I am trying to build the ptl weights for one of the simple SRE model. I tried to simplify the model so that it doesn’t contain any complex operations.
But it’s still throwing this ambiguous error.
RecursiveScriptModule(original_name=MainModel)
Traceback (most recent call last):
File "melspec_test.py", line 90, in <module>
optimized_model._save_for_lite_interpreter("vgg_sre.ptl")
File "C:\ProgramData\Anaconda3\envs\py38\lib\site-packages\torch\jit\_script.py", line 707, in _save_for_lite_interpreter
return self._c._save_for_mobile(*args, **kwargs)
RuntimeError: ENTER is not supported in mobile module.
My code:
import torch
import torchaudio
import torch.nn as nn
import torch.nn.functional as F
from torch.nn import Parameter
from torch.utils.mobile_optimizer import optimize_for_mobile
class MainModel(nn.Module):
def __init__(self, nOut = 1024):
super(MainModel, self).__init__()
self.netcnn = nn.Sequential(
nn.Conv2d(1, 96, kernel_size=(5,7), stride=(1,2), padding=(2,2)),
nn.BatchNorm2d(96),
nn.ReLU(inplace=True),
nn.MaxPool2d(kernel_size=(1,3), stride=(1,2)),
nn.Conv2d(96, 256, kernel_size=(5,5), stride=(2,2), padding=(1,1)),
nn.BatchNorm2d(256),
nn.ReLU(inplace=True),
nn.MaxPool2d(kernel_size=(3,3), stride=(2,2)),
nn.Conv2d(256, 384, kernel_size=(3,3), padding=(1,1)),
nn.BatchNorm2d(384),
nn.ReLU(inplace=True),
nn.Conv2d(384, 256, kernel_size=(3,3), padding=(1,1)),
nn.BatchNorm2d(256),
nn.ReLU(inplace=True),
nn.Conv2d(256, 256, kernel_size=(3,3), padding=(1,1)),
nn.BatchNorm2d(256),
nn.ReLU(inplace=True),
nn.MaxPool2d(kernel_size=(3,3), stride=(2,2)),
nn.Conv2d(256, 512, kernel_size=(4,1), padding=(0,0)),
nn.BatchNorm2d(512),
nn.ReLU(inplace=True),
)
self.sap_linear = nn.Linear(512, 512)
self.attention = self.new_parameter(512, 1)
out_dim = 512
self.fc = nn.Linear(out_dim, nOut)
def new_parameter(self, *size):
out = nn.Parameter(torch.FloatTensor(*size))
nn.init.xavier_normal_(out)
return out
def forward(self, x):
with torch.no_grad():
x = x.log()
x = x.unsqueeze(1)
x = self.netcnn(x)
x = x.permute(0, 2, 1, 3)
x = x.squeeze(dim=1).permute(0, 2, 1) # batch * L * D
h = torch.tanh(self.sap_linear(x))
w = torch.matmul(h, self.attention).squeeze(dim=2)
w = F.softmax(w, dim=1).view(x.size(0), x.size(1), 1)
x = torch.sum(x * w, dim=1)
x = self.fc(x)
return x
model = MainModel()
model.eval()
x = torch.rand(1, 40, 200)
traced_cell = torch.jit.trace(model, x)
quantized_model = torch.quantization.quantize_dynamic(
model, qconfig_spec={torch.nn.Linear}, dtype=torch.qint8)
scripted_model = torch.jit.script(quantized_model)
optimized_model = optimize_for_mobile(scripted_model)
print(optimized_model)
optimized_model._save_for_lite_interpreter("vgg_sre.ptl")
I am using pytorch torch-1.11.0.dev20220108+cpu (nightly) on windows.
I am not sure, for which operations it’s failing, there’s no clear error message.