[fix] Normalization compute step for non-nvidia hardware

[mikepapadim]

No description provided.

[Copilot]

Pull request overview

This PR adds a conditional finalization step for RMS normalization operations to support non-NVIDIA hardware. The fix ensures proper normalization computation across different GPU vendors by adding a hardware-specific final reduction step when needed.

• Adds conditional reductionFinalNormalization task after initial RMS reduction
• Uses shouldUseFinalNormalization() to determine when the additional step is required
• Applies the fix consistently across multiple model architectures and quantization formats

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
Qwen3Q8_0FFNLayers.java	Adds conditional finalization task for attention RMS normalization in Q8_0 quantized Qwen3 model
Phi3Q8_0FFNLayers.java	Adds conditional finalization task for attention RMS normalization in Q8_0 quantized Phi3 model
Qwen3FP16FFNLayers.java	Adds conditional finalization task for attention RMS normalization in FP16 Qwen3 model
Qwen2FP16FFNLayers.java	Adds conditional finalization task for attention RMS normalization in FP16 Qwen2 model
Phi3FP16FFNLayers.java	Adds conditional finalization task for attention RMS normalization in FP16 Phi3 model

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.