llama.cpp
381ee195 - finetune : fix ggml_allocr lifetimes (tmp workaround) (#5033)

Commit

1 year ago

finetune : fix ggml_allocr lifetimes (tmp workaround) (#5033) * Fix issue with alloc causing max_compute_size to be calculated * remove ggml_allocr_free as suggested in issue #4791

References

#5033 - Fix issue #4791 alloc causes compute_size to be calculated incorrectly in train-text-from-scratch, end result core dump

Author

bzuzo

Parents

a5cacb22

llama.cpp 381ee195 - finetune : fix ggml_allocr lifetimes (tmp workaround) (#5033)

llama.cpp
381ee195 - finetune : fix ggml_allocr lifetimes (tmp workaround) (#5033)