[pnnx]:torch.clamp_min convert failed

[deepage]

error log | 日志或报错信息 | ログ

model | 模型 | モデル

1. 定义模型

import torch.nn as nn
import torch.nn.functional as F

class Model(nn.Module):
    def __init__(self):
        super(Model, self).__init__()

    def forward(self, x):
        x = torch.clamp_min(x, min=0)
        return  x

model = Model()
model.eval()

input = torch.rand(1, 81, 512)
trace = torch.jit.trace(model, input)
trace.save('clamp_min.torchscript')

how to reproduce | 复现步骤 | 再現方法

1. pnnx clamp_min.torchscript inputshape=[1,81,512]
2. 转换时报警信息如下：

fallback batch axis 233 for operand 1
fallback batch axis 233 for operand 2
ignore pnnx.Expression pnnx_expr_0 param expr=0

3. clamp_min_pnnx.py中前向代码如下：

        v_1 = 0
        v_2 = aten::clamp_min(v_0, v_1)
        return v_2

可以通过修改v_2 = torch.clamp_min(v_0, v_1)得到正确的结果。