[ROCm][CI] Add "Qwen3-Next-80B-A3B-Instruct MTP Async EPLB Accuracy Test" Back Into AMD CI (#30590)

Signed-off-by: David Chen <530634352@qq.com>
Signed-off-by: WeiQing Chen <40507679+david6666666@users.noreply.github.com>
Signed-off-by: Micah Williamson <micah.williamson@amd.com>
Co-authored-by: WeiQing Chen <40507679+david6666666@users.noreply.github.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
This commit is contained in:
Micah Williamson
2025-12-14 06:56:26 +00:00
committed by GitHub
co-authored by WeiQing Chen Cyrus Leung
parent 29f7d97715
commit 97f2f160fd
3 changed files with 74 additions and 4 deletions
@@ -322,9 +322,6 @@ async def transfer_layer(
num_local_physical_experts = next(iter(expert_weights[0])).shape[0]
assert new_global_expert_indices.shape == (num_moe_layers, num_physical_experts)
assert num_physical_experts == ep_size * num_local_physical_experts
# A buffer to hold the expert weights in one layer during the exchange.
# NOTE: Currently we assume the same weights across different layers
# have the same shape.
is_unchanged, is_received_locally, experts_recv_loc = move_to_buffer(
num_local_experts=num_local_physical_experts,