Fix TFDWConv() c1 == c2 check

[glenn-jocher]

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Enhancement of Depthwise Convolution Layer in TensorFlow Models.

📊 Key Changes

• Removed the g parameter (number of groups) from TFDWConv layer initializer arguments.
• Modified the assertion check to only require input channels (c1) to be equal to the output channels (c2), instead of also checking against the g parameter.

🎯 Purpose & Impact

• 🧹 Simplifies the initialization of the depthwise convolution layer by removing an unnecessary parameter.
• 🛠️ Ensures the integrity of the model by checking that the number of input channels matches the number of output channels, a requirement for depthwise convolution.
• 🔧 Reduces potential for configuration errors by removing the group check, which is not needed in depthwise convolutions as the input and output channels are naturally equal.
• 👩‍💻 Helps developers to work with a cleaner interface when utilizing the TFDWConv layer in their machine learning models, potentially increasing development productivity and model stability.