[Migration] Migrate GGUF quantization support to plugin (#39612)

Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
This commit is contained in:
Isotr0py
2026-06-12 12:02:21 -07:00
committed by GitHub
parent d6fd7ce8da
commit 6635279d8a
57 changed files with 72 additions and 9048 deletions
-2
View File
@@ -168,8 +168,6 @@ class Exaone4Attention(nn.Module):
self.k_norm = RMSNorm(self.head_dim, eps=config.rms_norm_eps)
is_neox_style = True
if quant_config is not None and quant_config.get_name() == "gguf":
is_neox_style = False
layer_idx = extract_layer_index(prefix)
is_sliding = config.layer_types[layer_idx] == "sliding_attention"