who knows i just don't want to lose anything

gnpinkert · Sep 16, 2024 · 5874d30 · 5874d30
1 parent 5ce45eb
commit 5874d30
Showing 1 changed file with 4 additions and 0 deletions.
diff --git a/vllm/model_executor/layers/fused_moe/fused_moe.py b/vllm/model_executor/layers/fused_moe/fused_moe.py
@@ -550,6 +550,10 @@ def fused_experts(hidden_states: torch.Tensor,
         sorted_token_ids, expert_ids, num_tokens_post_padded = (
             moe_align_block_size(curr_topk_ids, config['BLOCK_SIZE_M'], E))
 
+        print(f"Top_k IDs: {curr_topk_ids}")
+        print(f"sorted_token_ids IDs: {sorted_token_ids}")
+        print(f"expert ids: {expert_ids}")
+
         invoke_fused_moe_kernel(curr_hidden_states,
                                 w1,
                                 intermediate_cache1,