn.k_proj.weight": "3.safetensors",
+    "encoder.layers.13.self_attn.k_proj.bias": "3.safetensors",
+    "encoder.layers.13.self_attn.v_proj.weight": "3.safetensors",
+    "encoder.layers.13.self_attn.v_proj.bias": "3.safetensors",
+    "encoder.layers.13.self_attn.o_proj.weight": "3.safetensors",
+    "encoder.layers.13.self_attn.rotary_emb.inv_freq": "3.safetensors",
+    "encoder.layers.13.post_attention_layernorm.weight": "3.safetensors",
+    "encoder.layers.13.mlp.gate_proj.weight": "3.safetensors",
+    "encoder.layers.13.mlp.up_proj.weight": "3.safetensors",
+    "encoder.layers.13.mlp.down_proj.weight": "3.safetensors",
+    "encoder.layers.14.input_layernorm.weight": "3.safetensors",
+    "encoder.layers.14.self_attn.q_proj.weight": "3.safetensors",
+    "encoder.layers.14.self_attn.q_proj.bias": "3.safetensors",
+    "encoder.layers.14.self_attn.k_proj.weight": "3.safetensors",
+    "encoder.layers.14.self_attn.k_proj.bias": "3.safetensors",
+    "encoder.layers.14.self_attn.v_proj.weight": "3.safetensors",
+    "encoder.layers.14.self_attn.v_proj.bias": "3.safetensors",
+    "encoder.layers.14.self_attn.o_proj.weight": "3.safetensors",
+    "encoder.layers.14.self_attn.rotary_emb.inv_freq": "3.safetensors",
+    "encoder.layers.14.post_attention_layernorm.weight": "3.safetensors",
+    "encoder.layers.14.mlp.gate_proj.weight": "3.safetensors",
+    "encoder.layers.14.mlp.up_proj.weight": "3.safetensors",
+    "encoder.layers.14.mlp.down_proj.weight": "3.safetensors",
+    "encoder.layers.15.input_layernorm.weight": "3.safetensors",
+    "encoder.layers.15.self_attn.q_proj.weight": "3.safetensors",
+    "encoder.layers.15.self_attn.q_proj.bias": "3.safetensors",
+    "encoder.layers.15.self_attn.k_proj.weight": "3.safetensors",
+    "encoder.layers.15.self_attn.k_proj.bias": "3.safetensors",
+    "encoder.layers.15.self_attn.v_proj.weight": "3.safetensors",
+    "encoder.layers.15.self_attn.v_proj.bias": "3.safetensors",
+    "encoder.layers.15.self_attn.o_proj.weight": "3.safetensors",
+    "encoder.layers.15.self_attn.rotary_emb.inv_freq": "3.safetensors",
+    "encoder.layers.15.post_attention_layernorm.weight": "3.safetensors",
+    "encoder.layers.15.mlp.gate_proj.weight": "3.safetensors",
+    "encoder.layers.15.mlp.up_proj.weight": "3.safetensors",
+    "encoder.layers.15.mlp.down_proj.weight": "4.safetensors",
+    "encoder.layers.16.input_layernorm.weight": "4.safetensors",
+    "encoder.layers.16.self_attn.q_proj.weight": "4.safetensors",
+    "encoder.layers.16.self_attn.q_proj.bias": "4.safetensors",
+    "encoder.layers.16.self_attn.k_proj.weight": "4.safetensors",
+    "encoder.layers.16.self_attn.k_proj.bias": "4.safetensors",
+    "encoder.layers.16.self_attn.v_proj.weight": "4.safetensors",
+    "encoder.layers.16.self_attn.v_proj.bias": "4.safetensors",
+    "encoder.layers.16.self_attn.o_proj.weight": "4.safetensors",
+    "encoder.layers.16.self_attn.rotary_emb.inv_freq": "4.safetensors",
+    "encoder.layers.16.post_attention_layernorm.weight": "4.safetensors",
+    "encoder.layers.16.mlp.gate_proj.weight": "4.safetensors",
+    "encoder.layers.16.mlp.up_proj.weight": "4.safetensors",
+    "encoder.layers.16.mlp.down_proj.weight": "4.safetensors",
+    "encoder.layers.17.input_layernorm.weight": "4.safetensors",
+    "encoder.layers.17.self_attn.q_proj.weight": "4.safetensors",
+    "encoder.layers.17.self_attn.q_proj.bias": "4.safetensors",
+    "encoder.layers.17.self_attn.k_proj.weight": "4.safetensors",
+    "encoder.layers.17.self_attn.k_proj.bias": "4.safetensors",
+    "encoder.layers.17.self_attn.v_proj.weight": "4.safetensors",
+    "encoder.layers.17.self_attn.v_proj.bias": "4.safetensors",
+    "encoder.layers.17.self_attn.o_proj.weight": "4.safetensors",
+    "encoder.layers.17.self_attn.rotary_emb.inv_freq": "4.safetensors",
+    "encoder.layers.17.post_attention_layernorm.weight": "4.safetensors",
+    "encoder.layers.17.mlp.gate_proj.weight": "4.safetensors",
+    "encoder.layers.17.mlp.up_proj.weight": "4.safetensors",
+    "encoder.layers.17.mlp.down_proj.weight": "4.safetensors",
+    "encoder.layers.18.input_layernorm.weight": "4.safetensors",
+    "encoder.layers.18.self_attn.q_proj.weight": "4.safetensors",
+    "encoder.layers.18.self_attn.q_proj.bias": "4.safetensors",
+    "encoder.layers.18.self_attn.k_proj.weight": "4.safetensors",
+    "encoder.layers.18.self_attn.k_proj.bias": "4.safetensors",
+    "encoder.layers.18.self_attn.v_proj.weight": "4.safetensors",
+    "encoder.layers.18.self_attn.v_proj.bias": "4.safetensors",
+    "encoder.layers.18.self_attn.o_proj.weight": "4.safetensors",
+    "encoder.layers.18.self_attn.rotary_emb.inv_freq": "4.safetensors",
+    "encoder.layers.18.post_attention_layernorm.weight": "4.safetensors",
+    "encoder.layers.18.mlp.gate_proj.weight": "4.safetensors",
+    "encoder.layers.18.mlp.up_proj.weight": "4.safetensors",
+    "encoder.layers.18.mlp.down_proj.weight": "4.safetensors",
+    "encoder.layers.19.input_layernorm.weight": "4.safetensors",
+    "encoder.layers.19.self_attn.q_proj.weight": "4.safetensors",
+    "encoder.layers.19.self_attn.q_proj.bias": "4.safetensors",
+    "encoder.layers.19.self_attn.k_proj.weight": "4.safetensors",
+    "encoder.layers.19.self_attn.k_proj.bias": "4.safetensors",
+    "encoder.layers.19.self_attn.v_proj.weight": "4.safetensors",
+    "encoder.layers.19.self_attn.v_proj.bias": "4.safetensors",
+    "encoder.layers.19.self_attn.o_proj.weight": "4.safetensors",
+    "encoder.layers.19.self_attn.rotary_emb.inv_freq": "4.safetensors",
+    "encoder.layers.19.post_attention_layernorm.weight": "4.safetensors",
+    "encoder.layers.19.mlp.gate_proj.weight": "4.safetensors",
+    "encoder.layers.19.mlp.up_proj.weight": "4.safetensors",
+    "encoder.layers.19.mlp.down_proj.weight": "4.safetensors",
+    "encoder.layers.20.input_layernorm.weight": "4.safetensors",
+    "encoder.layers.20.self_attn.q_proj.weight": "4.safetensors",
+    "encoder.layers.20.self_attn.q_proj.bias": "4.safetensors",
+    "encoder.layers.20.self_attn.k_proj.weight": "4.safetensors",
+    "encoder.layers.20.self_attn.k_proj.bias": "4.safetensors",
+    "encoder.layers.20.self_attn.v_proj.weight": "4.safetensors",
+    "encoder.layers.20.self_attn.v_proj.bias": "4.safetensors",
+    "encoder.layers.20.self_attn.o_proj.weight": "4.safetensors",
+    "encoder.layers.20.self_attn.rotary_emb.inv_freq": "4.safetensors",
+    "encoder.layers.20.post_attention_layernorm.weight": "4.safetensors",
+    "encoder.layers.20.mlp.gate_proj.weight": "5.safetensors",
+    "encoder.layers.20.mlp.up_proj.weight": "5.safetensors",
+    "encoder.layers.20.mlp.down_proj.weight": "5.safetensors",
+    "encoder.layers.21.input_layernorm.weight": "5.safetensors",
+    "encoder.layers.21.self_attn.q_proj.weight": "5.safetensors",
+    "encoder.layers.21.self_attn.q_proj.bias": "5.safetensors",
+    "encoder.layers.21.self_attn.k_proj.weight": "5.safetensors",
+    "encoder.layers.21.self_attn.k_proj.bias": "5.safetensors",
+    "encoder.layers.21.self_attn.v_proj.weight": "5.safetensors",
+    "encoder.layers.21.self_attn.v_proj.bias": "5.safetensors",
+    "encoder.layers.21.self_attn.o_proj.weight": "5.safetensors",
+    "encoder.layers.21.self_attn.rotary_emb.inv_freq": "5.safetensors",
+    "encoder.layers.21.post_attention_layernorm.weight": "5.safetensors",
+    "encoder.layers.21.mlp.gate_proj.weight": "5.safetensors",
+    "encoder.layers.21.mlp.up_proj.weight": "5.safetensors",
+    "encoder.layers.21.mlp.down_proj.weight": "5.safetensors",
+    "encoder.layers.22.input_layernorm.weight": "5.safetensors",
+    "encoder.layers.22.self_attn.q_proj.weight": "5.safetensors",
+    "encoder.layers.22.self_attn.q_proj.bias": "5.safetensors",
+    "encoder.layers.22.self_attn.k_proj.weight": "5.safetensors",
+    "encoder.layers.22.self_attn.k_proj.bias": "5.safetensors",
+    "encoder.layers.22.self_attn.v_proj.weight": "5.safetensors",
+    "encoder.layers.22.self_attn.v_proj.bias": "5.safetensors",
+    "encoder.layers.22.self_attn.o_proj.weight": "5.safetensors",
+    "encoder.layers.22.self_attn.rotary_emb.inv_freq": "5.safetensors",
+    "encoder.layers.22.post_attention_layernorm.weight": "5.safetensors",
+    "encoder.layers.22.mlp.gate_proj.weight": "5.safetensors",
+    "encoder.layers.22.mlp.up_proj.weight": "5.safetensors",
+    "encoder.layers.22.mlp.down_proj.weight": "5.safetensors",
+    "encoder.layers.23.input_layernorm.weight": "5.safetensors",
+    "encoder.layers.23.self_attn.q_proj.weight": "5.safetensors",
+    "encoder.layers.23.self_attn.q_proj.bias": "5.safetensors",
+    "encoder.layers.23.self_attn.k_proj.weight": "5.safetensors",
+    "encoder.layers.23.self_attn.k_proj.bias": "5.safetensors",
+    "encoder.layers.23.self_attn.v_proj.weight": "5.safetensors",
+    "encoder.layers.23.self_attn.v_proj.bias": "5.safetensors",
+    "encoder.layers.23.self_attn.o_proj.weight": "5.safetensors",
+    "encoder.layers.23.self_attn.rotary_emb.inv_freq": "5.safetensors",
+    "encoder.layers.23.post_attention_layernorm.weight": "5.safetensors",
+    "encoder.layers.23.mlp.gate_proj.weight": "5.safetensors",
+    "encoder.layers.23.mlp.up_proj.weight": "5.safetensors",
+    "encoder.layers.23.mlp.down_proj.weight": "5.safetensors",
+    "encoder.layers.24.input_layernorm.weight": "5.safetensors",
+    "encoder.layers.24.self_attn.q_proj.weight": "5.safetensors",
+    "encoder.layers.24.self_attn.q_proj.bias": "5.safetensors",
+    "encoder.layers.24.self_attn.k_proj.weight": "5.safetensors",
+    "encoder.layers.24.self_attn.k_proj.bias": "5.safetensors",
+    "encoder.layers.24.self_attn.v_proj.weight": "5.safetensors",
+    "encoder.layers.24.self_attn.v_proj.bias": "5.safetensors",
+    "encoder.layers.24.self_attn.o_proj.weight": "5.safetensors",
+    "encoder.layers.24.self_attn.rotary_emb.inv_freq": "5.safetensors",
+    "encoder.layers.24.post_attention_layernorm.weight": "5.safetensors",
+    "encoder.layers.24.mlp.gate_proj.weight": "5.safetensors",
+    "encoder.layers.24.mlp.up_proj.weight": "5.safetensors",
+    "encoder.layers.24.mlp.down_proj.weight": "6.safetensors",
+    "encoder.layers.25.input_layernorm.weight": "6.safetensors",
+    "encoder.layers.25.self_attn.q_proj.weight": "6.safetensors",
+    "encoder.layers.25.self_attn.q_proj.bias": "6.safetensors",
+    "encoder.layers.25.self_attn.k_proj.weight": "6.safetensors",
+    "encoder.layers.25.self_attn.k_proj.bias": "6.safetensors",
+    "encoder.layers.25.self_attn.v_proj.weight": "6.safetensors",
+    "encoder.layers.25.self_attn.v_proj.bias": "6.safetensors",
+    "encoder.layers.25.self_attn.o_proj.weight": "6.safetensors",
+    "encoder.layers.25.self_attn.rotary_emb.inv_freq": "6.safetensors",
+    "encoder.layers.25.post_attention_layernorm.weight": "6.safetensors",
+    "encoder.layers.25.mlp.gate_proj.weight": "6.safetensors",
+    "encoder.layers.25.mlp.up_proj.weight": "6.safetensors",
+    "encoder.layers.25.mlp.down_proj.weight": "6.safetensors",
+    "encoder.layers.26.input_layernorm.weight": "6.safetensors",
+    "encoder.layers.26.self_attn.q_proj.weight": "6.safetensors",
+    "encoder.layers.26.self_attn.q_proj.bias": "6.safetensors",
+    "encoder.layers.26.self_attn.k_proj.weight": "6.safetensors",
+    "encoder.layers.26.self_attn.k_proj.bias": "6.safetensors",
+    "encoder.layers.26.self_attn.v_proj.weight": "6.safetensors",
+    "encoder.layers.26.self_attn.v_proj.bias": "6.safetensors",
+    "encoder.layers.26.self_attn.o_proj.weight": "6.safetensors",
+    "encoder.layers.26.self_attn.rotary_emb.inv_freq": "6.safetensors",
+    "encoder.layers.26.post_attention_layernorm.weight": "6.safetensors",
+    "encoder.layers.26.mlp.gate_proj.weight": "6.safetensors",
+    "encoder.layers.26.mlp.up_proj.weight": "6.safetensors",
+    "encoder.layers.26.mlp.down_proj.weight": "6.safetensors",
+    "encoder.layers.27.input_layernorm.weight": "6.safetensors",
+    "encoder.layers.27.self_attn.q_proj.weight": "6.safetensors",
+    "encoder.layers.27.self_attn.q_proj.bias": "6.safetensors",
+    "encoder.layers.27.self_attn.k_proj.weight": "6.safetensors",
+    "encoder.layers.27.self_attn.k_proj.bias": "6.safetensors",
+    "encoder.layers.27.self_attn.v_proj.weight": "6.safetensors",
+    "encoder.layers.27.self_attn.v_proj.bias": "6.safetensors",
+    "encoder.layers.27.self_attn.o_proj.weight": "6.safetensors",
+    "encoder.layers.27.self_attn.rotary_emb.inv_freq": "6.safetensors",
+    "encoder.layers.27.post_attention_layernorm.weight": "6.safetensors",
+    "encoder.layers.27.mlp.gate_proj.weight": "6.safetensors",
+    "encoder.layers.27.mlp.up_proj.weight": "6.safetensors",
+    "encoder.layers.27.mlp.down_proj.weight": "6.safetensors",
+    "encoder.norm.weight": "6.safetensors",
+    "encoder.rotary_emb.inv_freq": "6.safetensors"
+  }
+}

tokenizer/chat_template.jinja

@@ -0,0 +1,54 @@
+{%- if tools %}
+    {{- '<|im_start|>system\n' }}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- messages[0]['content'] }}
+    {%- else %}
+        {{- 'You are a helpful assistant.' }}
+    {%- endif %}
+    {{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
+    {%- for tool in tools %}
+        {{- "\n" }}
+        {{- tool | tojson }}
+    {%- endfor %}
+    {{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
+{%- else %}
+    {%- if messages[0]['role'] == 'system' %}
+        {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
+    {%- else %}
+        {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in messages %}
+    {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content %}
+            {{- '\n' + message.content }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n{"name": "' }}
+            {{- tool_call.name }}
+            {{- '", "arguments": ' }}
+            {{- tool_call.arguments | tojson }}
+            {{- '}\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
+            {{- '<|im_start|>user' }}
+        {%- endif %}
+        {{- '\n<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>' }}
+        {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

tokenizer/tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67171c886afe270bc6bad965ab305bcba5e2cf3dc82ee3f61233d99faface637
+size 11416617

tokenizer/tokenizer_config.json

@@ -0,0 +1,11 @@
+{
+  "add_prefix_space": null,
+  "additional_special_tokens": null,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "eos_token": "<|endoftext|>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<|endoftext|>",
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": "<|endoftext|>"
+}

transformer/0.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b012ab4c8ae37dfc61f7e75a9b956d687442786ac587cf4be5bee266d228b0fc
+size 2119868330

transformer/1.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:859e1dde43776bf220488091ff7c667570d916945793c0d22130a866da205247
+size 2140792341

transformer/2.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84178307823d64cd8665089c92266f3800d3ee1f3bd62414909f820a8135733c
+size 2110074865

transformer/3.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52a62a2a7d309dd6ff4bc1ff92ddcad0638ff0f756e48eb86db790c8a02b6850
+size 2146660191

transformer/4.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:787e22981d7a7df83bd4a385bc4a64fa050d07a16ffcdbc42ba27d8e4813adfc
+size 2134925685

transformer/5.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b7b17f8fcd1b59fada08a211225606a9e77e76826abfeb8b01c037960040d73
+size 2140771143

transformer/6.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c3d949a0bb01382de7337f5f9a2588c79ffe7e4a1a2d01703665386de190156
+size 2140773685

transformer/7.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a556ee2a1cdc92d7a37b7a7cd83e9897d1d04d565496fb327e1b14f5aa803f22
+size 1672888373

vae/0.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59d4d01d43bba97ad5a22650a81df631e7779b5024e95d1d5d6ef9b90ec9fa0b
+size 253596517

vae/model.safetensors.index.json

@@ -0,0 +1,200 @@
+{
+  "metadata": {
+    "quantization_level": "6",
+    "mflux_version": "0.13.3"
+  },
+  "weight_map": {
+    "decoder.conv_in.conv3d.weight": "0.safetensors",
+    "decoder.conv_in.conv3d.bias": "0.safetensors",
+    "decoder.mid_block.attentions.0.norm.weight": "0.safetensors",
+    "decoder.mid_block.attentions.0.to_qkv.weight": "0.safetensors",
+    "decoder.mid_block.attentions.0.to_qkv.bias": "0.safetensors",
+    "decoder.mid_block.attentions.0.proj.weight": "0.safetensors",
+    "decoder.mid_block.attentions.0.proj.bias": "0.safetensors",
+    "decoder.mid_block.resnets.0.norm1.weight": "0.safetensors",
+    "decoder.mid_block.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "decoder.mid_block.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "decoder.mid_block.resnets.0.norm2.weight": "0.safetensors",
+    "decoder.mid_block.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "decoder.mid_block.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "decoder.mid_block.resnets.1.norm1.weight": "0.safetensors",
+    "decoder.mid_block.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "decoder.mid_block.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "decoder.mid_block.resnets.1.norm2.weight": "0.safetensors",
+    "decoder.mid_block.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "decoder.mid_block.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block0.resnets.0.norm1.weight": "0.safetensors",
+    "decoder.up_block0.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block0.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block0.resnets.0.norm2.weight": "0.safetensors",
+    "decoder.up_block0.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block0.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block0.resnets.1.norm1.weight": "0.safetensors",
+    "decoder.up_block0.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block0.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block0.resnets.1.norm2.weight": "0.safetensors",
+    "decoder.up_block0.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block0.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block0.resnets.2.norm1.weight": "0.safetensors",
+    "decoder.up_block0.resnets.2.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block0.resnets.2.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block0.resnets.2.norm2.weight": "0.safetensors",
+    "decoder.up_block0.resnets.2.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block0.resnets.2.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block0.upsamplers.0.time_conv.conv3d.weight": "0.safetensors",
+    "decoder.up_block0.upsamplers.0.time_conv.conv3d.bias": "0.safetensors",
+    "decoder.up_block0.upsamplers.0.resample_conv.weight": "0.safetensors",
+    "decoder.up_block0.upsamplers.0.resample_conv.bias": "0.safetensors",
+    "decoder.up_block1.resnets.0.norm1.weight": "0.safetensors",
+    "decoder.up_block1.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block1.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block1.resnets.0.norm2.weight": "0.safetensors",
+    "decoder.up_block1.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block1.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block1.resnets.0.skip_conv.conv3d.weight": "0.safetensors",
+    "decoder.up_block1.resnets.0.skip_conv.conv3d.bias": "0.safetensors",
+    "decoder.up_block1.resnets.1.norm1.weight": "0.safetensors",
+    "decoder.up_block1.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block1.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block1.resnets.1.norm2.weight": "0.safetensors",
+    "decoder.up_block1.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block1.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block1.resnets.2.norm1.weight": "0.safetensors",
+    "decoder.up_block1.resnets.2.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block1.resnets.2.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block1.resnets.2.norm2.weight": "0.safetensors",
+    "decoder.up_block1.resnets.2.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block1.resnets.2.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block1.upsamplers.0.time_conv.conv3d.weight": "0.safetensors",
+    "decoder.up_block1.upsamplers.0.time_conv.conv3d.bias": "0.safetensors",
+    "decoder.up_block1.upsamplers.0.resample_conv.weight": "0.safetensors",
+    "decoder.up_block1.upsamplers.0.resample_conv.bias": "0.safetensors",
+    "decoder.up_block2.resnets.0.norm1.weight": "0.safetensors",
+    "decoder.up_block2.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block2.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block2.resnets.0.norm2.weight": "0.safetensors",
+    "decoder.up_block2.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block2.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block2.resnets.1.norm1.weight": "0.safetensors",
+    "decoder.up_block2.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block2.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block2.resnets.1.norm2.weight": "0.safetensors",
+    "decoder.up_block2.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block2.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block2.resnets.2.norm1.weight": "0.safetensors",
+    "decoder.up_block2.resnets.2.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block2.resnets.2.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block2.resnets.2.norm2.weight": "0.safetensors",
+    "decoder.up_block2.resnets.2.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block2.resnets.2.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block2.upsamplers.0.resample_conv.weight": "0.safetensors",
+    "decoder.up_block2.upsamplers.0.resample_conv.bias": "0.safetensors",
+    "decoder.up_block3.resnets.0.norm1.weight": "0.safetensors",
+    "decoder.up_block3.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block3.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block3.resnets.0.norm2.weight": "0.safetensors",
+    "decoder.up_block3.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block3.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block3.resnets.1.norm1.weight": "0.safetensors",
+    "decoder.up_block3.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block3.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block3.resnets.1.norm2.weight": "0.safetensors",
+    "decoder.up_block3.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block3.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "decoder.up_block3.resnets.2.norm1.weight": "0.safetensors",
+    "decoder.up_block3.resnets.2.conv1.conv3d.weight": "0.safetensors",
+    "decoder.up_block3.resnets.2.conv1.conv3d.bias": "0.safetensors",
+    "decoder.up_block3.resnets.2.norm2.weight": "0.safetensors",
+    "decoder.up_block3.resnets.2.conv2.conv3d.weight": "0.safetensors",
+    "decoder.up_block3.resnets.2.conv2.conv3d.bias": "0.safetensors",
+    "decoder.norm_out.weight": "0.safetensors",
+    "decoder.conv_out.conv3d.weight": "0.safetensors",
+    "decoder.conv_out.conv3d.bias": "0.safetensors",
+    "encoder.conv_in.conv3d.weight": "0.safetensors",
+    "encoder.conv_in.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.0.resnets.0.norm1.weight": "0.safetensors",
+    "encoder.down_blocks.0.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.0.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.0.resnets.0.norm2.weight": "0.safetensors",
+    "encoder.down_blocks.0.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.0.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.0.resnets.1.norm1.weight": "0.safetensors",
+    "encoder.down_blocks.0.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.0.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.0.resnets.1.norm2.weight": "0.safetensors",
+    "encoder.down_blocks.0.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.0.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.0.downsamplers.0.resample_conv.weight": "0.safetensors",
+    "encoder.down_blocks.0.downsamplers.0.resample_conv.bias": "0.safetensors",
+    "encoder.down_blocks.1.resnets.0.norm1.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.1.resnets.0.norm2.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.1.resnets.0.skip_conv.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.0.skip_conv.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.1.resnets.1.norm1.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.1.resnets.1.norm2.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.1.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.1.downsamplers.0.resample_conv.weight": "0.safetensors",
+    "encoder.down_blocks.1.downsamplers.0.resample_conv.bias": "0.safetensors",
+    "encoder.down_blocks.2.resnets.0.norm1.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.2.resnets.0.norm2.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.2.resnets.0.skip_conv.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.0.skip_conv.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.2.resnets.1.norm1.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.2.resnets.1.norm2.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.2.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.2.downsamplers.0.time_conv.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.2.downsamplers.0.time_conv.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.2.downsamplers.0.resample_conv.weight": "0.safetensors",
+    "encoder.down_blocks.2.downsamplers.0.resample_conv.bias": "0.safetensors",
+    "encoder.down_blocks.3.resnets.0.norm1.weight": "0.safetensors",
+    "encoder.down_blocks.3.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.3.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.3.resnets.0.norm2.weight": "0.safetensors",
+    "encoder.down_blocks.3.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.3.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.3.resnets.1.norm1.weight": "0.safetensors",
+    "encoder.down_blocks.3.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.3.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "encoder.down_blocks.3.resnets.1.norm2.weight": "0.safetensors",
+    "encoder.down_blocks.3.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "encoder.down_blocks.3.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "encoder.mid_block.attentions.0.norm.weight": "0.safetensors",
+    "encoder.mid_block.attentions.0.to_qkv.weight": "0.safetensors",
+    "encoder.mid_block.attentions.0.to_qkv.bias": "0.safetensors",
+    "encoder.mid_block.attentions.0.proj.weight": "0.safetensors",
+    "encoder.mid_block.attentions.0.proj.bias": "0.safetensors",
+    "encoder.mid_block.resnets.0.norm1.weight": "0.safetensors",
+    "encoder.mid_block.resnets.0.conv1.conv3d.weight": "0.safetensors",
+    "encoder.mid_block.resnets.0.conv1.conv3d.bias": "0.safetensors",
+    "encoder.mid_block.resnets.0.norm2.weight": "0.safetensors",
+    "encoder.mid_block.resnets.0.conv2.conv3d.weight": "0.safetensors",
+    "encoder.mid_block.resnets.0.conv2.conv3d.bias": "0.safetensors",
+    "encoder.mid_block.resnets.1.norm1.weight": "0.safetensors",
+    "encoder.mid_block.resnets.1.conv1.conv3d.weight": "0.safetensors",
+    "encoder.mid_block.resnets.1.conv1.conv3d.bias": "0.safetensors",
+    "encoder.mid_block.resnets.1.norm2.weight": "0.safetensors",
+    "encoder.mid_block.resnets.1.conv2.conv3d.weight": "0.safetensors",
+    "encoder.mid_block.resnets.1.conv2.conv3d.bias": "0.safetensors",
+    "encoder.norm_out.weight": "0.safetensors",
+    "encoder.conv_out.conv3d.weight": "0.safetensors",
+    "encoder.conv_out.conv3d.bias": "0.safetensors",
+    "post_quant_conv.conv3d.weight": "0.safetensors",
+    "post_quant_conv.conv3d.bias": "0.safetensors",
+    "quant_conv.conv3d.weight": "0.safetensors",
+    "quant_conv.conv3d.bias": "0.safetensors"
+  }
+}