Init the main branch for aiter

85762c1a · Xiaowei.zhang · ae0b3521 · 85762c1a · 85762c1a · 85762c1a
Commit 85762c1a authored Apr 15, 2026 by Xiaowei.zhang
20 changed files
--- a/aiter/ops/triton/moe_routing_sigmoid_top1_fused.py
+++ b/aiter/ops/triton/moe_routing_sigmoid_top1_fused.py
--- a/aiter/ops/triton/norm.py
+++ b/aiter/ops/triton/norm.py
--- a/aiter/ops/triton/pa_decode.py
+++ b/aiter/ops/triton/pa_decode.py
--- a/aiter/ops/triton/pa_prefill.py
+++ b/aiter/ops/triton/pa_prefill.py
--- a/aiter/ops/triton/pod_attention.py
+++ b/aiter/ops/triton/pod_attention.py
--- a/aiter/ops/triton/prefill_attention.py
+++ b/aiter/ops/triton/prefill_attention.py
--- a/aiter/ops/triton/quant.py
+++ b/aiter/ops/triton/quant.py
--- a/aiter/ops/triton/rmsnorm.py
+++ b/aiter/ops/triton/rmsnorm.py
--- a/aiter/ops/triton/rope.py
+++ b/aiter/ops/triton/rope.py
--- a/aiter/ops/triton/routing.py
+++ b/aiter/ops/triton/routing.py
--- a/aiter/ops/triton/sage_attention.py
+++ b/aiter/ops/triton/sage_attention.py
--- a/aiter/ops/triton/sage_attention_qk_int8_per_block.py
+++ b/aiter/ops/triton/sage_attention_qk_int8_per_block.py
--- a/aiter/ops/triton/sage_attention_qk_int8_per_block_causal.py
+++ b/aiter/ops/triton/sage_attention_qk_int8_per_block_causal.py
--- a/aiter/ops/triton/sage_attention_quant_per_block.py
+++ b/aiter/ops/triton/sage_attention_quant_per_block.py
--- a/aiter/ops/triton/softmax.py
+++ b/aiter/ops/triton/softmax.py
--- a/aiter/ops/triton/topk.py
+++ b/aiter/ops/triton/topk.py
--- a/aiter/ops/triton/triton_decode_attention.py
+++ b/aiter/ops/triton/triton_decode_attention.py
--- a/aiter/ops/triton/unified_attention.py
+++ b/aiter/ops/triton/unified_attention.py
--- a/aiter/ops/triton/utils/__init__.py
+++ b/aiter/ops/triton/utils/__init__.py
--- a/aiter/ops/triton/utils/arch_info.py
+++ b/aiter/ops/triton/utils/arch_info.py