Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Sign up
Appearance settings

Commit fc337d5

Browse files
[docs] Models (#12248)
* init * fix * feedback * feedback
1 parent 32798bf commit fc337d5

File tree

3 files changed

+128
-33
lines changed

3 files changed

+128
-33
lines changed

‎docs/source/en/_toctree.yml

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -24,6 +24,8 @@
2424
title: Reproducibility
2525
- local: using-diffusers/schedulers
2626
title: Load schedulers and models
27+
- local: using-diffusers/models
28+
title: Models
2729
- local: using-diffusers/scheduler_features
2830
title: Scheduler features
2931
- local: using-diffusers/other-formats

‎docs/source/en/using-diffusers/loading.md

Lines changed: 6 additions & 33 deletions
Original file line numberDiff line numberDiff line change
@@ -108,50 +108,27 @@ print(pipeline.transformer.dtype, pipeline.vae.dtype)
108108

109109
The `device_map` argument determines individual model or pipeline placement on an accelerator like a GPU. It is especially helpful when there are multiple GPUs.
110110

111-
Diffusers currently provides three options to `device_map`, `"cuda"`, `"balanced"`and `"auto"`. Refer to the table below to compare the three placement strategies.
111+
A pipeline supports two options for `device_map`, `"cuda"`and `"balanced"`. Refer to the table below to compare the placement strategies.
112112

113113
| parameter | description |
114114
|---|---|
115-
| `"cuda"` | places model or pipeline on CUDA device |
116-
| `"balanced"` | evenly distributes model or pipeline on all GPUs |
117-
| `"auto"` | distribute model from fastest device first to slowest |
115+
| `"cuda"` | places pipeline on a supported accelerator device like CUDA |
116+
| `"balanced"` | evenly distributes pipeline on all GPUs |
118117

119118
Use the `max_memory` argument in [`~DiffusionPipeline.from_pretrained`] to allocate a maximum amount of memory to use on each device. By default, Diffusers uses the maximum amount available.
120119

121-
<hfoptions id="device_map">
122-
<hfoption id="pipeline">
123-
124120
```py
125121
import torch
126122
from diffusers import DiffusionPipeline
127123

124+
max_memory = {0: "16GB", 1: "16GB"}
128125
pipeline = DiffusionPipeline.from_pretrained(
129126
"Qwen/Qwen-Image",
130127
torch_dtype=torch.bfloat16,
131128
device_map="cuda",
132129
)
133130
```
134131

135-
</hfoption>
136-
<hfoption id="individual model">
137-
138-
```py
139-
import torch
140-
from diffusers import AutoModel
141-
142-
max_memory = {0: "16GB", 1: "16GB"}
143-
transformer = AutoModel.from_pretrained(
144-
"Qwen/Qwen-Image",
145-
subfolder="transformer",
146-
torch_dtype=torch.bfloat16
147-
device_map="cuda",
148-
max_memory=max_memory
149-
)
150-
```
151-
152-
</hfoption>
153-
</hfoptions>
154-
155132
The `hf_device_map` attribute allows you to access and view the `device_map`.
156133

157134
```py
@@ -189,22 +166,18 @@ pipeline = DiffusionPipeline.from_pretrained(
189166

190167
[`DiffusionPipeline`] is flexible and accommodates loading different models or schedulers. You can experiment with different schedulers to optimize for generation speed or quality, and you can replace models with more performant ones.
191168

192-
The example below swaps the default scheduler to generate higher quality images and a more stable VAE version. Pass the `subfolder` argument in [`~HeunDiscreteScheduler.from_pretrained`] to load the scheduler to the correct subfolder.
169+
The example below uses a more stable VAE version.
193170

194171
```py
195172
import torch
196-
from diffusers import DiffusionPipeline, HeunDiscreteScheduler, AutoModel
173+
from diffusers import DiffusionPipeline, AutoModel
197174

198-
scheduler = HeunDiscreteScheduler.from_pretrained(
199-
"stabilityai/stable-diffusion-xl-base-1.0", subfolder="scheduler"
200-
)
201175
vae = AutoModel.from_pretrained(
202176
"madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16
203177
)
204178

205179
pipeline = DiffusionPipeline.from_pretrained(
206180
"stabilityai/stable-diffusion-xl-base-1.0",
207-
scheduler=scheduler,
208181
vae=vae,
209182
torch_dtype=torch.float16,
210183
device_map="cuda"
Lines changed: 120 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,120 @@
1+
<!--Copyright 2025 The HuggingFace Team. All rights reserved.
2+
3+
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
4+
the License. You may obtain a copy of the License at
5+
6+
http://www.apache.org/licenses/LICENSE-2.0
7+
8+
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
9+
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
10+
specific language governing permissions and limitations under the License.
11+
-->
12+
13+
[[open-in-colab]]
14+
15+
# Models
16+
17+
A diffusion model relies on a few individual models working together to generate an output. These models are responsible for denoising, encoding inputs, and decoding latents into the actual outputs.
18+
19+
This guide will show you how to load models.
20+
21+
## Loading a model
22+
23+
All models are loaded with the [`~ModelMixin.from_pretrained`] method, which downloads and caches the latest model version. If the latest files are available in the local cache, [`~ModelMixin.from_pretrained`] reuses files in the cache.
24+
25+
Pass the `subfolder` argument to [`~ModelMixin.from_pretrained`] to specify where to load the model weights from. Omit the `subfolder` argument if the repository doesn't have a subfolder structure or if you're loading a standalone model.
26+
27+
```py
28+
from diffusers import QwenImageTransformer2DModel
29+
30+
model = QwenImageTransformer2DModel.from_pretrained("Qwen/Qwen-Image", subfolder="transformer")
31+
```
32+
33+
## AutoModel
34+
35+
[`AutoModel`] detects the model class from a `model_index.json` file or a model's `config.json` file. It fetches the correct model class from these files and delegates the actual loading to the model class. [`AutoModel`] is useful for automatic model type detection without needing to know the exact model class beforehand.
36+
37+
```py
38+
from diffusers import AutoModel
39+
40+
model = AutoModel.from_pretrained(
41+
"Qwen/Qwen-Image", subfolder="transformer"
42+
)
43+
```
44+
45+
## Model data types
46+
47+
Use the `torch_dtype` argument in [`~ModelMixin.from_pretrained`] to load a model with a specific data type. This allows you to load a model in a lower precision to reduce memory usage.
48+
49+
```py
50+
import torch
51+
from diffusers import QwenImageTransformer2DModel
52+
53+
model = QwenImageTransformer2DModel.from_pretrained(
54+
"Qwen/Qwen-Image",
55+
subfolder="transformer",
56+
torch_dtype=torch.bfloat16
57+
)
58+
```
59+
60+
[nn.Module.to](https://docs.pytorch.org/docs/stable/generated/torch.nn.Module.html#torch.nn.Module.to) can also convert to a specific data type on the fly. However, it converts *all* weights to the requested data type unlike `torch_dtype` which respects `_keep_in_fp32_modules`. This argument preserves layers in `torch.float32` for numerical stability and best generation quality (see example [_keep_in_fp32_modules](https://github.com/huggingface/diffusers/blob/f864a9a352fa4a220d860bfdd1782e3e5af96382/src/diffusers/models/transformers/transformer_wan.py#L374))
61+
62+
```py
63+
from diffusers import QwenImageTransformer2DModel
64+
65+
model = QwenImageTransformer2DModel.from_pretrained(
66+
"Qwen/Qwen-Image", subfolder="transformer"
67+
)
68+
model = model.to(dtype=torch.float16)
69+
```
70+
71+
## Device placement
72+
73+
Use the `device_map` argument in [`~ModelMixin.from_pretrained`] to place a model on an accelerator like a GPU. It is especially helpful where there are multiple GPUs.
74+
75+
Diffusers currently provides three options to `device_map` for individual models, `"cuda"`, `"balanced"` and `"auto"`. Refer to the table below to compare the three placement strategies.
76+
77+
| parameter | description |
78+
|---|---|
79+
| `"cuda"` | places pipeline on a supported accelerator (CUDA) |
80+
| `"balanced"` | evenly distributes pipeline on all GPUs |
81+
| `"auto"` | distribute model from fastest device first to slowest |
82+
83+
Use the `max_memory` argument in [`~ModelMixin.from_pretrained`] to allocate a maximum amount of memory to use on each device. By default, Diffusers uses the maximum amount available.
84+
85+
```py
86+
import torch
87+
from diffusers import QwenImagePipeline
88+
89+
max_memory = {0: "16GB", 1: "16GB"}
90+
pipeline = QwenImagePipeline.from_pretrained(
91+
"Qwen/Qwen-Image",
92+
torch_dtype=torch.bfloat16,
93+
device_map="cuda",
94+
max_memory=max_memory
95+
)
96+
```
97+
98+
The `hf_device_map` attribute allows you to access and view the `device_map`.
99+
100+
```py
101+
print(transformer.hf_device_map)
102+
# {'': device(type='cuda')}
103+
```
104+
105+
## Saving models
106+
107+
Save a model with the [`~ModelMixin.save_pretrained`] method.
108+
109+
```py
110+
from diffusers import QwenImageTransformer2DModel
111+
112+
model = QwenImageTransformer2DModel.from_pretrained("Qwen/Qwen-Image", subfolder="transformer")
113+
model.save_pretrained("./local/model")
114+
```
115+
116+
For large models, it is helpful to use `max_shard_size` to save a model as multiple shards. A shard can be loaded faster and save memory (refer to the [parallel loading](./loading#parallel-loading) docs for more details), especially if there is more than one GPU.
117+
118+
```py
119+
model.save_pretrained("./local/model", max_shard_size="5GB")
120+
```

0 commit comments

Comments
(0)

AltStyle によって変換されたページ (->オリジナル) /