【MambaSR复现】【Mamba环境配置】No Module named “causal_conv1d_cuda”、“selective_scan_cuda”报错解决

【注意：mamba就不是给windows玩的，linux下环境配置很方便，少有报错，基本都不是啥大问题，大伙别硬在windows上浪费时间了，装个Ubuntu复现起来很快的】

mamba_ssm和causal_conv1d模块的配置方法见博文：

【Windows系统下Mamba环境配置】triton 、causal conv1d和mamba_ssm模块配置保姆教程

继上述成功安装mamba_ssm和causal-conv1d模块后，运行环境发现此两模块仍有如下报错：

File "E:\install\Anaconda\envs\SR\Lib\site-packages\mamba_ssm\ops\selective_scan_interface.py", line 10, in <module>
    import causal_conv1d_cuda
ModuleNotFoundError: No module named 'causal_conv1d_cuda'

至今，未搞清楚此报错原因。但通过以下方式初步解决了这两个报错问题！

1.找不到“causal_conv1d_cuda”

（1）首先，进入conda环境，卸载causal-conv1d模块，保留本地路径下的安装文件用于后续修改和再次安装。
（2）从本地文件夹中找到causal_conv1d_interface.py文件，进入修改

（3）注释掉xx_cuda的模块导入

（4）将其causal_conv1d_fn部分函数代码用以下进行替换（注释部分为替换前源码，未注释部分为替换后）

# def causal_conv1d_fn(x, weight, bias=None, activation=None):
#     """
#     x: (batch, dim, seqlen)
#     weight: (dim, width)
#     bias: (dim,)
#     activation: either None or "silu" or "swish"
#
#     out: (batch, dim, seqlen)
#     """
#     return CausalConv1dFn.apply(x, weight, bias, activation)

def causal_conv1d_fn(x,weight,bias=None,seq_idx=None,initial_states=None,return_final_states=False,final_states_out=None,activation=None,):
    """
    x: (batch, dim, seqlen)
    weight: (dim, width)
    bias: (dim,)
    seq_idx: (batch, seqlen)
    initial_states: (batch, dim, width - 1)
    final_states_out: (batch, dim, width - 1), to be written to
    activation: either None or "silu" or "swish"
    out: (batch, dim, seqlen)
    """
    #### 修改在这里！！！！！
    return causal_conv1d_ref(x,weight,bias,seq_idx,initial_states,return_final_states,final_states_out,activation,)

（5）按照此前模块的安装方法，重新安装此模块即可。

2.selective_scan_cuda

（1）首先，在pytorch环境下卸载mamba_ssm模块（上次安装模块时候的文件保留，用于后续修改和重新安装mamba_ssm模块）
（2）根据文件存放路径，找到mamba_ssm - > ops - >selective_scan_interface.py文件，打开

（3）将文件模块导入部分，含xx_cuda的此两句注释掉

（4）将 selective_scan_fn函数部分代码做如下修改：（注释部分为修改前，未注释部分为修改后）

# def selective_scan_fn(u, delta, A, B, C, D=None, z=None, delta_bias=None, delta_softplus=False,
#                      return_last_state=False):
#     """if return_last_state is True, returns (out, last_state)
#     last_state has shape (batch, dim, dstate). Note that the gradient of the last state is
#     not considered in the backward pass.
#     """
#     return selective_scan_ref(u, delta, A, B, C, D, z, delta_bias, delta_softplus, return_last_state)

def selective_scan_fn(u, delta, A, B, C, D=None, z=None, delta_bias=None, delta_softplus=False,
                     return_last_state=False):
    """if return_last_state is True, returns (out, last_state)
    last_state has shape (batch, dim, dstate). Note that the gradient of the last state is
    not considered in the backward pass.
    """
    return selective_scan_ref(u, delta, A, B, C, D, z, delta_bias, delta_softplus, return_last_state)

同理，将mamba_inner_fn也做修改替换

# def mamba_inner_fn(
#     xz, conv1d_weight, conv1d_bias, x_proj_weight, delta_proj_weight,
#     out_proj_weight, out_proj_bias,
#     A, B=None, C=None, D=None, delta_bias=None, B_proj_bias=None,
#     C_proj_bias=None, delta_softplus=True):
#     return mamba_inner_ref(xz, conv1d_weight, conv1d_bias, x_proj_weight, delta_proj_weight,
#                               out_proj_weight, out_proj_bias,
#                               A, B, C, D, delta_bias, B_proj_bias, C_proj_bias, delta_softplus)

def mamba_inner_fn(
    xz, conv1d_weight, conv1d_bias, x_proj_weight, delta_proj_weight,
    out_proj_weight, out_proj_bias,
    A, B=None, C=None, D=None, delta_bias=None, B_proj_bias=None,
    C_proj_bias=None, delta_softplus=True):

    return mamba_inner_ref(xz, conv1d_weight, conv1d_bias, x_proj_weight, delta_proj_weight,
                              out_proj_weight, out_proj_bias,
                              A, B, C, D, delta_bias, B_proj_bias, C_proj_bias, delta_softplus)

（5）修改完成后，按照前一篇博文mamba_ssm的安装方法重新安装模块即可

3.模块验证

上述模块修改和重新安装完成后，运行代码会出现如下警告：

No Module named “causal_conv1d_cuda”、“selective_scan_cuda”报错问题初步解决，报错原因尚未弄清，有大佬请评论区指点！
后续弄清再进行更新！（给个三连吧！）
至于上图中的警告，后续继续更新！