1.78bit quant of full R1-0528 error
#10
by
d2rx
- opened
Running into this error when trying to load onto llama-server
llama_model_load: error loading model: corrupted model: 1086 tensors expected but 1044 found
Corrupted model file?
Running into this error when trying to load onto llama-server
llama_model_load: error loading model: corrupted model: 1086 tensors expected but 1044 found
Corrupted model file?
Did you update to the latest version of llamacpp? We re-uploaded a smaller version. Could you test it?
Just did that. All fixed now! Getting about 14~15 token per second on dual RTX Ada 6000 with 2TB system RAM