xiexiaozheng

> @xiexiaozheng, we don't support Conv1d for bias correction ATM. If you would like, please try the Adaround technique instead of bias correction, it has proven to give much better...

Hi, @quic-mangal The output of compute_encodings is calculated under the condition that the input of the layer has quantization noise. Where does the initial quantization parameter that compute_encodings needs to...

> > Where does the initial quantization parameter that compute_encodings needs to calculate the result come from? > > By parameters, I mean weights and bias. Since weights are bias...

Hi, @quic-akinlawo , for the firsht question, the hardswish layer is an activation layer, and when the model is running on the DSP, the conv layer and hardswish layer are...

> @xiexiaozheng, in the compute encoding Fn we do a forward pass wherein we pass data through the model, this data's min max are accumulated to get the min max...

@quic-mangal ,Hi, since in the forward pass stage, quantization nodes are enabled, where does the initial quantization stats inside the quantization nodes come from?

> This is the initial value- > > ``` > double min = std::numeric_limits::max(); > double max = -std::numeric_limits::max(); > ``` @quic-mangal ok, I see. Thank you for your patient...

> Hello @xiexiaozheng , if you're using per channel quantization for a given op then the quantization range needs to be symmetric i.e. offset should be zero or 2^bw -...

> @xiexiaozheng, which framework are you using? We support both CPU & GPU and for TF & ONNX framework, use_cuda argument needs to be set explicitly. For PT you need...

> Hello @xiexiaozheng, per channel quantization is recommended for convolution-like operations to increase the resolution. @quic-akinlawo But the accuracy of QAT model is ok, the accuracy decrease when the model...