Not able to use model in version 0.3.7
Error:
🥲 Failed to load the model
Failed to load model
llama.cpp error: 'error loading model vocabulary: unknown pre-tokenizer type: 'deepseek-r1-qwen''
+1
Model doesn't load in v0.3.6 either.
Try updating your runtime
Try updating your runtime
Thank you @bartowski - that resolve the issue for me.
I am getting the same error, what does "Try updating your runtime" mean. I'm on LM Studio 0.3.7.
Note I am using the model "DeepSeek-R1-Distill-Qwen-7B-GGUF/DeepSeek-R1-Distill-Qwen-7B-Q4_K_M.gguf" from lmstudio-community.
If you go to settings (cog wheel icon bottom right), and on the left of the settings window you select "Runtimes", you might see various options to update your runtime extension packs, or something similar. After I updated all three (for me) of those, I was able to load the model where I wasn't previously.
Updating Nvidia Cuda helps to fix this issue
How to update the runtime on the mac version?
To manage your runtime versions on Mac open the LMStudio app and go to Settings. Within the settings, locate the "Runtime" tab where you can see the runtimes which needs an update. Click on Download to update the runtimes you want.