lcm-lora-sd

モデルを手動でダウンロードして Raspberry Pi 5 にアップロードするか、または以下のコマンドでモデルリポジトリを取得します。

ヒント

git lfs がインストールされていない場合は、事前にgit lfs インストール手順を参照してインストールしてください。

git clone https://huggingface.co/AXERA-TECH/lcm-lora-sdv1-5

ファイル説明：

m5stack@raspberrypi:~/rsp/lcm-lora-sdv1-5 $ ls -lh
total 100K
drwxrwxr-x 2 m5stack m5stack 4.0K Aug 11 17:10 asserts
-rw-rw-r-- 1 m5stack m5stack    0 Aug 11 17:09 config.json
-rw-rw-r-- 1 m5stack m5stack 1.2K Aug 11 17:09 Disclaimer.md
-rw-rw-r-- 1 m5stack m5stack 1.5K Aug 11 17:09 LICENSE
drwxrwxr-x 4 m5stack m5stack 4.0K Aug 11 17:14 models
-rw-rw-r-- 1 m5stack m5stack 5.7K Aug 11 17:09 README.md
-rw-rw-r-- 1 m5stack m5stack  117 Aug 11 17:09 requirements.txt
-rw-rw-r-- 1 m5stack m5stack  23K Aug 11 17:09 run_img2img_axe_infer.py
-rw-rw-r-- 1 m5stack m5stack  23K Aug 11 17:09 run_img2img_onnx_infer.py
-rw-rw-r-- 1 m5stack m5stack 7.8K Aug 11 17:09 run_txt2img_axe_infer_new.py
-rw-rw-r-- 1 m5stack m5stack 7.3K Aug 11 17:09 run_txt2img_axe_infer.py
-rw-rw-r-- 1 m5stack m5stack 7.4K Aug 11 17:09 run_txt2img_onnx_infer.py

仮想環境を作成

python -m venv sd

仮想環境をアクティベート

source sd/bin/activate

依存パッケージをインストール

sudo apt install cmake -y
pip install -r requirements.txt
pip install https://github.com/AXERA-TECH/pyaxengine/releases/download/0.1.3.rc1/axengine-0.1.3-py3-none-any.whl

テキストから画像生成（文生図）

python run_txt2img_axe_infer.py

実行結果：

(sd) m5stack@raspberrypi:~/rsp/lcm-lora-sdv1-5$ python run_txt2img_axe_infer.py
[INFO] Available providers:  ['AXCLRTExecutionProvider']
prompt: Self-portrait oil painting, a beautiful cyborg with golden hair, 8k
text_tokenizer: ./models/tokenizer
text_encoder: ./models/text_encoder
unet_model: ./models/unet.axmodel
vae_decoder_model: ./models/vae_decoder.axmodel
time_input: ./models/time_input_txt2img.npy
save_dir: ./txt2img_output_axe.png
[INFO] Using provider: AXCLRTExecutionProvider
[INFO] SOC Name: AX650N
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Compiler version: 3.4 9215b7e5
text encoder take 4466.2ms
[INFO] Using provider: AXCLRTExecutionProvider
[INFO] SOC Name: AX650N
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Compiler version: 3.3 972f38ca
[INFO] Using provider: AXCLRTExecutionProvider
[INFO] SOC Name: AX650N
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Compiler version: 3.3 972f38ca
load models take 22582.9ms
unet once take 433.6ms
unet once take 433.5ms
unet once take 433.3ms
unet once take 433.5ms
unet loop take 1736.7ms
vae inference take 914.2ms
save image take 254.3ms 

画像から画像生成（図生図）

python run_txt2img_axe_infer.py

実行結果：

(sd) m5stack@raspberrypi:~/rsp/lcm-lora-sdv1-5$ python run_txt2img_axe_infer.py
[INFO] Available providers:  ['AXCLRTExecutionProvider']
prompt: Self-portrait oil painting, a beautiful cyborg with golden hair, 8k
text_tokenizer: ./models/tokenizer
text_encoder: ./models/text_encoder
unet_model: ./models/unet.axmodel
vae_decoder_model: ./models/vae_decoder.axmodel
time_input: ./models/time_input_txt2img.npy
save_dir: ./txt2img_output_axe.png
[INFO] Using provider: AXCLRTExecutionProvider
[INFO] SOC Name: AX650N
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Compiler version: 3.4 9215b7e5
text encoder take 2962.1ms
[INFO] Using provider: AXCLRTExecutionProvider
[INFO] SOC Name: AX650N
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Compiler version: 3.3 972f38ca
[INFO] Using provider: AXCLRTExecutionProvider
[INFO] SOC Name: AX650N
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Compiler version: 3.3 972f38ca
load models take 14483.6ms
unet once take 433.6ms
unet once take 433.3ms
unet once take 433.2ms
unet once take 433.4ms
unet loop take 1736.4ms
vae inference take 914.0ms
save image take 233.2ms 

Next Overview

Linux PC

CM4Stack

CoreMP135

AI Accelerator Card

LLM-8850 Card

Quick Start

Vision Models

Large Language Models

Multimodal Models

Audio Models

Generative Models

Advanced Usage

LLM

リアルタイム音声アシスタント

OpenAI ボイスアシスタント

XiaoZhi ボイスアシスタント

XiaoLing ボイスアシスタント

AtomS3R-M12 Volcengine Kit

オフライン音声認識

Unit ASR

Home Assistant

Home Assistant OS

ESPHome

Industrial Control

StamPLC

IoT Measuring Instruments

Air Quality

Module13.2 PPS

VAMeter

T-Lite

Ezdata

Ethernet Camera

PoECAM

Wi-Fi Camera

TimerCAM

Unit CamS3/-5MP

AI Camera

UnitV2

M5StickV/UnitV

LoRa & LoRaWAN

TTN (The Things Network)

Meshtastic

Motor Control

Unit Roller485/CAN

Develop Tools

Hobby Kit

ファームウェアの初期化