Hugging Face's logo

QuantFunc
/
Qwen-Image-Series

Text-to-Image
Diffusers
Safetensors
English
QuantFuncPipeline
custom_qwen_image
image-generation
diffusion
quantized
quantfunc
Model card Files
xet
 Community
QuantFunc commited on
Commit
32be0f3
·
1 Parent(s): e24556b

Upload folder using huggingface_hub

Browse files
Files changed (44) hide show
  1. README.md +175 -0
  2. assets/logo.webp +0 -0
  3. precision-config/50x-above-fp4-sample.json +12 -0
  4. precision-config/50x-below-int4-sample.json +12 -0
  5. prequant/qwen-image-2512-50x-above.safetensors +3 -0
  6. prequant/qwen-image-2512-50x-below.safetensors +3 -0
  7. prequant/qwen-image-50x-above.safetensors +3 -0
  8. prequant/qwen-image-50x-below.safetensors +3 -0
  9. qwen-image-series-50x-above-base-model/model_index.json +24 -0
  10. qwen-image-series-50x-above-base-model/quantfunc_config.json +15 -0
  11. qwen-image-series-50x-above-base-model/scheduler/scheduler_config.json +11 -0
  12. qwen-image-series-50x-above-base-model/text_encoder/config.json +132 -0
  13. qwen-image-series-50x-above-base-model/text_encoder/model.safetensors +3 -0
  14. qwen-image-series-50x-above-base-model/tokenizer/added_tokens.json +24 -0
  15. qwen-image-series-50x-above-base-model/tokenizer/chat_template.jinja +54 -0
  16. qwen-image-series-50x-above-base-model/tokenizer/merges.txt +0 -0
  17. qwen-image-series-50x-above-base-model/tokenizer/special_tokens_map.json +31 -0
  18. qwen-image-series-50x-above-base-model/tokenizer/tokenizer_config.json +207 -0
  19. qwen-image-series-50x-above-base-model/tokenizer/vocab.json +0 -0
  20. qwen-image-series-50x-above-base-model/vae/config.json +56 -0
  21. qwen-image-series-50x-above-base-model/vae/diffusion_pytorch_model.safetensors +3 -0
  22. qwen-image-series-50x-below-base-model/.quantfunc_keymap_cache/37e8f96ef1c64bf46879a40ccc33d348.bin +3 -0
  23. qwen-image-series-50x-below-base-model/.quantfunc_keymap_cache/ffee6342fd24d487e726656f39a6386b.bin +3 -0
  24. qwen-image-series-50x-below-base-model/.quantfunc_vram_cache_lighting.json +0 -0
  25. qwen-image-series-50x-below-base-model/.quantfunc_vram_cache_lighting_ed38d446_lora_39c6fde71e307d0495bd97d4ca504940.json +0 -0
  26. qwen-image-series-50x-below-base-model/.quantfunc_vram_cache_svdq.json +0 -0
  27. qwen-image-series-50x-below-base-model/model_index.json +24 -0
  28. qwen-image-series-50x-below-base-model/quantfunc_config.json +5 -0
  29. qwen-image-series-50x-below-base-model/scheduler/scheduler_config.json +11 -0
  30. qwen-image-series-50x-below-base-model/text_encoder/config.json +135 -0
  31. qwen-image-series-50x-below-base-model/text_encoder/model.safetensors +3 -0
  32. qwen-image-series-50x-below-base-model/tokenizer/added_tokens.json +24 -0
  33. qwen-image-series-50x-below-base-model/tokenizer/chat_template.jinja +54 -0
  34. qwen-image-series-50x-below-base-model/tokenizer/merges.txt +0 -0
  35. qwen-image-series-50x-below-base-model/tokenizer/special_tokens_map.json +31 -0
  36. qwen-image-series-50x-below-base-model/tokenizer/tokenizer_config.json +207 -0
  37. qwen-image-series-50x-below-base-model/tokenizer/vocab.json +0 -0
  38. qwen-image-series-50x-below-base-model/vae/config.json +56 -0
  39. qwen-image-series-50x-below-base-model/vae/diffusion_pytorch_model.safetensors +3 -0
  40. transformer/config.json +18 -0
  41. transformer/qwen-image-2512-50x-above-lighting-4steps-prequant.safetensors +3 -0
  42. transformer/qwen-image-2512-50x-above-lighting-4steps.safetensors +3 -0
  43. transformer/qwen-image-2512-50x-below-lighting-4steps-prequant.safetensors +3 -0
  44. transformer/qwen-image-2512-50x-below-lighting-4steps.safetensors +3 -0
README.md ADDED
@@ -0,0 +1,175 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ frameworks:
3
+ - pytorch
4
+ language:
5
+ - en
6
+ license: other
7
+ license_name: quantfunc-model-license
8
+ tags:
9
+ - image-generation
10
+ - text-to-image
11
+ - diffusion
12
+ - quantized
13
+ - quantfunc
14
+ ---
15
+
16
+ # QuantFunc
17
+
18
+ <div align="center" style="margin-top: 50px;">
19
+ <img src="assets/logo.webp" width="300" alt="Logo">
20
+ </div>
21
+
22
+ # Qwen-Image-Series
23
+
24
+ Pre-quantized **Qwen-Image-2512** text-to-image model series by [QuantFunc](https://github.com/user/quantfunc), with Lighting backend inference support.
25
+
26
+ ## Overview
27
+
28
+ Qwen-Image-2512 is a text-to-image diffusion model distilled from Alibaba Qwen team's image generation model.
29
+
30
+ With the latest QuantFunc ComfyUI plugin, inference achieves **2x–6x speedup** over mainstream frameworks.
31
+
32
+ ## Hardware Requirements
33
+
34
+ - Supports NVIDIA RTX 30 series and above
35
+ - RTX 20 series does not support BF16, which causes significant precision loss in Qwen series model quantization scenarios. Therefore, the 20 series currently only supports Z-Image models.
36
+
37
+ ## Compatibility
38
+
39
+ - The base models in this repository are compatible with **any version** of Qwen-Image transformer weights
40
+ - The QuantFunc code plugin and ComfyUI plugin are **100% compatible** with previous versions of Qwen-Image models
41
+
42
+ ## Directory Structure
43
+
44
+ ```
45
+ Qwen-Image-Series/
46
+ ├── qwen-image-series-50x-above-base-model/ # Base model, optimized for RTX 50 series and above
47
+ │ ├── text_encoder/ # Qwen2.5-VL text encoder (pre-quantized)
48
+ │ ├── vae/ # 3D VAE decoder (~242MB)
49
+ │ ├── tokenizer/ # Tokenizer
50
+ │ ├── scheduler/ # Scheduler config
51
+ │ ├── model_index.json
52
+ │ └── quantfunc_config.json
53
+ ├── qwen-image-series-50x-below-base-model/ # Base model, optimized for RTX 50 series and below
54
+ │ └── (same structure as above)
55
+ ├── transformer/
56
+ │ ├── config.json
57
+ │ ├── qwen-image-2512-50x-above-lighting-4steps.safetensors # RTX 50+ Lighting 4-step (~14GB)
58
+ │ ├── qwen-image-2512-50x-above-lighting-4steps-prequant.safetensors # RTX 50+ Lighting pre-quantized (~11GB)
59
+ │ ├── qwen-image-2512-50x-below-lighting-4steps.safetensors # RTX 30/40 Lighting 4-step (~14GB)
60
+ │ └── qwen-image-2512-50x-below-lighting-4steps-prequant.safetensors # RTX 30/40 Lighting pre-quantized (~11GB)
61
+ ├── prequant/ # Pre-quantized modulation weights
62
+ │ ├── qwen-image-2512-50x-above.safetensors # RTX 50+ mod weights (2512)
63
+ │ ├── qwen-image-2512-50x-below.safetensors # RTX 30/40 mod weights (2512)
64
+ │ ├── qwen-image-50x-above.safetensors # RTX 50+ mod weights (legacy)
65
+ │ └── qwen-image-50x-below.safetensors # RTX 30/40 mod weights (legacy)
66
+ └── precision-config/ # Lighting precision config samples
67
+ ├── 50x-above-fp4-sample.json # FP4 config for RTX 50+
68
+ └── 50x-below-int4-sample.json # INT4 config for RTX 30/40
69
+ ```
70
+
71
+ ## Model Variants
72
+
73
+ | Variant | base-model | transformer | Total Size | Target GPU |
74
+ |---------|-----------|-------------|------------|------------|
75
+ | **50x-above** | `qwen-image-series-50x-above-base-model` | `qwen-image-2512-50x-above-lighting-4steps.safetensors` | ~14GB | RTX 50 series and above |
76
+ | **50x-below** | `qwen-image-series-50x-below-base-model` | `qwen-image-2512-50x-below-lighting-4steps.safetensors` | ~14GB | RTX 30/40 series |
77
+
78
+ - **50x-above**: Optimized for RTX 50 series (Blackwell) and above
79
+ - **50x-below**: Optimized for RTX 30/40 series
80
+ - **4steps**: Distilled accelerated version — only 4 steps needed to generate images
81
+
82
+ > The base-model and transformer must use the **same variant** (both above or both below).
83
+
84
+ ## Quick Start
85
+
86
+ ### Download
87
+
88
+ ```bash
89
+ pip install huggingface_hub
90
+ ```
91
+
92
+ ```python
93
+ from huggingface_hub import snapshot_download
94
+ model_dir = snapshot_download('QuantFunc/Qwen-Image-Series')
95
+ ```
96
+
97
+ ### Inference
98
+
99
+ ```bash
100
+ # RTX 50 series
101
+ quantfunc \
102
+ --model-dir Qwen-Image-Series/qwen-image-series-50x-above-base-model \
103
+ --transformer Qwen-Image-Series/transformer/qwen-image-2512-50x-above-lighting-4steps.safetensors \
104
+ --auto-optimize --model-backend lighting \
105
+ --prompt "a beautiful sunset over the ocean with dramatic clouds" \
106
+ --output output.png --steps 4
107
+
108
+ # RTX 30/40 series
109
+ quantfunc \
110
+ --model-dir Qwen-Image-Series/qwen-image-series-50x-below-base-model \
111
+ --transformer Qwen-Image-Series/transformer/qwen-image-2512-50x-below-lighting-4steps.safetensors \
112
+ --auto-optimize --model-backend lighting \
113
+ --prompt "a beautiful sunset over the ocean with dramatic clouds" \
114
+ --output output.png --steps 4
115
+ ```
116
+
117
+ `--auto-optimize` automatically configures VRAM management, attention backend, and offload strategies based on your GPU.
118
+
119
+ ## SVDQ && Lighting Backend
120
+
121
+ This repository provides **Lighting** backend models. Differences between the two backends:
122
+
123
+ | Feature | Lighting | SVDQ |
124
+ |---------|----------|------|
125
+ | **Quantization** | Per-layer mixed precision (FP4/INT4/FP8/INT8) | Nunchaku-based holistic pre-quantization |
126
+ | **LoRA Integration** | Real-time quantization — build a custom model in 5 minutes with zero speed loss, integrating any number of LoRAs | Runtime low-rank pathway |
127
+ | **Ecosystem** | QuantFunc native | Compatible with the widely-adopted Nunchaku ecosystem, enhanced with Rotation quantization and Auto Rank dynamic rank optimization |
128
+ | **Flexibility** | Per-layer/sub-layer precision control | Precision fixed at export time |
129
+ | **Use Cases** | Rapid personal model customization, batch LoRA integration | Leverage Nunchaku ecosystem, runtime dynamic LoRA |
130
+
131
+ ## Pre-quantized Modulation Weights (prequant/)
132
+
133
+ The `prequant/` directory contains **pre-quantized modulation weights** extracted from SVDQ models. Use them with the Lighting backend for high-quality modulation without runtime quantization overhead.
134
+
135
+ ```bash
136
+ # From FP16 with mod weights (first run quantizes on-the-fly)
137
+ quantfunc \
138
+ --model-dir Qwen-Image-Series/qwen-image-series-50x-above-base-model \
139
+ --model-backend lighting \
140
+ --precision-config Qwen-Image-Series/precision-config/50x-above-fp4-sample.json \
141
+ --mod-weights Qwen-Image-Series/prequant/qwen-image-2512-50x-above.safetensors \
142
+ --rotation-block-size 256 \
143
+ --prompt "a beautiful sunset" --steps 4 --auto-optimize
144
+ ```
145
+
146
+ Alternatively, use the **pre-quantized Lighting transformer** for instant loading (no runtime quantization):
147
+
148
+ ```bash
149
+ quantfunc \
150
+ --model-dir Qwen-Image-Series/qwen-image-series-50x-above-base-model \
151
+ --transformer Qwen-Image-Series/transformer/qwen-image-2512-50x-above-lighting-4steps-prequant.safetensors \
152
+ --model-backend lighting \
153
+ --prompt "a beautiful sunset" --steps 4 --auto-optimize
154
+ ```
155
+
156
+ ## Precision Config (precision-config/)
157
+
158
+ Sample per-layer precision configurations for the Lighting backend:
159
+
160
+ | File | Target GPU | Precision |
161
+ |------|-----------|-----------|
162
+ | `50x-above-fp4-sample.json` | RTX 50+ | FP4 attention + AF8WF4 MLP fc2 + INT8 modulation |
163
+ | `50x-below-int4-sample.json` | RTX 30/40 | INT4 all layers + INT8 modulation |
164
+
165
+ ## Related Repositories
166
+
167
+ - [QuantFunc/Z-Image-Series](https://huggingface.co/QuantFunc/Z-Image-Series) — Z-Image-Turbo text-to-image (lightweight, fast)
168
+ - [QuantFunc/Qwen-Image-Edit-Series](https://huggingface.co/QuantFunc/Qwen-Image-Edit-Series) — Qwen-Image-Edit image editing
169
+
170
+ ## License
171
+
172
+ The pre-quantized model weights in this repository are derived from the original models. Users must comply with the original model's license agreement. The QuantFunc inference engine and its plugins (including the ComfyUI plugin) are licensed separately — see official QuantFunc channels for details.
173
+
174
+ For models quantized from commercially licensed models, users are responsible for obtaining the necessary commercial licenses from the original model providers.
175
+
assets/logo.webp ADDED
precision-config/50x-above-fp4-sample.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "transformer_blocks.attn.to_qkv": "f4",
3
+ "transformer_blocks.attn.add_qkv_proj": "f4",
4
+ "transformer_blocks.attn.to_out": "f4",
5
+ "transformer_blocks.attn.to_add_out": "f4",
6
+ "transformer_blocks.img_mlp.net.0.proj": "f4",
7
+ "transformer_blocks.img_mlp.net.2": "af8wf4",
8
+ "transformer_blocks.txt_mlp.net.0.proj": "f4",
9
+ "transformer_blocks.txt_mlp.net.2": "af8wf4",
10
+ "transformer_blocks.img_mod": "i8",
11
+ "transformer_blocks.txt_mod": "i8"
12
+ }
precision-config/50x-below-int4-sample.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "transformer_blocks.attn.to_qkv": "i4",
3
+ "transformer_blocks.attn.add_qkv_proj": "i4",
4
+ "transformer_blocks.attn.to_out": "i4",
5
+ "transformer_blocks.attn.to_add_out": "i4",
6
+ "transformer_blocks.img_mlp.net.0.proj": "i4",
7
+ "transformer_blocks.img_mlp.net.2": "i4",
8
+ "transformer_blocks.txt_mlp.net.0.proj": "i4",
9
+ "transformer_blocks.txt_mlp.net.2": "i4",
10
+ "transformer_blocks.img_mod": "i8",
11
+ "transformer_blocks.txt_mod": "i8"
12
+ }
prequant/qwen-image-2512-50x-above.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8280a6fe7c2d8a78a2dc8198175589207a1c488c9bafa0d18d2174e17ff93398
3
+ size 3826539208
prequant/qwen-image-2512-50x-below.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10c894c79b780cc0787f286f26c5f62d94e41f55b6a471d5698bd136a59c8396
3
+ size 3826539160
prequant/qwen-image-50x-above.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a70ff77a22c4ebf86d89c9400ff89746a16312cba630510a77c30e4b34e9fde0
3
+ size 3826539224
prequant/qwen-image-50x-below.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c3c9ff574d3c8754a6501ac0a6de06b190de73a91d3b138e71f9ac467ce7f8e
3
+ size 3826539176
qwen-image-series-50x-above-base-model/model_index.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "QwenImagePipeline",
3
+ "_diffusers_version": "0.36.0.dev0",
4
+ "scheduler": [
5
+ "diffusers",
6
+ "FlowMatchEulerDiscreteScheduler"
7
+ ],
8
+ "text_encoder": [
9
+ "transformers",
10
+ "Qwen2_5_VLForConditionalGeneration"
11
+ ],
12
+ "tokenizer": [
13
+ "transformers",
14
+ "Qwen2Tokenizer"
15
+ ],
16
+ "transformer": [
17
+ "diffusers",
18
+ "QwenImageTransformer2DModel"
19
+ ],
20
+ "vae": [
21
+ "diffusers",
22
+ "AutoencoderKLQwenImage"
23
+ ]
24
+ }
qwen-image-series-50x-above-base-model/quantfunc_config.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "lighting",
3
+ "model_id": "70bf7036-ffaf-48fb-9c89-c1be346f07c4",
4
+ "obfuscated": true,
5
+ "text_encoder": {
6
+ "prequantized": true,
7
+ "text_precision": "fp4",
8
+ "use_rotation": true
9
+ },
10
+ "vision_encoder": {
11
+ "prequantized": true,
12
+ "vision_quant": "fp4",
13
+ "vision_rotation": true
14
+ }
15
+ }
qwen-image-series-50x-above-base-model/scheduler/scheduler_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "FlowMatchEulerDiscreteScheduler",
3
+ "base_image_seq_len": 256,
4
+ "base_shift": 1.0986122886681098,
5
+ "max_image_seq_len": 8192,
6
+ "max_shift": 1.0986122886681098,
7
+ "num_train_timesteps": 1000,
8
+ "shift": 1.0,
9
+ "time_shift_type": "exponential",
10
+ "use_dynamic_shifting": true
11
+ }
qwen-image-series-50x-above-base-model/text_encoder/config.json ADDED
@@ -0,0 +1,132 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2_5_VLForConditionalGeneration"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "dtype": "bfloat16",
8
+ "eos_token_id": 151645,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 3584,
11
+ "image_token_id": 151655,
12
+ "initializer_range": 0.02,
13
+ "intermediate_size": 18944,
14
+ "max_position_embeddings": 128000,
15
+ "max_window_layers": 28,
16
+ "model_type": "qwen2_5_vl",
17
+ "num_attention_heads": 28,
18
+ "num_hidden_layers": 28,
19
+ "num_key_value_heads": 4,
20
+ "rms_norm_eps": 1e-06,
21
+ "rope_scaling": {
22
+ "mrope_section": [
23
+ 16,
24
+ 24,
25
+ 24
26
+ ],
27
+ "rope_type": "default",
28
+ "type": "default"
29
+ },
30
+ "rope_theta": 1000000.0,
31
+ "sliding_window": 32768,
32
+ "text_config": {
33
+ "_name_or_path": "/cpfs01/haoyangzhang/pretrained_weights/Qwen2.5-VL",
34
+ "architectures": [
35
+ "Qwen2_5_VLForConditionalGeneration"
36
+ ],
37
+ "attention_dropout": 0.0,
38
+ "bos_token_id": 151643,
39
+ "dtype": "float32",
40
+ "eos_token_id": 151645,
41
+ "hidden_act": "silu",
42
+ "hidden_size": 3584,
43
+ "initializer_range": 0.02,
44
+ "intermediate_size": 18944,
45
+ "layer_types": [
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention",
51
+ "full_attention",
52
+ "full_attention",
53
+ "full_attention",
54
+ "full_attention",
55
+ "full_attention",
56
+ "full_attention",
57
+ "full_attention",
58
+ "full_attention",
59
+ "full_attention",
60
+ "full_attention",
61
+ "full_attention",
62
+ "full_attention",
63
+ "full_attention",
64
+ "full_attention",
65
+ "full_attention",
66
+ "full_attention",
67
+ "full_attention",
68
+ "full_attention",
69
+ "full_attention",
70
+ "full_attention",
71
+ "full_attention",
72
+ "full_attention",
73
+ "full_attention"
74
+ ],
75
+ "max_position_embeddings": 128000,
76
+ "max_window_layers": 28,
77
+ "model_type": "qwen2_5_vl_text",
78
+ "num_attention_heads": 28,
79
+ "num_hidden_layers": 28,
80
+ "num_key_value_heads": 4,
81
+ "rms_norm_eps": 1e-06,
82
+ "rope_scaling": {
83
+ "mrope_section": [
84
+ 16,
85
+ 24,
86
+ 24
87
+ ],
88
+ "rope_type": "default",
89
+ "type": "default"
90
+ },
91
+ "rope_theta": 1000000.0,
92
+ "sliding_window": null,
93
+ "use_cache": true,
94
+ "use_sliding_window": false,
95
+ "vision_token_id": 151654,
96
+ "vocab_size": 152064
97
+ },
98
+ "tie_word_embeddings": false,
99
+ "transformers_version": "4.57.1",
100
+ "use_cache": true,
101
+ "use_sliding_window": false,
102
+ "video_token_id": 151656,
103
+ "vision_config": {
104
+ "depth": 32,
105
+ "dtype": "float32",
106
+ "fullatt_block_indexes": [
107
+ 7,
108
+ 15,
109
+ 23,
110
+ 31
111
+ ],
112
+ "hidden_act": "silu",
113
+ "hidden_size": 1280,
114
+ "in_channels": 3,
115
+ "in_chans": 3,
116
+ "initializer_range": 0.02,
117
+ "intermediate_size": 3420,
118
+ "model_type": "qwen2_5_vl",
119
+ "num_heads": 16,
120
+ "out_hidden_size": 3584,
121
+ "patch_size": 14,
122
+ "spatial_merge_size": 2,
123
+ "spatial_patch_size": 14,
124
+ "temporal_patch_size": 2,
125
+ "tokens_per_second": 2,
126
+ "window_size": 112
127
+ },
128
+ "vision_end_token_id": 151653,
129
+ "vision_start_token_id": 151652,
130
+ "vision_token_id": 151654,
131
+ "vocab_size": 152064
132
+ }
qwen-image-series-50x-above-base-model/text_encoder/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:48a4e5e25c2cbfa88a5e69263f70c252079c14aea320223fbd68045560ec12d7
3
+ size 4761171924
qwen-image-series-50x-above-base-model/tokenizer/added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|box_end|>": 151649,
5
+ "<|box_start|>": 151648,
6
+ "<|endoftext|>": 151643,
7
+ "<|file_sep|>": 151664,
8
+ "<|fim_middle|>": 151660,
9
+ "<|fim_pad|>": 151662,
10
+ "<|fim_prefix|>": 151659,
11
+ "<|fim_suffix|>": 151661,
12
+ "<|im_end|>": 151645,
13
+ "<|im_start|>": 151644,
14
+ "<|image_pad|>": 151655,
15
+ "<|object_ref_end|>": 151647,
16
+ "<|object_ref_start|>": 151646,
17
+ "<|quad_end|>": 151651,
18
+ "<|quad_start|>": 151650,
19
+ "<|repo_name|>": 151663,
20
+ "<|video_pad|>": 151656,
21
+ "<|vision_end|>": 151653,
22
+ "<|vision_pad|>": 151654,
23
+ "<|vision_start|>": 151652
24
+ }
qwen-image-series-50x-above-base-model/tokenizer/chat_template.jinja ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0]['role'] == 'system' %}
4
+ {{- messages[0]['content'] }}
5
+ {%- else %}
6
+ {{- 'You are a helpful assistant.' }}
7
+ {%- endif %}
8
+ {{- "\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>" }}
9
+ {%- for tool in tools %}
10
+ {{- "\n" }}
11
+ {{- tool | tojson }}
12
+ {%- endfor %}
13
+ {{- "\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" }}
14
+ {%- else %}
15
+ {%- if messages[0]['role'] == 'system' %}
16
+ {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
17
+ {%- else %}
18
+ {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
19
+ {%- endif %}
20
+ {%- endif %}
21
+ {%- for message in messages %}
22
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
23
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
24
+ {%- elif message.role == "assistant" %}
25
+ {{- '<|im_start|>' + message.role }}
26
+ {%- if message.content %}
27
+ {{- '\n' + message.content }}
28
+ {%- endif %}
29
+ {%- for tool_call in message.tool_calls %}
30
+ {%- if tool_call.function is defined %}
31
+ {%- set tool_call = tool_call.function %}
32
+ {%- endif %}
33
+ {{- '\n<tool_call>\n{"name": "' }}
34
+ {{- tool_call.name }}
35
+ {{- '", "arguments": ' }}
36
+ {{- tool_call.arguments | tojson }}
37
+ {{- '}\n</tool_call>' }}
38
+ {%- endfor %}
39
+ {{- '<|im_end|>\n' }}
40
+ {%- elif message.role == "tool" %}
41
+ {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
42
+ {{- '<|im_start|>user' }}
43
+ {%- endif %}
44
+ {{- '\n<tool_response>\n' }}
45
+ {{- message.content }}
46
+ {{- '\n</tool_response>' }}
47
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
48
+ {{- '<|im_end|>\n' }}
49
+ {%- endif %}
50
+ {%- endif %}
51
+ {%- endfor %}
52
+ {%- if add_generation_prompt %}
53
+ {{- '<|im_start|>assistant\n' }}
54
+ {%- endif %}
qwen-image-series-50x-above-base-model/tokenizer/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
qwen-image-series-50x-above-base-model/tokenizer/special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
qwen-image-series-50x-above-base-model/tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ }
181
+ },
182
+ "additional_special_tokens": [
183
+ "<|im_start|>",
184
+ "<|im_end|>",
185
+ "<|object_ref_start|>",
186
+ "<|object_ref_end|>",
187
+ "<|box_start|>",
188
+ "<|box_end|>",
189
+ "<|quad_start|>",
190
+ "<|quad_end|>",
191
+ "<|vision_start|>",
192
+ "<|vision_end|>",
193
+ "<|vision_pad|>",
194
+ "<|image_pad|>",
195
+ "<|video_pad|>"
196
+ ],
197
+ "bos_token": null,
198
+ "clean_up_tokenization_spaces": false,
199
+ "eos_token": "<|im_end|>",
200
+ "errors": "replace",
201
+ "extra_special_tokens": {},
202
+ "model_max_length": 131072,
203
+ "pad_token": "<|endoftext|>",
204
+ "split_special_tokens": false,
205
+ "tokenizer_class": "Qwen2Tokenizer",
206
+ "unk_token": null
207
+ }
qwen-image-series-50x-above-base-model/tokenizer/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
qwen-image-series-50x-above-base-model/vae/config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "AutoencoderKLQwenImage",
3
+ "_diffusers_version": "0.36.0.dev0",
4
+ "attn_scales": [],
5
+ "base_dim": 96,
6
+ "dim_mult": [
7
+ 1,
8
+ 2,
9
+ 4,
10
+ 4
11
+ ],
12
+ "dropout": 0.0,
13
+ "latents_mean": [
14
+ -0.7571,
15
+ -0.7089,
16
+ -0.9113,
17
+ 0.1075,
18
+ -0.1745,
19
+ 0.9653,
20
+ -0.1517,
21
+ 1.5508,
22
+ 0.4134,
23
+ -0.0715,
24
+ 0.5517,
25
+ -0.3632,
26
+ -0.1922,
27
+ -0.9497,
28
+ 0.2503,
29
+ -0.2921
30
+ ],
31
+ "latents_std": [
32
+ 2.8184,
33
+ 1.4541,
34
+ 2.3275,
35
+ 2.6558,
36
+ 1.2196,
37
+ 1.7708,
38
+ 2.6052,
39
+ 2.0743,
40
+ 3.2687,
41
+ 2.1526,
42
+ 2.8652,
43
+ 1.5579,
44
+ 1.6382,
45
+ 1.1253,
46
+ 2.8251,
47
+ 1.916
48
+ ],
49
+ "num_res_blocks": 2,
50
+ "temperal_downsample": [
51
+ false,
52
+ true,
53
+ true
54
+ ],
55
+ "z_dim": 16
56
+ }
qwen-image-series-50x-above-base-model/vae/diffusion_pytorch_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c8bc8b758c649abef9ea407b95408389a3b2f610d0d10fcb054fe171d0a8344
3
+ size 253806966
qwen-image-series-50x-below-base-model/.quantfunc_keymap_cache/37e8f96ef1c64bf46879a40ccc33d348.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:02df71a347fc4010098264ad082d7d96ccacefe46472b726ba602623dba62955
3
+ size 34344
qwen-image-series-50x-below-base-model/.quantfunc_keymap_cache/ffee6342fd24d487e726656f39a6386b.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7407700e82bdcc89d643e13817b3467be0df660cfa5348e30837a9d366391159
3
+ size 112900
qwen-image-series-50x-below-base-model/.quantfunc_vram_cache_lighting.json ADDED
The diff for this file is too large to render. See raw diff
 
qwen-image-series-50x-below-base-model/.quantfunc_vram_cache_lighting_ed38d446_lora_39c6fde71e307d0495bd97d4ca504940.json ADDED
The diff for this file is too large to render. See raw diff
 
qwen-image-series-50x-below-base-model/.quantfunc_vram_cache_svdq.json ADDED
The diff for this file is too large to render. See raw diff
 
qwen-image-series-50x-below-base-model/model_index.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "QwenImagePipeline",
3
+ "_diffusers_version": "0.34.0.dev0",
4
+ "scheduler": [
5
+ "diffusers",
6
+ "FlowMatchEulerDiscreteScheduler"
7
+ ],
8
+ "text_encoder": [
9
+ "transformers",
10
+ "Qwen2_5_VLForConditionalGeneration"
11
+ ],
12
+ "tokenizer": [
13
+ "transformers",
14
+ "Qwen2Tokenizer"
15
+ ],
16
+ "transformer": [
17
+ "diffusers",
18
+ "QwenImageTransformer2DModel"
19
+ ],
20
+ "vae": [
21
+ "diffusers",
22
+ "AutoencoderKLQwenImage"
23
+ ]
24
+ }
qwen-image-series-50x-below-base-model/quantfunc_config.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "backend": "lighting",
3
+ "model_id": "df4b338b-c4ff-483e-9f68-6f6f4c46201e",
4
+ "obfuscated": true
5
+ }
qwen-image-series-50x-below-base-model/scheduler/scheduler_config.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "FlowMatchEulerDiscreteScheduler",
3
+ "base_image_seq_len": 256,
4
+ "base_shift": 1.0986122886681098,
5
+ "max_image_seq_len": 8192,
6
+ "max_shift": 1.0986122886681098,
7
+ "num_train_timesteps": 1000,
8
+ "shift": 1.0,
9
+ "time_shift_type": "exponential",
10
+ "use_dynamic_shifting": true
11
+ }
qwen-image-series-50x-below-base-model/text_encoder/config.json ADDED
@@ -0,0 +1,135 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen2_5_VLForConditionalGeneration"
4
+ ],
5
+ "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
7
+ "eos_token_id": 151645,
8
+ "hidden_act": "silu",
9
+ "hidden_size": 3584,
10
+ "image_token_id": 151655,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 18944,
13
+ "max_position_embeddings": 128000,
14
+ "max_window_layers": 28,
15
+ "model_type": "qwen2_5_vl",
16
+ "num_attention_heads": 28,
17
+ "num_hidden_layers": 28,
18
+ "num_key_value_heads": 4,
19
+ "rms_norm_eps": 1e-06,
20
+ "rope_scaling": {
21
+ "mrope_section": [
22
+ 16,
23
+ 24,
24
+ 24
25
+ ],
26
+ "rope_type": "default",
27
+ "type": "default"
28
+ },
29
+ "rope_theta": 1000000.0,
30
+ "sliding_window": 32768,
31
+ "text_config": {
32
+ "architectures": [
33
+ "Qwen2_5_VLForConditionalGeneration"
34
+ ],
35
+ "attention_dropout": 0.0,
36
+ "bos_token_id": 151643,
37
+ "eos_token_id": 151645,
38
+ "hidden_act": "silu",
39
+ "hidden_size": 3584,
40
+ "image_token_id": null,
41
+ "initializer_range": 0.02,
42
+ "intermediate_size": 18944,
43
+ "layer_types": [
44
+ "full_attention",
45
+ "full_attention",
46
+ "full_attention",
47
+ "full_attention",
48
+ "full_attention",
49
+ "full_attention",
50
+ "full_attention",
51
+ "full_attention",
52
+ "full_attention",
53
+ "full_attention",
54
+ "full_attention",
55
+ "full_attention",
56
+ "full_attention",
57
+ "full_attention",
58
+ "full_attention",
59
+ "full_attention",
60
+ "full_attention",
61
+ "full_attention",
62
+ "full_attention",
63
+ "full_attention",
64
+ "full_attention",
65
+ "full_attention",
66
+ "full_attention",
67
+ "full_attention",
68
+ "full_attention",
69
+ "full_attention",
70
+ "full_attention",
71
+ "full_attention"
72
+ ],
73
+ "max_position_embeddings": 128000,
74
+ "max_window_layers": 28,
75
+ "model_type": "qwen2_5_vl_text",
76
+ "num_attention_heads": 28,
77
+ "num_hidden_layers": 28,
78
+ "num_key_value_heads": 4,
79
+ "rms_norm_eps": 1e-06,
80
+ "rope_scaling": {
81
+ "mrope_section": [
82
+ 16,
83
+ 24,
84
+ 24
85
+ ],
86
+ "rope_type": "default",
87
+ "type": "default"
88
+ },
89
+ "rope_theta": 1000000.0,
90
+ "sliding_window": null,
91
+ "torch_dtype": "float32",
92
+ "use_cache": true,
93
+ "use_sliding_window": false,
94
+ "video_token_id": null,
95
+ "vision_end_token_id": 151653,
96
+ "vision_start_token_id": 151652,
97
+ "vision_token_id": 151654,
98
+ "vocab_size": 152064
99
+ },
100
+ "tie_word_embeddings": false,
101
+ "torch_dtype": "bfloat16",
102
+ "transformers_version": "4.53.1",
103
+ "use_cache": true,
104
+ "use_sliding_window": false,
105
+ "video_token_id": 151656,
106
+ "vision_config": {
107
+ "depth": 32,
108
+ "fullatt_block_indexes": [
109
+ 7,
110
+ 15,
111
+ 23,
112
+ 31
113
+ ],
114
+ "hidden_act": "silu",
115
+ "hidden_size": 1280,
116
+ "in_channels": 3,
117
+ "in_chans": 3,
118
+ "initializer_range": 0.02,
119
+ "intermediate_size": 3420,
120
+ "model_type": "qwen2_5_vl",
121
+ "num_heads": 16,
122
+ "out_hidden_size": 3584,
123
+ "patch_size": 14,
124
+ "spatial_merge_size": 2,
125
+ "spatial_patch_size": 14,
126
+ "temporal_patch_size": 2,
127
+ "tokens_per_second": 2,
128
+ "torch_dtype": "float32",
129
+ "window_size": 112
130
+ },
131
+ "vision_end_token_id": 151653,
132
+ "vision_start_token_id": 151652,
133
+ "vision_token_id": 151654,
134
+ "vocab_size": 152064
135
+ }
qwen-image-series-50x-below-base-model/text_encoder/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1421bae43e820c7b5ac8a55450bee6f998cb3aaf198f8978c80c6f8ed9ac4ba2
3
+ size 4557255810
qwen-image-series-50x-below-base-model/tokenizer/added_tokens.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</tool_call>": 151658,
3
+ "<tool_call>": 151657,
4
+ "<|box_end|>": 151649,
5
+ "<|box_start|>": 151648,
6
+ "<|endoftext|>": 151643,
7
+ "<|file_sep|>": 151664,
8
+ "<|fim_middle|>": 151660,
9
+ "<|fim_pad|>": 151662,
10
+ "<|fim_prefix|>": 151659,
11
+ "<|fim_suffix|>": 151661,
12
+ "<|im_end|>": 151645,
13
+ "<|im_start|>": 151644,
14
+ "<|image_pad|>": 151655,
15
+ "<|object_ref_end|>": 151647,
16
+ "<|object_ref_start|>": 151646,
17
+ "<|quad_end|>": 151651,
18
+ "<|quad_start|>": 151650,
19
+ "<|repo_name|>": 151663,
20
+ "<|video_pad|>": 151656,
21
+ "<|vision_end|>": 151653,
22
+ "<|vision_pad|>": 151654,
23
+ "<|vision_start|>": 151652
24
+ }
qwen-image-series-50x-below-base-model/tokenizer/chat_template.jinja ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {%- if tools %}
2
+ {{- '<|im_start|>system\n' }}
3
+ {%- if messages[0]['role'] == 'system' %}
4
+ {{- messages[0]['content'] }}
5
+ {%- else %}
6
+ {{- 'You are a helpful assistant.' }}
7
+ {%- endif %}
8
+ {{- "\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>" }}
9
+ {%- for tool in tools %}
10
+ {{- "\n" }}
11
+ {{- tool | tojson }}
12
+ {%- endfor %}
13
+ {{- "\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" }}
14
+ {%- else %}
15
+ {%- if messages[0]['role'] == 'system' %}
16
+ {{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
17
+ {%- else %}
18
+ {{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
19
+ {%- endif %}
20
+ {%- endif %}
21
+ {%- for message in messages %}
22
+ {%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
23
+ {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
24
+ {%- elif message.role == "assistant" %}
25
+ {{- '<|im_start|>' + message.role }}
26
+ {%- if message.content %}
27
+ {{- '\n' + message.content }}
28
+ {%- endif %}
29
+ {%- for tool_call in message.tool_calls %}
30
+ {%- if tool_call.function is defined %}
31
+ {%- set tool_call = tool_call.function %}
32
+ {%- endif %}
33
+ {{- '\n<tool_call>\n{"name": "' }}
34
+ {{- tool_call.name }}
35
+ {{- '", "arguments": ' }}
36
+ {{- tool_call.arguments | tojson }}
37
+ {{- '}\n</tool_call>' }}
38
+ {%- endfor %}
39
+ {{- '<|im_end|>\n' }}
40
+ {%- elif message.role == "tool" %}
41
+ {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
42
+ {{- '<|im_start|>user' }}
43
+ {%- endif %}
44
+ {{- '\n<tool_response>\n' }}
45
+ {{- message.content }}
46
+ {{- '\n</tool_response>' }}
47
+ {%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
48
+ {{- '<|im_end|>\n' }}
49
+ {%- endif %}
50
+ {%- endif %}
51
+ {%- endfor %}
52
+ {%- if add_generation_prompt %}
53
+ {{- '<|im_start|>assistant\n' }}
54
+ {%- endif %}
qwen-image-series-50x-below-base-model/tokenizer/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
qwen-image-series-50x-below-base-model/tokenizer/special_tokens_map.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<|im_start|>",
4
+ "<|im_end|>",
5
+ "<|object_ref_start|>",
6
+ "<|object_ref_end|>",
7
+ "<|box_start|>",
8
+ "<|box_end|>",
9
+ "<|quad_start|>",
10
+ "<|quad_end|>",
11
+ "<|vision_start|>",
12
+ "<|vision_end|>",
13
+ "<|vision_pad|>",
14
+ "<|image_pad|>",
15
+ "<|video_pad|>"
16
+ ],
17
+ "eos_token": {
18
+ "content": "<|im_end|>",
19
+ "lstrip": false,
20
+ "normalized": false,
21
+ "rstrip": false,
22
+ "single_word": false
23
+ },
24
+ "pad_token": {
25
+ "content": "<|endoftext|>",
26
+ "lstrip": false,
27
+ "normalized": false,
28
+ "rstrip": false,
29
+ "single_word": false
30
+ }
31
+ }
qwen-image-series-50x-below-base-model/tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,207 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_bos_token": false,
3
+ "add_prefix_space": false,
4
+ "added_tokens_decoder": {
5
+ "151643": {
6
+ "content": "<|endoftext|>",
7
+ "lstrip": false,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false,
11
+ "special": true
12
+ },
13
+ "151644": {
14
+ "content": "<|im_start|>",
15
+ "lstrip": false,
16
+ "normalized": false,
17
+ "rstrip": false,
18
+ "single_word": false,
19
+ "special": true
20
+ },
21
+ "151645": {
22
+ "content": "<|im_end|>",
23
+ "lstrip": false,
24
+ "normalized": false,
25
+ "rstrip": false,
26
+ "single_word": false,
27
+ "special": true
28
+ },
29
+ "151646": {
30
+ "content": "<|object_ref_start|>",
31
+ "lstrip": false,
32
+ "normalized": false,
33
+ "rstrip": false,
34
+ "single_word": false,
35
+ "special": true
36
+ },
37
+ "151647": {
38
+ "content": "<|object_ref_end|>",
39
+ "lstrip": false,
40
+ "normalized": false,
41
+ "rstrip": false,
42
+ "single_word": false,
43
+ "special": true
44
+ },
45
+ "151648": {
46
+ "content": "<|box_start|>",
47
+ "lstrip": false,
48
+ "normalized": false,
49
+ "rstrip": false,
50
+ "single_word": false,
51
+ "special": true
52
+ },
53
+ "151649": {
54
+ "content": "<|box_end|>",
55
+ "lstrip": false,
56
+ "normalized": false,
57
+ "rstrip": false,
58
+ "single_word": false,
59
+ "special": true
60
+ },
61
+ "151650": {
62
+ "content": "<|quad_start|>",
63
+ "lstrip": false,
64
+ "normalized": false,
65
+ "rstrip": false,
66
+ "single_word": false,
67
+ "special": true
68
+ },
69
+ "151651": {
70
+ "content": "<|quad_end|>",
71
+ "lstrip": false,
72
+ "normalized": false,
73
+ "rstrip": false,
74
+ "single_word": false,
75
+ "special": true
76
+ },
77
+ "151652": {
78
+ "content": "<|vision_start|>",
79
+ "lstrip": false,
80
+ "normalized": false,
81
+ "rstrip": false,
82
+ "single_word": false,
83
+ "special": true
84
+ },
85
+ "151653": {
86
+ "content": "<|vision_end|>",
87
+ "lstrip": false,
88
+ "normalized": false,
89
+ "rstrip": false,
90
+ "single_word": false,
91
+ "special": true
92
+ },
93
+ "151654": {
94
+ "content": "<|vision_pad|>",
95
+ "lstrip": false,
96
+ "normalized": false,
97
+ "rstrip": false,
98
+ "single_word": false,
99
+ "special": true
100
+ },
101
+ "151655": {
102
+ "content": "<|image_pad|>",
103
+ "lstrip": false,
104
+ "normalized": false,
105
+ "rstrip": false,
106
+ "single_word": false,
107
+ "special": true
108
+ },
109
+ "151656": {
110
+ "content": "<|video_pad|>",
111
+ "lstrip": false,
112
+ "normalized": false,
113
+ "rstrip": false,
114
+ "single_word": false,
115
+ "special": true
116
+ },
117
+ "151657": {
118
+ "content": "<tool_call>",
119
+ "lstrip": false,
120
+ "normalized": false,
121
+ "rstrip": false,
122
+ "single_word": false,
123
+ "special": false
124
+ },
125
+ "151658": {
126
+ "content": "</tool_call>",
127
+ "lstrip": false,
128
+ "normalized": false,
129
+ "rstrip": false,
130
+ "single_word": false,
131
+ "special": false
132
+ },
133
+ "151659": {
134
+ "content": "<|fim_prefix|>",
135
+ "lstrip": false,
136
+ "normalized": false,
137
+ "rstrip": false,
138
+ "single_word": false,
139
+ "special": false
140
+ },
141
+ "151660": {
142
+ "content": "<|fim_middle|>",
143
+ "lstrip": false,
144
+ "normalized": false,
145
+ "rstrip": false,
146
+ "single_word": false,
147
+ "special": false
148
+ },
149
+ "151661": {
150
+ "content": "<|fim_suffix|>",
151
+ "lstrip": false,
152
+ "normalized": false,
153
+ "rstrip": false,
154
+ "single_word": false,
155
+ "special": false
156
+ },
157
+ "151662": {
158
+ "content": "<|fim_pad|>",
159
+ "lstrip": false,
160
+ "normalized": false,
161
+ "rstrip": false,
162
+ "single_word": false,
163
+ "special": false
164
+ },
165
+ "151663": {
166
+ "content": "<|repo_name|>",
167
+ "lstrip": false,
168
+ "normalized": false,
169
+ "rstrip": false,
170
+ "single_word": false,
171
+ "special": false
172
+ },
173
+ "151664": {
174
+ "content": "<|file_sep|>",
175
+ "lstrip": false,
176
+ "normalized": false,
177
+ "rstrip": false,
178
+ "single_word": false,
179
+ "special": false
180
+ }
181
+ },
182
+ "additional_special_tokens": [
183
+ "<|im_start|>",
184
+ "<|im_end|>",
185
+ "<|object_ref_start|>",
186
+ "<|object_ref_end|>",
187
+ "<|box_start|>",
188
+ "<|box_end|>",
189
+ "<|quad_start|>",
190
+ "<|quad_end|>",
191
+ "<|vision_start|>",
192
+ "<|vision_end|>",
193
+ "<|vision_pad|>",
194
+ "<|image_pad|>",
195
+ "<|video_pad|>"
196
+ ],
197
+ "bos_token": null,
198
+ "clean_up_tokenization_spaces": false,
199
+ "eos_token": "<|im_end|>",
200
+ "errors": "replace",
201
+ "extra_special_tokens": {},
202
+ "model_max_length": 131072,
203
+ "pad_token": "<|endoftext|>",
204
+ "split_special_tokens": false,
205
+ "tokenizer_class": "Qwen2Tokenizer",
206
+ "unk_token": null
207
+ }
qwen-image-series-50x-below-base-model/tokenizer/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
qwen-image-series-50x-below-base-model/vae/config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "AutoencoderKLQwenImage",
3
+ "_diffusers_version": "0.34.0.dev0",
4
+ "attn_scales": [],
5
+ "base_dim": 96,
6
+ "dim_mult": [
7
+ 1,
8
+ 2,
9
+ 4,
10
+ 4
11
+ ],
12
+ "dropout": 0.0,
13
+ "latents_mean": [
14
+ -0.7571,
15
+ -0.7089,
16
+ -0.9113,
17
+ 0.1075,
18
+ -0.1745,
19
+ 0.9653,
20
+ -0.1517,
21
+ 1.5508,
22
+ 0.4134,
23
+ -0.0715,
24
+ 0.5517,
25
+ -0.3632,
26
+ -0.1922,
27
+ -0.9497,
28
+ 0.2503,
29
+ -0.2921
30
+ ],
31
+ "latents_std": [
32
+ 2.8184,
33
+ 1.4541,
34
+ 2.3275,
35
+ 2.6558,
36
+ 1.2196,
37
+ 1.7708,
38
+ 2.6052,
39
+ 2.0743,
40
+ 3.2687,
41
+ 2.1526,
42
+ 2.8652,
43
+ 1.5579,
44
+ 1.6382,
45
+ 1.1253,
46
+ 2.8251,
47
+ 1.916
48
+ ],
49
+ "num_res_blocks": 2,
50
+ "temperal_downsample": [
51
+ false,
52
+ true,
53
+ true
54
+ ],
55
+ "z_dim": 16
56
+ }
qwen-image-series-50x-below-base-model/vae/diffusion_pytorch_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c8bc8b758c649abef9ea407b95408389a3b2f610d0d10fcb054fe171d0a8344
3
+ size 253806966
transformer/config.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_class_name": "QwenImageTransformer2DModel",
3
+ "_diffusers_version": "0.34.0.dev0",
4
+ "attention_head_dim": 128,
5
+ "axes_dims_rope": [
6
+ 16,
7
+ 56,
8
+ 56
9
+ ],
10
+ "guidance_embeds": false,
11
+ "in_channels": 64,
12
+ "joint_attention_dim": 3584,
13
+ "num_attention_heads": 24,
14
+ "num_layers": 60,
15
+ "out_channels": 16,
16
+ "patch_size": 2,
17
+ "pooled_projection_dim": 768
18
+ }
transformer/qwen-image-2512-50x-above-lighting-4steps-prequant.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ffa41bd340432b954c67807cd04386aed4c1255a30951f5a623be19ccb8d69fa
3
+ size 11420095396
transformer/qwen-image-2512-50x-above-lighting-4steps.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d3b7354a3bf6eafd3ed193ec9e2af7504e494ebe86b54e994d235f62961c324
3
+ size 14498964653
transformer/qwen-image-2512-50x-below-lighting-4steps-prequant.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ec13d077253fe7126adfe9ea6a5c08ed8c726230f3f1bfdccd3f3e349f13b575
3
+ size 11139201766
transformer/qwen-image-2512-50x-below-lighting-4steps.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4492b33c783d169e056f8c1bd95d2ad9f673d864ec0be83e9c9ef236634fc1e
3
+ size 14218071081