fix(nyz): fix pytorch1.9.0 compatibility bug and change naive buffer log freq

ding/entry/tests/test_serial_entry_il.py

@@ -60,7 +60,7 @@ def test_serial_pipeline_il_ppo():
 
     # il training 1
     il_config = [deepcopy(cartpole_ppo_offpolicy_config), deepcopy(cartpole_ppo_offpolicy_create_config)]
-    il_config[0].policy.learn.train_epoch = 10
+    il_config[0].policy.learn.train_epoch = 20
     il_config[0].policy.type = 'ppo_il'
     _, converge_stop_flag = serial_pipeline_il(il_config, seed=314, data_path=expert_data_path)
     assert converge_stop_flag

ding/utils/data/collate_fn.py

@@ -3,8 +3,10 @@ from typing import List, Dict, Union, Any
 
 import torch
 import re
-from torch._six import container_abcs, string_classes, int_classes
+from torch._six import string_classes
+import collections.abc as container_abcs
 
+int_classes = int
 np_str_obj_array_pattern = re.compile(r'[SaUO]')
 
 default_collate_err_msg_format = (

ding/worker/replay_buffer/naive_buffer.py

@@ -77,7 +77,7 @@ class NaiveReplayBuffer(IBuffer):
             )
         # Periodic thruput. Here by default, monitor range is 60 seconds. You can modify it for free.
         self._periodic_thruput_monitor = PeriodicThruputMonitor(
-            self._instance_name, EasyDict(seconds=3), self._logger, self._tb_logger
+            self._instance_name, EasyDict(seconds=60), self._logger, self._tb_logger
         )
 
     def start(self) -> None:

setup.py

@@ -48,7 +48,7 @@ setup(
         'requests>=2.25.1',
         'six',
         'gym>=0.20.0',  # pypy incompatible
-        'torch>=1.3.1,<=1.8.0',  # PyTorch 1.9.0 is available, but you need to do something like https://github.com/opendilab/DI-engine/discussions/81
+        'torch>=1.3.1,<=1.9.0',  # PyTorch 1.9.0 is available, if some errors, you need to do something like https://github.com/opendilab/DI-engine/discussions/81
         'pyyaml',
         'easydict==1.9',
         'tensorboardX>=2.1,<=2.2',