add fedsp

Author: xieyxclack

LICENSE

@@ -661,3 +661,46 @@ distributed under the License is distributed on an "AS IS" BASIS,
 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 See the License for the specific language governing permissions and
 limitations under the License.
+
+--------------------------------------------------------------------------------
+
+The code in federatedscope/nlp/fedsp/dataset and federatedscope/nlp/fedsp/metric
+is partially adapted from https://github.com/mit-han-lab/offsite-tuning (MIT License)
+
+Copyright (c) 2023 MIT HAN Lab
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+---------------------------------------------------------------------------------
+The code in federatedscope/nlp/fedsp/model and federatedscope/nlp/fedsp/worker/server.py
+is partially adapted from  https://github.com/THUDM/P-tuning-v2  (Apache License)
+
+Copyright 2021 Xiao Liu
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+   http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.

federatedscope/core/auxiliaries/metric_builder.py

@@ -5,6 +5,7 @@
 
 try:
     from federatedscope.contrib.metrics import *
+    from federatedscope.nlp.hetero_tasks.metric import *
 except ImportError as error:
     logger.warning(
         f'{error} in `federatedscope.contrib.metrics`, some modules are not '

federatedscope/core/auxiliaries/worker_builder.py

@@ -72,10 +72,6 @@ def get_client_cls(cfg):
         from federatedscope.nlp.hetero_tasks.worker import ATCClient
         return ATCClient
 
-    if cfg.data.type.lower() == 'hetero_nlp_tasks':
-        from federatedscope.nlp.hetero_tasks.worker import ATCClient
-        return ATCClient
-
     if cfg.trainer.type.lower() == 'fedsp_trainer':
         from federatedscope.nlp.fedsp.worker import FedSPClient
         return FedSPClient

federatedscope/core/configs/cfg_aggregator.py

@@ -1,9 +1,11 @@
 import logging
+
 from federatedscope.core.configs.config import CN
 from federatedscope.register import register_config
 
 
 def extend_aggregator_cfg(cfg):
+
     # ---------------------------------------------------------------------- #
     # aggregator related options
     # fs has supported the robust aggregation rules of 'krum', 'median',
@@ -26,6 +28,7 @@ def extend_aggregator_cfg(cfg):
 
 
 def assert_aggregator_cfg(cfg):
+
     if cfg.aggregator.byzantine_node_num == 0 and \
             cfg.aggregator.robust_rule in \
             ['krum', 'normbounding', 'median', 'trimmedmean', 'bulyan']:

federatedscope/core/configs/cfg_data.py

@@ -100,7 +100,7 @@ def extend_data_cfg(cfg):
     cfg.feat_engr.secure.dp = CN()  # under dev
 
     # prompt tuning
-    cfg.data.dataset_name = ''
+    cfg.data.dataset_name = '' #TODO
     cfg.data.train_frac = 0.9
     cfg.data.num_train_per_client = -1
 

federatedscope/core/configs/cfg_fl_setting.py

@@ -1,7 +1,8 @@
 import logging
-import torch
+
 from federatedscope.core.configs.config import CN
 from federatedscope.register import register_config
+import torch
 
 logger = logging.getLogger(__name__)
 

federatedscope/nlp/fedsp/README.md

@@ -1,8 +1,8 @@
 ## Tunable Soft Prompts are Messengers in Federated Learning
 The implementation of *Tunable Soft Prompts are Messengers in Federated Learning*.
 
-In this study, we propose a novel FL training approach that accomplishes information exchange among participants via tunable soft prompts.
-These soft prompts are updated and transmitted between the server and clients, taking over the duty of the global model parameters and serving as messengers to deliver useful knowledge in the local data and global model.
+In this study, we propose a novel FL training approach, named FedSP, that accomplishes information exchange among participants via tunable soft prompts. 
+These soft prompts, updated and transmitted between the server and clients, assume the role of the global model parameters and serve as messengers to deliver useful knowledge from the local data and global model.
 
 ### Installation
 First of all, you need to install FederatedScope, please refer to [installation](https://github.com/alibaba/FederatedScope#step-1-installation).
@@ -25,19 +25,7 @@ bash run_gpt_prefix.sh $DEVICE  # gpt2-xl
 bash run_opt_prefix.sh $DEVICE  # opt-1.3b
 ```
 
-**FedPrompt**
-```bash
-bash run_gpt_fedprompt.sh $DEVICE  # gpt2-xl
-bash run_opt_fedprompt.sh $DEVICE  # opt-1.3b
-```
-
-**FedPrompt-LSR**
-```bash
-bash run_gpt_fedprompt_lsr.sh $DEVICE  # gpt2-xl
-bash run_opt_fedprompt_lsr.sh $DEVICE  # opt-1.3b
-```
-
-**FedSP (ours)**
+**FedSP**
 ```bash
 bash run_gpt_fedsp.sh $DEVICE  # gpt2-xl
 bash run_opt_fedsp.sh $DEVICE  # opt-1.3b