Quantization occurs with RuntimeError: zero_point
must be between quant_min
and quant_max
.
#126266
Labels
oncall: quantization
Quantization support in PyTorch
馃悰 Describe the bug
The strange thing is that when I train only 100 epochs with FP32, the model can quantize normally, when I train 200 or more epochs and then try to do the quantization - the model reports the following error.
It says: RuntimeError:
zero_point
must be betweenquant_min
andquant_max
.I double checked #89619 but got no help.
Here is my code,
Versions
Here is the version info.
cc @jerryzh168 @jianyuh @raghuramank100 @jamesr66a @vkuzo @jgong5 @Xia-Weiwen @leslie-fang-intel
The text was updated successfully, but these errors were encountered: