Can someone provide magnet link to 30B model please?

[thedmdim]

I cannot download from huggingface cause the size is too big and condition speed is too slow

[erhathaway]

On, linux:

wget https://huggingface.co/Pi3141/alpaca-30B-ggml/resolve/main/ggml-model-q4_0.bin -O ggml-model-q4_0.bin

[Castaa]

magnet:?xt=urn:btih:6K5O4J7DCKAMMMAJHWXQU72OYFXPZQJG&dn=ggml-alpaca-30b-q4.bin&xl=20333638921&tr=udp%3A%2F%2Ftracker.openbittorrent.com%3A80%2Fannounce

I hope this magnet link works properly. I've never created one before This is the alpaca.cpp 30B 4-bit weight file. Same file downloaded from huggingface. Let me know if it doesn't work. Apologies if it doesn't. I can confirm it would be far faster to simply download it from huggingface if you have a choice.

Also, I can confirm the file itself works with the current version of alpaca.cpp. Which is good news.

If anyone is able to download the entire file, could you seed it for a few days because I fear my upload speed is quite slow. I have the jank tier xfinity.

Update: It's being seeded by multiple users now. So the torrent download speed should be much improved.

[mallory303]

Unfortunately I have problems with it
Anybody has idea what is the problem?

llama_model_load: loading model from 'ggml-model-q4_0.bin' - please wait ...
llama_model_load: ggml ctx size = 25631.50 MB
llama_model_load: memory_size = 6240.00 MB, n_mem = 122880
llama_model_load: loading model part 1/4 from 'ggml-model-q4_0.bin'
llama_model_load: llama_model_load: tensor 'tok_embeddings.weight' has wrong size in model file
main: failed to load model from 'ggml-model-q4_0.bin'

[jellomaster]

make sure you get the latest release and recompile.
this should fix things:
git pull
make chat

[mallory303]

It's works now, thanks :)