diff --git a/server/text_generation_server/models/custom_modeling/mamba_modeling.py b/server/text_generation_server/models/custom_modeling/mamba_modeling.py index 07284e6a..5a9c0588 100644 --- a/server/text_generation_server/models/custom_modeling/mamba_modeling.py +++ b/server/text_generation_server/models/custom_modeling/mamba_modeling.py @@ -212,7 +212,7 @@ class MambaModel(nn.Module): try: self.lm_head = SpeculativeHead.load(config, f"{prefix}.embeddings", weights) except RuntimeError: - self.lm_head = SpeculativeHead.load(config, f"{prefix}.embeddings", weights) + self.lm_head = SpeculativeHead.load(config, f"{prefix}.embedding", weights) self.config = config def forward(