Thanks for sending an issue! Please fill in the following template to help quickly solve your problem.
matmul/quant_batch_matmul_v3/docs/aclnnQuantMatmulV4.md matmul/quant_batch_matmul_v3/docs/aclnnQuantMatmulWeightNz.md
x1 x2 int8输入,scale为float32和bfloat32场景,且为perchannel和pertoken场景支持大于65535,当时资料没有体现。
Thanks for sending an issue! Please fill in the following template to help quickly solve your problem.
Document Link(文档链接)
matmul/quant_batch_matmul_v3/docs/aclnnQuantMatmulV4.md
matmul/quant_batch_matmul_v3/docs/aclnnQuantMatmulWeightNz.md
Issues Section(问题文档片段)
Existing Issues(存在的问题)
x1 x2 int8输入,scale为float32和bfloat32场景,且为perchannel和pertoken场景支持大于65535,当时资料没有体现。