pppeng 0f289fa2a8 Add patch_qwen3_5 for triton ops fused_recurrent_gated_delta_rule (#7109) ...

### What this PR does / why we need it?

The ops `torch_npu.npu_recurrent_gated_delta_rule` currently does not
support `ssm_state` inputs in float32 format,
we temporarily retain the _forward_core implementation with triton for
Qwen3_5

---------

Signed-off-by: pppeng <zepengliu912@qq.com>
Signed-off-by: pppeng <60355449+ppppeng@users.noreply.github.com>

2026-03-10 23:28:58 +08:00

..

platform

[KVCache]Qwen3.5 supports contiguous tensor hybrid-attn kv-cache (#6887)

2026-03-09 15:28:40 +08:00

worker

Add patch_qwen3_5 for triton ops fused_recurrent_gated_delta_rule (#7109)

2026-03-10 23:28:58 +08:00

__init__.py

Add patch_qwen3_5 for triton ops fused_recurrent_gated_delta_rule (#7109)

2026-03-10 23:28:58 +08:00