This is not the first time I've heard of checkpoints being used to distribute malware. In fact, I've heard this was a popular vector from shady international groups.
I wouldn't expect this from Bilibili's Index Team, though, given how high profile they are. It's probably(?) a false positive. Though I wouldn't use it personally, just to be safe.
The safetensors format should be used by everyone. Raw pth files and pickle files should be shunned and abandoned by the industry. It's a bad format.
I imagine it has more to do with whether or not the file appears to have executable python code in it, as a .pth file is usually just a a pickled python object and these can be manipulated to load arbitrary python code when loaded.
This is cool, but it only seems to work on the host that has the /etc/hosts loopback redirect to *.localhost. I run my app on a home server and access it from multiple PCs on the LAN. I have several apps, each associated with a different port number. Right now, I rely on a start page (https://github.com/dh1011/start-ichi) to keep track of all those ports. I’m wondering if there’s an easy way to resolve custom domains to each of those apps?
You need to install a DNS server in your network and configure everything to use it (probably via DHCP). Then you can configure whatever you like. Dnsmasq is quite easy to get started. There are handy integrated solutions like pihole that combine DHCP and DNS. I run things like this on my router (opnsense).
I'm working on a VSCode extension that lets me quickly copy code from a workspace into a structured format for LLMs.
This makes it easy to provide context so the LLM can continue developing the codebase.
There are other similar tools out there—mostly web apps or CLI-based—but I found a VSCode extension to be the fastest and most convenient option for VSCode users.
chatgpt did a stellar job parsing the "books on hard things" thread from a little while ago. my prompt was:
Can you identify all the books here, sorted by a weight which is determined based on a combo of the number of votes the comment has, the number of sub-comments, or the number of repeat mentions.
I completely agree that bookmarks should prioritize content over metadata. This is actually what led me to develop the "semantic-bookmark-manager" [1]. It uses LLM to summarize the content of bookmarked pages and generate relevant tags. Additionally, it utilizes RAG to facilitate semantic searching within your bookmarks.
I am working on "Subscription Manager," a simple single-page web application that allows users to manage their recurring payments. It provides a summary of your expenses, including weekly, monthly, and yearly totals for all your current subscriptions. Additionally, it features notifications through NTFY to notify when a subscription is dued for payment.
