Andreas Rozek
Andreas Rozek
Thank you very much for your explanation. However, I somewhere found the "recipe" to first nullify an object and then to write a new one using the same context (I...
Very, very strange... If I change the above code to ``` const Gun = new GUN() const GunUser = Gun.user() // creates an environment for an anonymous user /**** find...
a small note: if I change `RAD_write` and set the `trivialRADStore_Store` directly (i.e., avoid to call `Rad`), the whole loop completes almost immediately - thus, its `RAD` that cause the...
Ok, my current observations so far: * entry "!" seems to collect written entries until they get stored away at their own keys * I have no idea what entry...
RAD also strangely reads the previously written key again - and writes it out again, unchanged! How useless is that?
I meanwhile set the `chunk` option to `1` for speedup Here is a transcript of the first few operations on my in-memory store: ``` entry #1: 0 trivialRADStore: reading %1C...
well, entry "%1C" seems to contain, what is called the "radix tree": the JSON string of an object with the hierarchy of keys. I have no idea what that is...
That's good to know - I already lost a day on this because I searched the problem on my side...will give up on RAD then!
I just learned that RWKV-5 models are not yet supported by rwkv.cpp. So I tried RWKV-4 instead - took the .pth model and converted it to .bin following the docs....
how do I check my GGML version? I'm using the current version of [rwkv.cpp](https://github.com/saharNooby/rwkv.cpp)