}
}
+ // need to zero dst since we are accumulating into it
+ memset(dst->data, 0, ggml_nbytes(dst));
+
return;
}
const float * const src = (float *)((char *) src0->data + i02*nb02 + i01*nb01);
float * dst_data = wdata + i01*ne00*ne02;
for (int64_t i00 = 0; i00 < ne00; i00++) {
- dst_data[i01*ne00*ne02 + i00*ne02 + i02] = src[i00];
+ dst_data[i00*ne02 + i02] = src[i00];
}
}
}
}
}
+ // need to zero dst since we are accumulating into it
+ memset(dst->data, 0, ggml_nbytes(dst));
+
return;
}
}
}
+ memset(dst->data, 0, ggml_nbytes(dst));
+
return;
}