Instructions to use AXERA-TECH/lcm-lora-sdv1-5-ax620e with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AXERA-TECH/lcm-lora-sdv1-5-ax620e with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("AXERA-TECH/lcm-lora-sdv1-5-ax620e", dtype="auto") - Notebooks
- Google Colab
- Kaggle
SD1.5-LCM AX620E
本仓库是 lcm-lora-sdv1-5 的 AX620E 拆分版本,只包含 AX620E 可用的 axmodel 文件。客户只需要下载本仓库即可在 AX620E 板端运行对应分辨率模型。
AX650N 模型保留在原仓库:
AXERA-TECH/lcm-lora-sdv1-5
发布内容只保留 *.axmodel 推理所需文件。*.onnx 和 *.safetensors 不属于当前发布路径。
模型目录
| OpenAI model id | 目录 | 分辨率 | txt2img | img2img |
|---|---|---|---|---|
lcm-lora-sdv1-5-256x256 |
models/ |
256x256 | 支持 | 支持 |
lcm-lora-sdv1-5-320x320 |
models_320x320/ |
320x320 | 支持 | 支持 |
config.json 中的 image_variants 是 axllm serve 的模型发现入口。前端应先请求 /v1/models,再选择返回的真实模型 ID 发起生图请求。
启动服务
axllm serve /path/to/lcm-lora-sdv1-5-ax620e --port 8000
接口:
GET /v1/modelsPOST /v1/images/generationsPOST /v1/images/edits
文生图示例
curl -X POST http://127.0.0.1:8000/v1/images/generations \
-H 'Content-Type: application/json' \
-d '{
"model": "lcm-lora-sdv1-5-256x256",
"prompt": "a cat wearing sunglasses",
"size": "256x256",
"seed": 0,
"response_format": "b64_json"
}'
图生图示例
curl -X POST http://127.0.0.1:8000/v1/images/edits \
-F model=lcm-lora-sdv1-5-320x320 \
-F prompt="Astronauts in a jungle, cold color palette, muted colors, detailed, 8k" \
-F size=320x320 \
-F seed=0 \
-F response_format=b64_json \
-F image=@assets/img2img-init.png
Python Demo
pip3 install -r requirements.txt
python3 launcher.py --model_dir models --isize 256
python3 launcher.py --model_dir models_320x320 --isize 320
- Downloads last month
- 3
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for AXERA-TECH/lcm-lora-sdv1-5-ax620e
Base model
Lykon/dreamshaper-7