[Bugfix] Missing quant_config in deepseek embedding layer (#12836)

2025-02-07 06:35:09 +01:00
parent 433c4a4923
commit aa375dca9f
1 changed files with 2 additions and 1 deletions
--- a/vllm/model_executor/models/deepseek_v2.py
+++ b/vllm/model_executor/models/deepseek_v2.py
@ -581,7 +581,8 @@ class DeepseekV2Model(nn.Module):
            self.embed_tokens = VocabParallelEmbedding(
                config.vocab_size,
                config.hidden_size,
-            )
+                quant_config=quant_config,
+                prefix=f"{prefix}.embed_tokens")
        else:
            self.embed_tokens = PPMissingLayer()