search

thisserand / llama2_local

MIT License
82 stars 35 forks source link

Always loads the same model #2

Closed kevinkaw closed 1 year ago

kevinkaw commented 1 year ago

llama.py always loads the model llama-2-7b-chat.ggmlv3.q4_K_M.bin even if I change models.

alexng88 commented 1 year ago

modify the llama.py for your favorite model.

thisserand commented 1 year ago

Yes, you're right. I forgot to remove the constant initialisation of the models. Thanks for pointing it out, I fixed now. :-)