apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-27 02:20:31 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Feature] support pool (#3827)

Browse Source 
* support pool

* update pooling

* add pooler_config and check

* update

* support AutoWeightsLoader load weight

* fix

* update

* delete print

* update pre-commit

* fix

* fix xpu

* fix ModelRegistry->model_registry

* fix Copilot review

* fix pooler.py

* delete StepPooler

* fix abstract

* fix default_loader_v1

* fix Pre Commit

* support torch qwen3 dense

* add test and fix torch-qwen

* fix

* fix

* adapter ci:

* fix review

* fix pooling_params.py

* fix

* fix tasks.py 2025

* fix print and logger

* Modefy ModelRegistry and delete AutoWeightsLoader

* fix logger

* fix test_embedding

* fix ci bug

* ernie4_5 model_registry

* fix test

* support Qwen3-Embedding-0.6B tp=1 load

* fix extra code

* fix

* delete fix vocab_size

* delete prepare_params_dict

* fix:
This commit is contained in:
lizexu123
2025-09-22 14:09:09 +08:00
committed by GitHub
parent da74a5f0b3
commit c86945ef49
36 changed files with 2371 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

24
fastdeploy/model_executor/models/ernie4_5_moe.py
View File

@@ -45,7 +45,11 @@ from fastdeploy.model_executor.layers.linear import (
from fastdeploy.model_executor.layers.lm_head import ParallelLMHead
from fastdeploy.model_executor.layers.moe.moe import FusedMoE
from fastdeploy.model_executor.layers.normalization import RMSNorm
from fastdeploy.model_executor.models.model_base import ModelForCasualLM
from fastdeploy.model_executor.models.model_base import (
ModelCategory,
ModelForCasualLM,
ModelRegistry,
)
from fastdeploy.model_executor.models.tp_utils import TensorSplitMode as tsm
from fastdeploy.model_executor.models.utils import LayerIdPlaceholder as layerid
from fastdeploy.model_executor.models.utils import WeightMeta
@@ -471,6 +475,12 @@ class Ernie4_5_Model(nn.Layer):
return out
@ModelRegistry.register_model_class(
architecture="Ernie4_5_MoeForCausalLM",
module_path="ernie4_5_moe",
category=ModelCategory.TEXT_GENERATION,
primary_use=ModelCategory.TEXT_GENERATION,
)
class Ernie4_5_MoeForCausalLM(ModelForCasualLM):
"""
Ernie4_5_MoeForCausalLM
@@ -646,6 +656,12 @@ class Ernie4_5_MoeForCausalLM(ModelForCasualLM):
self.ernie.clear_grpah_opt_backend(fd_config=self.fd_config)
@ModelRegistry.register_model_class(
architecture="Ernie4_5_ForCausalLM",
module_path="ernie4_5_moe",
category=ModelCategory.TEXT_GENERATION,
primary_use=ModelCategory.TEXT_GENERATION,
)
class Ernie4_5_ForCausalLM(Ernie4_5_MoeForCausalLM):
"""
Ernie4_5_ForCausalLM
@@ -659,6 +675,12 @@ class Ernie4_5_ForCausalLM(Ernie4_5_MoeForCausalLM):
return "Ernie4_5_ForCausalLM"
@ModelRegistry.register_model_class(
architecture="Ernie4_5ForCausalLM",
module_path="ernie4_5_moe",
category=ModelCategory.TEXT_GENERATION,
primary_use=ModelCategory.TEXT_GENERATION,
)
class Ernie4_5ForCausalLM(Ernie4_5_ForCausalLM):
"""
Ernie4_5ForCausalLM 0.3B-PT