apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-05 08:37:06 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

qwen loader (#3057)

Browse Source
This commit is contained in:
bukejiyu
2025-07-30 19:09:38 +08:00
committed by GitHub
parent 28fff1b035
commit db698bda01
22 changed files with 494 additions and 92 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
fastdeploy/model_executor/layers/embeddings.py
View File

@@ -22,6 +22,7 @@ from paddle import nn
from paddle.distributed import fleet
from fastdeploy.config import FDConfig
from fastdeploy.model_executor.models.utils import set_weight_attrs
from .utils import get_tensor
@@ -80,6 +81,7 @@ class VocabParallelEmbedding(nn.Layer):
initializer=nn.initializer.Normal(mean=0.0, std=self.initializer_range),
),
)
set_weight_attrs(self.embeddings.weight, {"output_dim": False})
else:
# column cut embedding
self.embeddings = nn.Embedding(
@@ -89,6 +91,7 @@ class VocabParallelEmbedding(nn.Layer):
self.embeddings.weight.is_distributed = True
self.embeddings.weight.split_axis = 1
set_weight_attrs(self.embeddings.weight, {"output_dim": True})
self.prefix = prefix
self.dropout = nn.Dropout(self.hidden_dropout_prob)