dssm train

recommend-model-produce/src/main/python/tools/static_ps_infer_v3.py

@@ -0,0 +1,362 @@
+# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
+#
+# 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.
+
+from __future__ import print_function
+import os
+os.environ['FLAGS_enable_pir_api'] = '0'
+from utils.static_ps.reader_helper_hdfs import get_infer_reader
+from utils.static_ps.program_helper import get_model, get_strategy, set_dump_config
+from utils.static_ps.metric_helper import set_zero, get_global_auc
+from utils.static_ps.common_ps import YamlHelper, is_distributed_env
+import argparse
+import time
+import sys
+import paddle.distributed.fleet as fleet
+import paddle.distributed.fleet.base.role_maker as role_maker
+import paddle
+from paddle.base.executor import FetchHandler
+import queue
+import threading
+
+import warnings
+import logging
+import ast
+import numpy as np
+import struct
+from utils.utils_single import auc
+from utils.oss_client import HangZhouOSSClient
+import utils.compress as compress
+
+__dir__ = os.path.dirname(os.path.abspath(__file__))
+sys.path.append(os.path.abspath(os.path.join(__dir__, '..')))
+
+root_loger = logging.getLogger()
+for handler in root_loger.handlers[:]:
+    root_loger.removeHandler(handler)
+logging.basicConfig(
+    format='%(asctime)s - %(levelname)s - %(message)s', level=logging.INFO)
+logger = logging.getLogger(__name__)
+
+
+import json
+
+class InferenceFetchHandler(FetchHandler):
+    def __init__(self, var_dict, output_file, batch_size=1000):
+        super().__init__(var_dict=var_dict, period_secs=1)
+        self.output_file = output_file
+        self.batch_size = batch_size
+        self.result_queue = queue.Queue()
+        self.writer_thread = threading.Thread(target=self._writer)
+        self.writer_thread.daemon = True  # 设置为守护线程
+        self.writer_thread.start()
+        
+        # 创建输出目录（如果不存在）
+        output_dir = os.path.dirname(output_file)
+        if not os.path.exists(output_dir):
+            os.makedirs(output_dir)
+        # 创建或清空输出文件
+        with open(self.output_file, 'w') as f:
+            f.write('')
+    
+    def handler(self, fetch_vars):
+        """处理每批次的推理结果"""
+        result_dict = {}
+
+        for key in fetch_vars:
+            # 转换数据类型
+            if type(fetch_vars[key]) is np.ndarray:
+                result = fetch_vars[key][0].tolist()
+            else:
+                result = fetch_vars[key]
+            result_dict[key] = result
+        self.result_queue.put(result_dict)  # 将结果放入队列
+    
+    def _writer(self):
+        batch = []
+        while True:
+            try:
+                result_dict = self.result_queue.get(timeout=1)  # 非阻塞获取
+                logger.info("write vector {} {}".format(json.dumps(result_dict), len(batch)))
+                batch.append(result_dict)
+                if len(batch) >= self.batch_size:
+                    logger.info("write vector")
+                    with open(self.output_file, 'a') as f:
+                        for result in batch:
+                            f.write(json.dumps(result) + '\n')
+                    batch = []
+            except queue.Empty:
+                pass
+    
+    def _write_batch(self, batch):
+        with open(self.output_file, 'a') as f:
+            for result in batch:
+                f.write(json.dumps(result) + '\n')
+
+    def flush(self):
+        """确保所有结果都被写入文件"""
+        # 等待队列中剩余的结果被处理
+        self.result_queue.join()
+        # 写入最后一批结果
+        self._write_batch(self.result_queue.queue)
+    
+
+
+def parse_args():
+    parser = argparse.ArgumentParser("PaddleRec train script")
+    parser.add_argument("-o", "--opt", nargs='*', type=str)
+    parser.add_argument(
+        '-m',
+        '--config_yaml',
+        type=str,
+        required=True,
+        help='config file path')
+    parser.add_argument(
+        '-bf16',
+        '--pure_bf16',
+        type=ast.literal_eval,
+        default=False,
+        help="whether use bf16")
+    args = parser.parse_args()
+    args.abs_dir = os.path.dirname(os.path.abspath(args.config_yaml))
+    yaml_helper = YamlHelper()
+    config = yaml_helper.load_yaml(args.config_yaml)
+    # modify config from command
+    if args.opt:
+        for parameter in args.opt:
+            parameter = parameter.strip()
+            key, value = parameter.split("=")
+            if type(config.get(key)) is int:
+                value = int(value)
+            if type(config.get(key)) is float:
+                value = float(value)
+            if type(config.get(key)) is bool:
+                value = (True if value.lower() == "true" else False)
+            config[key] = value
+    config["yaml_path"] = args.config_yaml
+    config["config_abs_dir"] = args.abs_dir
+    config["pure_bf16"] = args.pure_bf16
+    yaml_helper.print_yaml(config)
+    return config
+
+
+def bf16_to_fp32(val):
+    return np.float32(struct.unpack('<f', struct.pack('<I', val << 16))[0])
+
+
+class Main(object):
+    def __init__(self, config):
+        self.metrics = {}
+        self.config = config
+        self.input_data = None
+        self.reader = None
+        self.exe = None
+        self.train_result_dict = {}
+        self.train_result_dict["speed"] = []
+        self.train_result_dict["auc"] = []
+        self.model = None
+        self.pure_bf16 = self.config['pure_bf16']
+
+    def run(self):
+        self.init_fleet_with_gloo()
+        self.network()
+        if fleet.is_server():
+            self.run_server()
+        elif fleet.is_worker():
+            self.run_worker()
+            fleet.stop_worker()
+            self.record_result()
+        logger.info("Run Success, Exit.")
+
+    def init_fleet_with_gloo(self,use_gloo=True):
+        if use_gloo:
+            os.environ["PADDLE_WITH_GLOO"] = "0"
+            role = role_maker.PaddleCloudRoleMaker(
+                is_collective=False,
+                init_gloo=False
+            )
+            fleet.init(role)
+        else:
+            fleet.init()
+
+
+
+    def network(self):
+        self.model = get_model(self.config)
+        self.input_data = self.model.create_feeds(is_infer=True)
+        self.inference_feed_var = self.input_data
+        self.init_reader()
+        self.metrics = self.model.net(self.inference_feed_var,is_infer=True)
+        self.inference_target_var = self.model.inference_target_var
+        logger.info("cpu_num: {}".format(os.getenv("CPU_NUM")))
+        self.model.create_optimizer(get_strategy(self.config),is_infer=True)
+
+    def run_server(self):
+        logger.info("Run Server Begin")
+        fleet.init_server(config.get("runner.warmup_model_path"))
+        fleet.run_server()
+
+    def run_worker(self):
+        logger.info("Run Worker Begin")
+        use_cuda = int(config.get("runner.use_gpu"))
+        use_auc = config.get("runner.infer_use_auc", False)
+        place = paddle.CUDAPlace(0) if use_cuda else paddle.CPUPlace()
+        self.exe = paddle.static.Executor(place)
+
+        with open("./{}_worker_main_program.prototxt".format(
+                fleet.worker_index()), 'w+') as f:
+            f.write(str(paddle.static.default_main_program()))
+        with open("./{}_worker_startup_program.prototxt".format(
+                fleet.worker_index()), 'w+') as f:
+            f.write(str(paddle.static.default_startup_program()))
+
+        self.exe.run(paddle.static.default_startup_program())
+        if self.pure_bf16:
+            self.model.optimizer.amp_init(self.exe.place)
+        fleet.init_worker()
+
+        init_model_path = config.get("runner.infer_load_path")
+        model_mode = config.get("runner.model_mode", 0)
+        client = HangZhouOSSClient("art-recommend")
+        oss_object_name = self.config.get("runner.oss_object_name", "dyp/model.tar.gz")
+        client.get_object_to_file(oss_object_name, "model.tar.gz")
+        compress.uncompress_tar("model.tar.gz", init_model_path)
+        assert os.path.exists(init_model_path)
+
+        #if fleet.is_first_worker():
+        #fleet.load_inference_model(init_model_path, mode=int(model_mode))
+        #fleet.barrier_worker()
+
+
+        reader_type = self.config.get("runner.reader_type", "QueueDataset")
+        epochs = int(self.config.get("runner.epochs"))
+        sync_mode = self.config.get("runner.sync_mode")
+        opt_info = paddle.static.default_main_program()._fleet_opt
+        if use_auc is True:
+            opt_info['stat_var_names'] = [
+                self.model.stat_pos.name, self.model.stat_neg.name
+            ]
+        else:
+            opt_info['stat_var_names'] = []
+
+        if reader_type == "InmemoryDataset":
+            self.reader.load_into_memory()
+
+        fleet.load_inference_model(
+            init_model_path,
+            mode=int(model_mode))
+        epoch_start_time = time.time()
+        epoch = 0
+        if sync_mode == "heter":
+            self.heter_train_loop(epoch)
+        elif reader_type == "QueueDataset":
+            self.dataset_train_loop(epoch)
+        elif reader_type == "InmemoryDataset":
+            self.dataset_train_loop(epoch)
+
+        epoch_time = time.time() - epoch_start_time
+        logger.info(
+            "using time {} second, ips  {}/sec.".format(epoch_time, self.count_method))
+        while True:
+            time.sleep(300)
+            continue;
+        if reader_type == "InmemoryDataset":
+            self.reader.release_memory()
+
+    def init_reader(self):
+        if fleet.is_server():
+            return
+        self.config["runner.reader_type"] = self.config.get(
+            "runner.reader_type", "QueueDataset")
+        self.reader, self.file_list = get_infer_reader(self.input_data, config)
+        self.example_nums = 0
+        self.count_method = self.config.get("runner.example_count_method",
+                                            "example")
+
+    def dataset_train_loop(self, epoch=0):
+        logger.info("Epoch: {}, Running Dataset Begin.".format(epoch))
+
+        fetch_info = [
+            "Epoch {} Var {}".format(epoch, var_name)
+            for var_name in self.metrics
+        ]
+
+        fetch_vars = [var for _, var in self.metrics.items()]
+        input_data_names = [data.name for data in self.input_data]
+        test_dataloader = self.reader
+
+        for batch_id, batch_data in enumerate(test_dataloader()):
+            fetch_batch_var = exe.run(
+                program=paddle.static.default_main_program(),
+                feed=dict(zip(input_data_names, batch_data)),
+                fetch_list=fetch_vars)
+
+            logger.info("fetch_batch_var : {}".format(fetch_batch_var))
+
+
+    def heter_train_loop(self, epoch):
+        logger.info(
+            "Epoch: {}, Running Begin. Check running metrics at heter_log".
+            format(epoch))
+        reader_type = self.config.get("runner.reader_type")
+        if reader_type == "QueueDataset":
+            self.exe.infer_from_dataset(
+                program=paddle.static.default_main_program(),
+                dataset=self.reader,
+                debug=config.get("runner.dataset_debug"))
+        elif reader_type == "DataLoader":
+            batch_id = 0
+            train_run_cost = 0.0
+            total_examples = 0
+            self.reader.start()
+            while True:
+                try:
+                    train_start = time.time()
+                    # --------------------------------------------------- #
+                    self.exe.run(program=paddle.static.default_main_program())
+                    # --------------------------------------------------- #
+                    train_run_cost += time.time() - train_start
+                    total_examples += self.config.get("runner.batch_size")
+                    batch_id += 1
+                    print_step = int(config.get("runner.print_period"))
+                    if batch_id % print_step == 0:
+                        profiler_string = ""
+                        profiler_string += "avg_batch_cost: {} sec, ".format(
+                            format((train_run_cost) / print_step, '.5f'))
+                        profiler_string += "avg_samples: {}, ".format(
+                            format(total_examples / print_step, '.5f'))
+                        profiler_string += "ips: {} {}/sec ".format(
+                            format(total_examples / (train_run_cost), '.5f'),
+                            self.count_method)
+                        logger.info("Epoch: {}, Batch: {}, {}".format(
+                            epoch, batch_id, profiler_string))
+                        train_run_cost = 0.0
+                        total_examples = 0
+                except paddle.core.EOFException:
+                    self.reader.reset()
+                    break
+
+    def record_result(self):
+        logger.info("train_result_dict: {}".format(self.train_result_dict))
+        with open("./train_result_dict.txt", 'w+') as f:
+            f.write(str(self.train_result_dict))
+
+
+if __name__ == "__main__":
+    paddle.enable_static()
+
+    config = parse_args()
+    os.environ["CPU_NUM"] = str(config.get("runner.thread_num"))
+    benchmark_main = Main(config)
+
+    benchmark_main.run()