diff options
| author | Jared Van Bortel <jared@nomic.ai> | 2024-02-22 17:05:23 -0500 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-02-22 17:05:23 -0500 |
| commit | 15499eb94227401bdc8875da6eb85c15d37068f7 (patch) | |
| tree | 304ceb65978864af454d9b5ad7ad08fc4f673326 /convert-hf-to-gguf.py | |
| parent | 96633eeca1265ed03e57230de54032041c58f9cd (diff) | |
mpt : do not duplicate token_embd.weight on disk (#5670)
Diffstat (limited to 'convert-hf-to-gguf.py')
| -rwxr-xr-x | convert-hf-to-gguf.py | 5 |
1 files changed, 0 insertions, 5 deletions
diff --git a/convert-hf-to-gguf.py b/convert-hf-to-gguf.py index 481198da..9bdfce07 100755 --- a/convert-hf-to-gguf.py +++ b/convert-hf-to-gguf.py @@ -622,11 +622,6 @@ class MPTModel(Model): 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): |