Part Number:TDA4VM
在使用edge-ai-tools编译onnx模型时发现,有的模型的maxpool和add层是支持的,有的模型的maxpool和add层不支持,请问这是什么原因。
这个模型是支持的
Supported TIDL layer type — Conv — Conv_0Supported TIDL layer type — Relu — Relu_1Supported TIDL layer type — MaxPool — MaxPool_2Supported TIDL layer type — Conv — Conv_3Supported TIDL layer type — Relu — Relu_4Supported TIDL layer type — Conv — Conv_5Supported TIDL layer type — Add — Add_6
别的模型却不支持
Unsupported (TIDL check) TIDL layer type — MaxPool
Unsupported (TIDL check) TIDL layer type — Add
Shine:
已咨询e2e工程师,请关注下面帖子的回复。https://e2e.ti.com/support/processors-group/processors/f/processors-forum/1274063/tda4vm-unsupported-tidl-check-tidl-layer-type—–maxpool
,
Shine:
请看下面e2e工程师的回复。
The maxpooling kernel specs are different for different models.
We have list of supported operators present here, please check the same.
https://github.com/TexasInstruments/edgeai-tidl-tools/blob/master/docs/supported_ops_rts_versions.md