精品课程 网站建设质量,wordpress百度分享插件下载,怎么创建一个视频网站,网站建设html模板下载本地部署文生图模型 Flux 0. 引言1. 本地部署1-1. 创建虚拟环境1-2. 安装依赖模块1-3. 创建 Web UI1-4. 启动 Web UI1-5. 访问 Web UI 0. 引言
2024年8月1日#xff0c;blackforestlabs.ai发布了 FLUX.1 模型套件。
FLUX.1 文本到图像模型套件#xff0c;该套件定义了文本到… 本地部署文生图模型 Flux 0. 引言1. 本地部署1-1. 创建虚拟环境1-2. 安装依赖模块1-3. 创建 Web UI1-4. 启动 Web UI1-5. 访问 Web UI 0. 引言
2024年8月1日blackforestlabs.ai发布了 FLUX.1 模型套件。
FLUX.1 文本到图像模型套件该套件定义了文本到图像合成的图像细节、提示依从性、样式多样性和场景复杂性的新技术。
为了在可访问性和模型功能之间取得平衡FLUX.1 有三种变体FLUX.1 [pro]、FLUX.1 [dev] 和 FLUX.1 [schnell]
FLUX.1 [pro]FLUX.1 的佼佼者提供最先进的性能图像生成具有顶级的提示跟随、视觉质量、图像细节和输出多样性。在此处通过我们的 API 注册 FLUX.1 [pro] 访问权限。FLUX.1 [pro] 也可通过 Replicate 和 fal.ai 获得。FLUX.1 [dev]FLUX.1 [dev] 是一个用于非商业应用的开放权重、指导蒸馏模型。FLUX.1 [dev] 直接从 FLUX.1 [pro] 蒸馏而来获得了相似的质量和快速粘附能力同时比相同尺寸的标准模型效率更高。FLUX.1 [dev] 权重在 HuggingFace 上可用可以直接在 Replicate 或 Fal.ai 上试用。FLUX.1 [schnell]我们最快的模型是为本地开发和个人使用量身定制的。FLUX.1 [schnell] 在 Apache2.0 许可下公开可用。类似FLUX.1 [dev]权重在Hugging Face上可用推理代码可以在GitHub和HuggingFace的Diffusers中找到。
1. 本地部署
1-1. 创建虚拟环境
conda create -n flux python3.11 -y
conda activate flux1-2. 安装依赖模块
git clone https://github.com/black-forest-labs/flux; cd flux
pip install -e .[all]
pip install accelerate
pip install githttps://github.com/huggingface/diffusers.git
pip install optimum-quanto
pip install gradio1-3. 创建 Web UI
import torchimport gradio as grfrom optimum.quanto import freeze, qfloat8, quantizefrom diffusers import FlowMatchEulerDiscreteScheduler, AutoencoderKL
from diffusers.models.transformers.transformer_flux import FluxTransformer2DModel
from diffusers.pipelines.flux.pipeline_flux import FluxPipeline
from transformers import CLIPTextModel, CLIPTokenizer,T5EncoderModel, T5TokenizerFastdtype torch.bfloat16# schnell is the distilled turbo model. For the CFG distilled model, use:
# bfl_repo black-forest-labs/FLUX.1-dev
# revision refs/pr/3
#
# The undistilled model that uses CFG (pro) which can use negative prompts
# was not released.
bfl_repo black-forest-labs/FLUX.1-schnell
revision refs/pr/1
# bfl_repo black-forest-labs/FLUX.1-dev
# revision mainscheduler FlowMatchEulerDiscreteScheduler.from_pretrained(bfl_repo, subfolderscheduler, revisionrevision)
text_encoder CLIPTextModel.from_pretrained(openai/clip-vit-large-patch14, torch_dtypedtype)
tokenizer CLIPTokenizer.from_pretrained(openai/clip-vit-large-patch14, torch_dtypedtype)
text_encoder_2 T5EncoderModel.from_pretrained(bfl_repo, subfoldertext_encoder_2, torch_dtypedtype, revisionrevision)
tokenizer_2 T5TokenizerFast.from_pretrained(bfl_repo, subfoldertokenizer_2, torch_dtypedtype, revisionrevision)
vae AutoencoderKL.from_pretrained(bfl_repo, subfoldervae, torch_dtypedtype, revisionrevision)
transformer FluxTransformer2DModel.from_pretrained(bfl_repo, subfoldertransformer, torch_dtypedtype, revisionrevision)# Experimental: Try this to load in 4-bit for 16GB cards.
#
# from optimum.quanto import qint4
# quantize(transformer, weightsqint4, exclude[proj_out, x_embedder, norm_out, context_embedder])
# freeze(transformer)
quantize(transformer, weightsqfloat8)
freeze(transformer)quantize(text_encoder_2, weightsqfloat8)
freeze(text_encoder_2)pipe FluxPipeline(schedulerscheduler,text_encodertext_encoder,tokenizertokenizer,text_encoder_2None,tokenizer_2tokenizer_2,vaevae,transformerNone,
)
pipe.text_encoder_2 text_encoder_2
pipe.transformer transformer
pipe.enable_model_cpu_offload()def generate(prompt, steps, guidance, width, height, seed):if seed -1:seed torch.seed()generator torch.Generator().manual_seed(int(seed))image pipe(promptprompt,widthwidth,heightheight,num_inference_stepssteps,generatorgenerator,guidance_scaleguidance,).images[0]return imagedemo gr.Interface(fngenerate, inputs[textbox, gr.Number(value4), gr.Number(value3.5), gr.Slider(0, 1920, value1024, step2), gr.Slider(0, 1920, value1024, step2), gr.Number(value-1)], outputsimage)demo.launch(server_name0.0.0.0)1-4. 启动 Web UI
python flux_on_potato.py1-5. 访问 Web UI
使用浏览器打开 http://localhost:7860 就可以访问了。 reference:
https://blackforestlabs.ai/announcing-black-forest-labs/https://github.com/black-forest-labs/flux/https://github.com/black-forest-labs/flux/issues/7https://gist.github.com/AmericanPresidentJimmyCarter/873985638e1f3541ba8b00137e7dacd9
