happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

minor changes

This commit is contained in:
Yee Man Chan 2026-02-03 19:25:33 +08:00
parent 000fded1ea
commit 8ec5b08409
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
convert_hf_to_gguf.py
View File

@ -5189,7 +5189,7 @@ class KimiLinearModel(TextModel):
# process the experts separately
if name.find("block_sparse_moe.experts") != -1:
n_experts = self.find_hparam(["num_experts"], optional=False)
n_experts = self.find_hparam(["num_local_experts", "num_experts"], optional=False)
assert bid is not None
if self._experts is None:
@ -5218,7 +5218,7 @@ class KimiLinearModel(TextModel):
name_kb = name.replace("kv_b_proj", "k_b_proj")
name_vb = name.replace("kv_b_proj", "v_b_proj")
n_head_kv = self.hparams["num_key_value_heads"]
v_head_dim = self.hparams["v_head_dim"]
v_head_dim = self.find_hparam(["n_embd_head_v_mla", "v_head_dim"], optional=False)
qk_nope_head_dim = self.hparams["qk_nope_head_dim"]
logger.info("Split kv_b n_head_kv %d\n" % n_head_kv)
assert data_torch.shape[0] == n_head_kv * (v_head_dim + qk_nope_head_dim)