[Migration] Migrate GGUF quantization support to plugin (#39612)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
This commit is contained in:
@@ -238,9 +238,6 @@ class Llama4Attention(nn.Module):
|
||||
prefix=f"{prefix}.o_proj",
|
||||
)
|
||||
is_neox_style = True
|
||||
is_gguf = quant_config and quant_config.get_name() == "gguf"
|
||||
if is_gguf and config.model_type == "llama":
|
||||
is_neox_style = False
|
||||
|
||||
self.rotary_emb = (
|
||||
get_rope(
|
||||
|
||||
Reference in New Issue
Block a user