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

remove blank lines to make lint happy

This commit is contained in:
Yee Man Chan 2026-01-11 22:58:44 +08:00
parent 4f6ef2c085
commit 719d374bf6
1 changed files with 0 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
convert_hf_to_gguf.py
View File

@ -5088,7 +5088,6 @@ class KimiLinearModel(TextModel):
merged = QwenModel.bpe(mergeable_ranks, token, max_rank=rank) merged = QwenModel.bpe(mergeable_ranks, token, max_rank=rank)
if len(merged) == 2: if len(merged) == 2:
merges.append(' '.join(map(QwenModel.token_bytes_to_string, merged))) merges.append(' '.join(map(QwenModel.token_bytes_to_string, merged)))
# Build token list # Build token list
vocab_size = self.hparams["vocab_size"] vocab_size = self.hparams["vocab_size"]
special_tokens = tokenizer.special_tokens special_tokens = tokenizer.special_tokens
@ -5316,14 +5315,11 @@ class KimiLinearModel(TextModel):
if name.endswith("kv_b_proj.weight"): if name.endswith("kv_b_proj.weight"):
name_kb = name.replace("kv_b_proj", "k_b_proj") name_kb = name.replace("kv_b_proj", "k_b_proj")
name_vb = name.replace("kv_b_proj", "v_b_proj") name_vb = name.replace("kv_b_proj", "v_b_proj")
n_head_kv = self.hparams["num_key_value_heads"] n_head_kv = self.hparams["num_key_value_heads"]
v_head_dim = self.hparams["v_head_dim"] v_head_dim = self.hparams["v_head_dim"]
qk_nope_head_dim = self.hparams["qk_nope_head_dim"] qk_nope_head_dim = self.hparams["qk_nope_head_dim"]
logger.info("Split kv_b n_head_kv %d\n" % n_head_kv) 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) assert data_torch.shape[0] == n_head_kv * (v_head_dim + qk_nope_head_dim)
kv_b = data_torch.view(n_head_kv, v_head_dim + qk_nope_head_dim, data_torch.shape[-1]) kv_b = data_torch.view(n_head_kv, v_head_dim + qk_nope_head_dim, data_torch.shape[-1])
k_b, v_b = torch.split(kv_b, [qk_nope_head_dim, v_head_dim], dim=1) k_b, v_b = torch.split(kv_b, [qk_nope_head_dim, v_head_dim], dim=1)
k_b = k_b.transpose(1, 2) k_b = k_b.transpose(1, 2)