Local Dream
正式版 实用工具 · AI工具io.github.xororz.localdream
2.4.0_with_filter
·
Build 58
rate_review
写点评
bookmark_border
60
下载
702
大小
63.98 MB
架构
[arm64-v8a]
兼容
Android 9.0+
提取机型
Redmi K70 Pro
warning_amber
https://github.com/xororz/local-dream
如果觉得好用,可以支持一下软件作者,项目主页最下面赞助方式有国内的爱发电
smartphone截图
notes更新日志
This release introduces experimental SDXL support, currently limited to Snapdragon 8 Elite and Snapdragon 8 Elite Gen 5 / 8 Gen 5 devices. Output resolution is fixed at 1024×1024.
Two models are built into the app: SDXL Base and Anything XL
To convert your own SDXL models, refer to the SDXL conversion guide.
Low RAM Mode is enabled by default for SDXL — each model is loaded before inference and automatically released afterward. If your device has ≥ 16 GB RAM, you may disable this for faster performance.
Textual inversion embeddings and prompt weighting both work as expected on SDXL.
Added the LCM scheduler — pair with an LCM-distilled checkpoint for high-quality output in far fewer steps.
CLIP outputs are cached between generations. If the positive and negative prompts are unchanged, the text encoder is skipped entirely (QNN, MNN hybrid, and the SDXL dual-encoder path).
On QNN the unconditional UNet pass is skipped when CFG = 1, roughly halving step time in that mode.
Added an optional log collection feature. When enabled, logs are displayed after exiting the inference screen — useful for bug reports and troubleshooting. Keep this disabled under normal use.
Fixed an issue where custom model IDs could collide with built-in model IDs.
Fixed a potential memory leak in the QNN model lifecycle.
Removed the upscaler resolution limit on the GitHub version.
Variants :
LocalDream_xxx_with_filter: Block NSFW Results. Same as Google Play version.
LocalDream_xxx: No filter.
Co-Authored-By: Claude Opus 4.7
info应用介绍
本地运行文生图模型应用
————————————
使用 Snapdragon NPU 加速在 Android 上运行 Stable Diffusion。还支持CPU/gpu推理。
NPU 支持的设备:
- 骁龙 8gen1
- 骁龙8+第一代
- 骁龙 8gen2
- 骁龙 8gen3
- 骁龙 8gen4
CPU支持的设备:
- 全部
技术实施
NPU加速
SDK:利用 Hexagon NPU 的高通 QNN SDK
量化:W8A16静态量化以实现最佳性能
分辨率:固定 512×512 模型形状
性能:极快的推理速度
CPU/GPU 推理
框架:由 MNN 框架提供支持
量化:W8 动态量化
分辨率:灵活尺寸(128×128、256×256、384×384、512×512)
性能:速度适中,兼容性高
NPU 高分辨率支持
下载 512 分辨率模型后,您可以下载补丁以启用 768×768 和 1024×1024 图像生成。请注意,量化高分辨率模型可能会产生布局不佳的图像。我们建议首先以 512 分辨率生成,然后使用 img2img 的高分辨率模型(本质上是 Highres.fix)。建议的 img2img denoise_strength约为 0.75。
设备兼容性
NPU 加速支持
兼容具有以下特点的设备:
骁龙 8 Gen 1
骁龙 8+ 第 1 代
骁龙 8 Gen 2
骁龙 8 第 3 代
骁龙8精英
注意:其他设备无法下载 NPU 模型
CPU/GPU 支持
RAM 要求:~2GB 可用内存
兼容性:近年来的大多数 Android 设备
forum口碑与点评
--
推荐率
thumb_up0 推荐
thumb_down0 不推荐
chat_bubble_outline0 点评
加载中...
1.8.4
正式版