Skip to content

[v0.21.0] Fix accuracy issue in minimax_m2 with TP > 1 #1506

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
skavulya wants to merge 2 commits into from
+1 −4
Closed

[v0.21.0] Fix accuracy issue in minimax_m2 with TP > 1 #1506

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b09e8d0
Fix accuracy issue in minimax_m2 with TP > 1
skavulya May 16, 2026
13fd528
Fix lint format issues
skavulya May 27, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 1 addition & 4 deletions vllm_gaudi/models/minimax_m2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,7 +33,7 @@
from vllm.model_executor.layers.attention import Attention
from vllm.compilation.decorators import support_torch_compile
from vllm.config import CacheConfig, ModelConfig, VllmConfig
from vllm.distributed import (get_pp_group, get_tensor_model_parallel_world_size, tensor_model_parallel_all_reduce)
from vllm.distributed import (get_pp_group, get_tensor_model_parallel_world_size)
from vllm.model_executor.layers.fused_moe import FusedMoE
from vllm.model_executor.layers.mamba.linear_attn import MiniMaxText01RMSNormTP
from vllm.model_executor.layers.layernorm import RMSNorm
Expand Down Expand Up @@ -107,9 +107,6 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
# router_logits: (bs * seq_len, n_experts)
router_logits, _ = self.gate(hidden_states.to(torch.float32))
final_hidden_states = self.experts(hidden_states=hidden_states, router_logits=router_logits)
final_hidden_states = final_hidden_states
if self.tp_size > 1:
final_hidden_states = tensor_model_parallel_all_reduce(final_hidden_states)

return final_hidden_states.view(bs, seq_len, hidden_dim)

Expand Down