mpt : do not duplicate token_embd.weight on disk (#5670)

ggerganov · Feb 22, 2024 · 15499eb · 15499eb
1 parent 96633ee
commit 15499eb
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 7 deletions.
diff --git a/convert-hf-to-gguf.py b/convert-hf-to-gguf.py
@@ -622,11 +622,6 @@ def write_tensors(self):
 
             self.gguf_writer.add_tensor(new_name, data)
 
-            # note: MPT output is tied to (same as) wte in original model;
-            # for easier implementation in llama.cpp it's duplicated in GGUF, though :/
-            if new_name == "token_embd.weight":
-                self.gguf_writer.add_tensor("output.weight", data)
-
 
 class OrionModel(Model):
     def set_vocab(self):

diff --git a/llama.cpp b/llama.cpp
@@ -509,7 +509,6 @@ static std::map<llm_arch, std::map<llm_tensor, std::string>> LLM_TENSOR_NAMES =
         {
             { LLM_TENSOR_TOKEN_EMBD,      "token_embd" },
             { LLM_TENSOR_OUTPUT_NORM,     "output_norm" },
-            { LLM_TENSOR_OUTPUT,          "output" },
             { LLM_TENSOR_ATTN_NORM,       "blk.%d.attn_norm" },
             { LLM_TENSOR_FFN_NORM,        "blk.%d.ffn_norm" },
             { LLM_TENSOR_ATTN_QKV,        "blk.%d.attn_qkv" },
@@ -4056,7 +4055,10 @@ static bool llm_load_tensors(
                         model.output_norm   = ml.create_tensor(ctx_output,       tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd});
                         model.output_norm_b = ml.create_tensor(ctx_output,       tn(LLM_TENSOR_OUTPUT_NORM, "bias"),   {n_embd}, false);
 
-                        model.output        = ml.create_tensor(ctx_output_split, tn(LLM_TENSOR_OUTPUT,      "weight"), {n_embd, n_vocab});
+                        // same as tok_embd, duplicated to allow offloading
+                        model.output        = ml.create_tensor(ctx_output_split, tn(LLM_TENSOR_TOKEN_EMBD,  "weight"), {n_embd, n_vocab});
+                        ml.n_created--; // artificial tensor
+                        ml.size_data += ggml_nbytes(model.output);
                     }
 
                     for (int i = 0; i < n_layer; ++i) {