Do not duplicate embedding weights for output.weight

Author: mitmul

convert_hf_to_gguf.py

@@ -3594,18 +3594,7 @@ def set_gguf_parameters(self):
     def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
         del bid  # unused
 
-        if name.endswith(".embed_tokens.weight"):
-            # If there is no lm_head, we need to map the token embedding to the output layer
-            assert self.tensor_names is not None
-            if all(['lm_head' not in name for name in self.tensor_names]):
-                name_base = name.replace(".embed_tokens.weight", "")
-                output_name = "lm_head"
-
-                embed_tokens_mapped = self.map_tensor_name(name)
-                output_mapped = self.map_tensor_name(output_name) + ".weight"
-
-                return [(embed_tokens_mapped, data_torch), (output_mapped, data_torch)]
-        elif name.endswith(".A_log"):
+        if name.endswith(".A_log"):
             data_torch = -torch.exp(data_torch)
         elif name.endswith(".dt_bias"):
             name = name.rpartition(".dt_bias")[0] + ".dt_proj.bias"

src/llama-model.cpp

@@ -2865,7 +2865,11 @@ bool llama_model::load_tensors(llama_model_loader & ml) {
 
                     // output
                     output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
-                    output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, 0);
+                    output      = create_tensor(tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED);
+                    // if output is NULL, init from the input tok embed
+                    if (output == NULL) {
+                        output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED);
+                    }
 
                     for (int i = 0; i < n_layer; ++i) {
                         auto & layer = layers[i];