CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-27 08:21:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

convert : correctly handle LLaMA tokenizer for Jamba (#16470)

Browse Source 
* fix: convert_hf_to_gguf - change Jamba non-sentencepiece mode (tokenizer.json) vocab construction

* fix: convert_hf_to_gguf - jamba non-sentencepiece tokenizer to use _set_vocab_llama_hf func

* fix: convert_hf_to_gguf - removed get_vocab_base_pre from jamba
This commit is contained in:
amirai21
2025-10-11 11:33:41 +03:00
committed by GitHub
parent e60f01d941
commit 477a66b035
1 changed files with 2 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
convert_hf_to_gguf.py
View File

@@ -5966,20 +5966,12 @@ class Mamba2Model(TextModel):
class JambaModel(TextModel):
model_arch = gguf.MODEL_ARCH.JAMBA
def get_vocab_base_pre(self, tokenizer) -> str:
del tokenizer # unused
return "gpt-2"
def set_vocab(self):
if (self.dir_model / "tokenizer.model").is_file():
# Using Jamba's tokenizer.json causes errors on model load
# (something about "byte not found in vocab"),
# but there's a working tokenizer.model
self._set_vocab_sentencepiece()
else:
# Some Jamba models only have a tokenizer.json, which works.
self._set_vocab_gpt2()
self._set_vocab_llama_hf()
self.gguf_writer.add_add_space_prefix(False)
def set_gguf_parameters(self):
d_model = self.find_hparam(["hidden_size", "mamba_d_model"])