class <lambda>(torch.nn.Module):
    def forward(self, arg0_1: "f16[4624, 1, 256]", arg1_1: "f16[4624, 1, 256]", arg2_1: "f32[4624, 1, 256]", arg3_1: "f32[256]", arg4_1: "f32[256]", arg5_1: "f32[1024, 256]", arg6_1: "f32[1024]", arg7_1: "f32[1024]", arg8_1: "f32[1024]", arg9_1: "f32[1024, 1, 5, 5]", arg10_1: "f32[256, 1024]", arg11_1: "f32[256]"):
        # File: /workspace/networks/layers/transformer.py:841 in torch_dynamo_resume_in_forward_at_836, code: tgt = tgt + self.lst_dropout(tgt2 + tgt3)
        add: "f16[4624, 1, 256]" = torch.ops.aten.add.Tensor(arg1_1, arg0_1);  arg1_1 = arg0_1 = None
        add_1: "f32[4624, 1, 256]" = torch.ops.aten.add.Tensor(arg2_1, add);  arg2_1 = add = None
        
        # File: /opt/conda/lib/python3.11/site-packages/torch/nn/modules/normalization.py:201 in forward, code: return F.layer_norm(
        var_mean = torch.ops.aten.var_mean.correction(add_1, [2], correction = 0, keepdim = True)
        getitem: "f32[4624, 1, 1]" = var_mean[0]
        getitem_1: "f32[4624, 1, 1]" = var_mean[1];  var_mean = None
        
        # File: /opt/conda/lib/python3.11/site-packages/torch/nn/modules/linear.py:116 in forward, code: return F.linear(input, self.weight, self.bias)
        convert_element_type: "f16[1024]" = torch.ops.prims.convert_element_type.default(arg6_1, torch.float16);  arg6_1 = None
        
        # File: /opt/conda/lib/python3.11/site-packages/torch/nn/modules/normalization.py:201 in forward, code: return F.layer_norm(
        sub: "f32[4624, 1, 256]" = torch.ops.aten.sub.Tensor(add_1, getitem_1);  getitem_1 = None
        add_2: "f32[4624, 1, 1]" = torch.ops.aten.add.Tensor(getitem, 1e-05);  getitem = None
        rsqrt: "f32[4624, 1, 1]" = torch.ops.aten.rsqrt.default(add_2);  add_2 = None
        mul: "f32[4624, 1, 256]" = torch.ops.aten.mul.Tensor(sub, rsqrt);  sub = rsqrt = None
        mul_1: "f32[4624, 1, 256]" = torch.ops.aten.mul.Tensor(mul, arg3_1);  mul = arg3_1 = None
        add_3: "f32[4624, 1, 256]" = torch.ops.aten.add.Tensor(mul_1, arg4_1);  mul_1 = arg4_1 = None
        
        # File: /opt/conda/lib/python3.11/site-packages/torch/nn/modules/linear.py:116 in forward, code: return F.linear(input, self.weight, self.bias)
        convert_element_type_2: "f16[4624, 1, 256]" = torch.ops.prims.convert_element_type.default(add_3, torch.float16);  add_3 = None
        view: "f16[4624, 256]" = torch.ops.aten.reshape.default(convert_element_type_2, [4624, 256]);  convert_element_type_2 = None
        convert_element_type_1: "f16[1024, 256]" = torch.ops.prims.convert_element_type.default(arg5_1, torch.float16);  arg5_1 = None
        permute: "f16[256, 1024]" = torch.ops.aten.permute.default(convert_element_type_1, [1, 0]);  convert_element_type_1 = None
        
        # No stacktrace found for following nodes
        mm_default: "f16[4624, 1024]" = torch.ops.aten.mm.default(view, permute);  view = permute = None
        add_tensor: "f16[4624, 1024]" = torch.ops.aten.add.Tensor(mm_default, convert_element_type);  mm_default = convert_element_type = None
        
        # File: /opt/conda/lib/python3.11/site-packages/torch/nn/modules/linear.py:116 in forward, code: return F.linear(input, self.weight, self.bias)
        view_1: "f16[4624, 1, 1024]" = torch.ops.aten.reshape.default(add_tensor, [4624, 1, 1024]);  add_tensor = None
        
        # File: /workspace/networks/layers/basic.py:30 in forward, code: x = x.view(h, w, bs, c).permute(2, 3, 0, 1)
        view_2: "f16[68, 68, 1, 1024]" = torch.ops.aten.reshape.default(view_1, [68, 68, 1, 1024]);  view_1 = None
        permute_1: "f16[1, 1024, 68, 68]" = torch.ops.aten.permute.default(view_2, [2, 3, 0, 1]);  view_2 = None
        
        # File: /opt/conda/lib/python3.11/site-packages/torch/nn/modules/normalization.py:287 in forward, code: return F.group_norm(
        convert_element_type_6: "f32[1, 1024, 68, 68]" = torch.ops.prims.convert_element_type.default(permute_1, torch.float32);  permute_1 = None
        clone: "f32[1, 1024, 68, 68]" = torch.ops.aten.clone.default(convert_element_type_6, memory_format = torch.contiguous_format);  convert_element_type_6 = None
        view_3: "f32[1, 32, 32, 4624]" = torch.ops.aten.reshape.default(clone, [1, 32, 32, 4624]);  clone = None
        var_mean_1 = torch.ops.aten.var_mean.correction(view_3, [2, 3], correction = 0, keepdim = True)
        getitem_2: "f32[1, 32, 1, 1]" = var_mean_1[0]
        getitem_3: "f32[1, 32, 1, 1]" = var_mean_1[1];  var_mean_1 = None
        sub_1: "f32[1, 32, 32, 4624]" = torch.ops.aten.sub.Tensor(view_3, getitem_3);  view_3 = getitem_3 = None
        add_4: "f32[1, 32, 1, 1]" = torch.ops.aten.add.Tensor(getitem_2, 1e-05);  getitem_2 = None
        rsqrt_1: "f32[1, 32, 1, 1]" = torch.ops.aten.rsqrt.default(add_4);  add_4 = None
        mul_2: "f32[1, 32, 32, 4624]" = torch.ops.aten.mul.Tensor(sub_1, rsqrt_1);  sub_1 = rsqrt_1 = None
        view_4: "f32[1, 1024, 68, 68]" = torch.ops.aten.reshape.default(mul_2, [1, 1024, 68, 68]);  mul_2 = None
        unsqueeze_3: "f32[1, 1024]" = torch.ops.aten.unsqueeze.default(arg7_1, 0);  arg7_1 = None
        unsqueeze_4: "f32[1, 1024, 1]" = torch.ops.aten.unsqueeze.default(unsqueeze_3, 2);  unsqueeze_3 = None
        unsqueeze_5: "f32[1, 1024, 1, 1]" = torch.ops.aten.unsqueeze.default(unsqueeze_4, 3);  unsqueeze_4 = None
        mul_3: "f32[1, 1024, 68, 68]" = torch.ops.aten.mul.Tensor(view_4, unsqueeze_5);  view_4 = unsqueeze_5 = None
        unsqueeze: "f32[1, 1024]" = torch.ops.aten.unsqueeze.default(arg8_1, 0);  arg8_1 = None
        unsqueeze_1: "f32[1, 1024, 1]" = torch.ops.aten.unsqueeze.default(unsqueeze, 2);  unsqueeze = None
        unsqueeze_2: "f32[1, 1024, 1, 1]" = torch.ops.aten.unsqueeze.default(unsqueeze_1, 3);  unsqueeze_1 = None
        add_5: "f32[1, 1024, 68, 68]" = torch.ops.aten.add.Tensor(mul_3, unsqueeze_2);  mul_3 = unsqueeze_2 = None
        
        # File: /workspace/networks/layers/basic.py:32 in forward, code: x = F.gelu(x)
        mul_4: "f32[1, 1024, 68, 68]" = torch.ops.aten.mul.Tensor(add_5, 0.5)
        mul_5: "f32[1, 1024, 68, 68]" = torch.ops.aten.mul.Tensor(add_5, 0.7071067811865476);  add_5 = None
        erf: "f32[1, 1024, 68, 68]" = torch.ops.aten.erf.default(mul_5);  mul_5 = None
        add_6: "f32[1, 1024, 68, 68]" = torch.ops.aten.add.Tensor(erf, 1);  erf = None
        mul_6: "f32[1, 1024, 68, 68]" = torch.ops.aten.mul.Tensor(mul_4, add_6);  mul_4 = add_6 = None
        
        # File: /opt/conda/lib/python3.11/site-packages/torch/nn/modules/conv.py:453 in _conv_forward, code: return F.conv2d(input, weight, bias, self.stride,
        convert_element_type_8: "f16[1, 1024, 68, 68]" = torch.ops.prims.convert_element_type.default(mul_6, torch.float16);  mul_6 = None
        convert_element_type_7: "f16[1024, 1, 5, 5]" = torch.ops.prims.convert_element_type.default(arg9_1, torch.float16);  arg9_1 = None
        convolution: "f16[1, 1024, 68, 68]" = torch.ops.aten.convolution.default(convert_element_type_8, convert_element_type_7, None, [1, 1], [2, 2], [1, 1], False, [0, 0], 1024);  convert_element_type_8 = convert_element_type_7 = None
        
        # File: /workspace/networks/layers/basic.py:34 in forward, code: x = x.view(bs, c, h * w).permute(2, 0, 1)
        view_5: "f16[1, 1024, 4624]" = torch.ops.aten.reshape.default(convolution, [1, 1024, 4624]);  convolution = None
        permute_2: "f16[4624, 1, 1024]" = torch.ops.aten.permute.default(view_5, [2, 0, 1]);  view_5 = None
        
        # File: /opt/conda/lib/python3.11/site-packages/torch/nn/modules/linear.py:116 in forward, code: return F.linear(input, self.weight, self.bias)
        expand: "f16[4624, 1, 1024]" = torch.ops.aten.expand.default(permute_2, [4624, 1, 1024]);  permute_2 = None
        convert_element_type_10: "f16[256, 1024]" = torch.ops.prims.convert_element_type.default(arg10_1, torch.float16);  arg10_1 = None
        permute_3: "f16[1024, 256]" = torch.ops.aten.permute.default(convert_element_type_10, [1, 0]);  convert_element_type_10 = None
        expand_1: "f16[4624, 1024, 256]" = torch.ops.aten.expand.default(permute_3, [4624, 1024, 256]);  permute_3 = None
        bmm: "f16[4624, 1, 256]" = torch.ops.aten.bmm.default(expand, expand_1);  expand = expand_1 = None
        convert_element_type_9: "f16[256]" = torch.ops.prims.convert_element_type.default(arg11_1, torch.float16);  arg11_1 = None
        add_7: "f16[4624, 1, 256]" = torch.ops.aten.add.Tensor(bmm, convert_element_type_9);  bmm = convert_element_type_9 = None
        
        # File: /workspace/networks/layers/transformer.py:848 in torch_dynamo_resume_in_forward_at_836, code: tgt = tgt + self.droppath(tgt2)
        add_8: "f32[4624, 1, 256]" = torch.ops.aten.add.Tensor(add_1, add_7);  add_1 = add_7 = None
        return (add_8,)