Revert "Revert "[FEAT] Support GGUF format"" (#2287)

2024-11-30 22:14:48 -08:00
parent 1bfa511b95
commit 4936be8acc
41 changed files with 229 additions and 132 deletions
@@ -966,7 +966,7 @@ class MllamaForConditionalGeneration(nn.Module):
            skip_cross_attention=skip_cross_attention,
        )
        return self.logits_processor(
-            input_ids, hidden_states, self.language_model.lm_head.weight, forward_batch
+            input_ids, hidden_states, self.language_model.lm_head, forward_batch
        )

    def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):