monkeypatch.models.qwen3.fused_attn
monkeypatch.models.qwen3.fused_attn
Fuse q_norm/k_norm + RoPE in Qwen3Attention.forward via one Triton kernel.
monkeypatch.models.qwen3.fused_attn
Fuse q_norm/k_norm + RoPE in Qwen3Attention.forward via one Triton kernel.