llama.cpp
f696428c - graph : add clamping to ffn_moe_weights_sum to avoid div-by-zero (#16655)

Commit

55 days ago

graph : add clamping to ffn_moe_weights_sum to avoid div-by-zero (#16655) * add missing norm topk bias * use clamping instead, update number and add comment

References

#16655 - graph : add clamping to ffn_moe_weights_sum to avoid div-by-zero

Author

CISC

Parents

7cce4f81

llama.cpp f696428c - graph : add clamping to ffn_moe_weights_sum to avoid div-by-zero (#16655)

llama.cpp
f696428c - graph : add clamping to ffn_moe_weights_sum to avoid div-by-zero (#16655)