[Quant][Kernel] Remove GPTQ Group/Dynamic Activation Ordering (#54809)
Signed-off-by: Roderick-Wu <roderickwu2003@gmail.com> Signed-off-by: Roderick Wu <roderickwu2003@gmail.com> Signed-off-by: mgoin <mgoin64@gmail.com> Co-authored-by: mgoin <mgoin64@gmail.com>
This commit is contained in:
@@ -463,8 +463,7 @@ class VocabParallelEmbedding(PluggableLayer):
|
||||
output_dim = getattr(param, "output_dim", None)
|
||||
packed_dim = getattr(param, "packed_dim", None)
|
||||
|
||||
# If parameter does not have output dim, then it should
|
||||
# be copied onto all gpus (e.g. g_idx for act_order gptq).
|
||||
# Parameters without an output dimension are copied onto all GPUs.
|
||||
if output_dim is None:
|
||||
if (
|
||||
loaded_weight.ndim == 0
|
||||
|
||||
Reference in New Issue
Block a user