In the syno storage management, it can’t show the disk information.
So first check the “invalid ELF” log in the “/var/log”
root@DiskStation:/var/log# grep -R "invalid ELF" *
grep -R "2024-05-01.*invalid ELF" *messages:2024-05-01T22:48:55+08:00 DiskStation Refer the content from the link, it has the complete list for different Security Companies, so backup the content here.
How to remove your website from Security Blacklists
Last Updated 9 months ago
There are so many large language models that you can run it locally, e.g. llama, mixtral.
The most popular open source LLM framework that support run many LLM locally.
git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
make
./examples/chat-13B.shBefore you …
If you want to add the Nginx proxy for OpenAI API
You can use the following nginx configs.
location /v1/chat/completions {
proxy_pass https://api.openai.com/v1/chat/completions;
proxy_ssl_name api.openai.com;
proxy_ssl_server_name on;
# Forward all headers
proxy_pass_request_headers on;
# proxy_http_version 1.1;
}And then restart …
Refer the GoDaddy document
Find the settings for you domain want to transfer
In the CloudFlare, it says, so I need to wait for some time.
…Registry status: Client transfer prohibited. Please unlock and allow a few hours
fatal error: 'openssl/opensslv.h' file not found
brew install openssl
env LDFLAGS=”-L$(brew –prefix openssl)/lib” CFLAGS=”-I$(brew –prefix openssl)/include” pip install cryptography
…
昨天GPT-4发布，除了惊艳还是惊艳，OpenAI的进化速度超乎想象。
除此之外，最近我也在关注LLaMa((Large Language Model Meta AI)
Facebook的大型语言模型，现在已经有很多基于它进行优化的项目
比如Standford, 号称7B model已经达到 OpenAI text-davinci-003效果
而我这里选择的是llama.cpp, 作者刚把whisper用cpp实现了一遍叫whisper.cpp, 还真是直接
整个安装过程还参考了这篇文章
LLaMa的优势
先看看我本地用了30B model的效果
git clone git@github.com:ggerganov/llama.cpp.git
cd llama.cpp
python3 关注AI的都应该知道最近AI界，各种技术进化的速度，已经超过想象了。
比如说
各种模型迭代的速度，正在很疯狂的往前逛奔，每天早上都能看到一些意想不到的新奇的东西。
而今天看到的是有关大脑的
这篇论文提出了一种基于渗透模型的图像重建方法，以重建人脑活动的视觉体验。研究者使用了一种名为Stable Diffusion的潜在渗透模型，该模型减少了渗透模型的计算成本，同时保留了其高生成性能。他们进一步研究了不同组件如何与不同的大脑功能相关联，并表明该方法可以直接重建高分辨率图像，不需要额外的复杂深度学习模型的训练和微调。研究为重建人脑活动图像提供了一种有前途的方法，并为理解渗透模型提供了一个新的框架。
简单的说就是可以通过仪器扫描大脑的活动，然后通过Stable Diffusion重现大脑的景象，相当于视觉你的大脑所想。
这个潘多拉魔盒一旦打开，以后人类是不是就没有秘密可言了
Try the Stable Diffusion, and ChilloutMix, and LoRA to generate the images on Apple M1.
I follow this guideline to setup the Stable Diffusion running on my Apple M1.
brew install cmake protobuf rust python@3.10 
This python script is used to interact with OpenAI’s ChatGPT model via the MiSpeaker. The script fetches questions from the MiSpeaker, passes the questions to the ChatGPT model, and returns the answers back to the MiSpeaker for …