Search results

torch_npu 2.10 + CANN 9.0 on soc_version 104: torch.nn.functional.linear fails with MatMul/TransData

Status: Open.

#144 In Ascend/pytorch;

· wakaka6 opened on Jun 8, 2026

算子launch线程数量有误，全部由acl_thread进行算子下发

Status: Open.

#140 In Ascend/pytorch;

· AcK1ng opened on May 24, 2026

RFC: 为 torch_npu MemPool 增加 sub_block_alignment 参数

Status: Open.

#139 In Ascend/pytorch;

· whybeyoung opened on May 22, 2026

torch_npu profiler 输出的 trace JSON 与 Perfetto 不兼容：ts 为字符串类型 + 绝对时间戳导致精度丢失

Status: Open.

#133 In Ascend/pytorch;

· JiaryCoder opened on May 13, 2026

CANN 9.0 Support

Status: Open.

#130 In Ascend/pytorch;

· chengchen666 opened on May 8, 2026

torch.zero_ in compile

Status: Open.

#128 In Ascend/pytorch;

· arabel1a opened on May 4, 2026

[Bug] ATC allow_mix_precision / allow_fp32_to_fp16 在 SAM3 vision encoder FPN neck 上造成灾难性精度崩溃 (与 torch_npu fp16 行为不一致)

Status: Open.

#125 In Ascend/pytorch;

· ZhaoyuZhong0913 opened on Apr 24, 2026

torch_npu/_inductor 中 MLIRProcessor 使用 shell=True 执行由环境变量拼接的命令，存在条件性命令注入风险

Status: Open.

#123 In Ascend/pytorch;

· ycaibb opened on Apr 22, 2026

【feature】当配置为cuda和nccl时，ptx 不支持 创建通信域

Status: Open.

#121 In Ascend/pytorch;

· littleyellowbicycle opened on Apr 20, 2026

[Usage] AclnnGroupedMatmulWeightNz does not support x dtype int8 and weight dtype float16

Status: Open.

#110 In Ascend/pytorch;

· ZhongsJie opened on Mar 23, 2026

Question: 在torch.npu的使用环境下，通过pytorch官方文档实现的自定义的Process Group Backend会有效果吗？

Status: Open.

#109 In Ascend/pytorch;

· 525309178 opened on Mar 19, 2026

执行deepseekv3.2-exp模型量化报错：torch_npu.memory: [ERROR] [269128] Get a block from the existing pool failed. Try to free cached blocks and reallocate. This error log can be ignored.

Status: Open.

#108 In Ascend/pytorch;

· mominhua opened on Mar 17, 2026