593 lines
23 KiB
Python

import logging
import time
from abc import ABC, abstractmethod
from collections import defaultdict, deque
from collections.abc import Callable
from typing import Any
import numpy as np
import tqdm
from tianshou.data import AsyncCollector, Collector, ReplayBuffer
from tianshou.policy import BasePolicy
from tianshou.trainer.utils import gather_info, test_episode
from tianshou.utils import (
BaseLogger,
DummyTqdm,
LazyLogger,
MovAvg,
deprecation,
tqdm_config,
)
log = logging.getLogger(__name__)
class BaseTrainer(ABC):
"""An iterator base class for trainers.
Returns an iterator that yields a 3-tuple (epoch, stats, info) of train results
on every epoch.
:param policy: an instance of the :class:`~tianshou.policy.BasePolicy` class.
:param batch_size: the batch size of sample data, which is going to feed in
the policy network.
:param train_collector: the collector used for training.
:param test_collector: the collector used for testing. If it's None,
then no testing will be performed.
:param buffer: the replay buffer used for off-policy algorithms or for pre-training.
If a policy overrides the ``process_buffer`` method, the replay buffer will
be pre-processed before training.
:param max_epoch: the maximum number of epochs for training. The training
process might be finished before reaching ``max_epoch`` if ``stop_fn``
is set.
:param step_per_epoch: the number of transitions collected per epoch.
:param repeat_per_collect: the number of repeat time for policy learning,
for example, set it to 2 means the policy needs to learn each given batch
data twice. Only used in on-policy algorithms
:param episode_per_test: the number of episodes for one policy evaluation.
:param update_per_step: only used in off-policy algorithms.
How many gradient steps to perform per step in the environment
(i.e., per sample added to the buffer).
:param step_per_collect: the number of transitions the collector would
collect before the network update, i.e., trainer will collect
"step_per_collect" transitions and do some policy network update repeatedly
in each epoch.
:param episode_per_collect: the number of episodes the collector would
collect before the network update, i.e., trainer will collect
"episode_per_collect" episodes and do some policy network update repeatedly
in each epoch.
:param train_fn: a hook called at the beginning of training in each
epoch. It can be used to perform custom additional operations, with the
signature ``f(num_epoch: int, step_idx: int) -> None``.
:param test_fn: a hook called at the beginning of testing in each
epoch. It can be used to perform custom additional operations, with the
signature ``f(num_epoch: int, step_idx: int) -> None``.
:param save_best_fn: a hook called when the undiscounted average mean
reward in evaluation phase gets better, with the signature
``f(policy: BasePolicy) -> None``. It was ``save_fn`` previously.
:param save_checkpoint_fn: a function to save training process and
return the saved checkpoint path, with the signature ``f(epoch: int,
env_step: int, gradient_step: int) -> str``; you can save whatever you want.
:param resume_from_log: resume env_step/gradient_step and other metadata
from existing tensorboard log.
:param stop_fn: a function with signature ``f(mean_rewards: float) ->
bool``, receives the average undiscounted returns of the testing result,
returns a boolean which indicates whether reaching the goal.
:param reward_metric: a function with signature
``f(rewards: np.ndarray with shape (num_episode, agent_num)) -> np.ndarray
with shape (num_episode,)``, used in multi-agent RL. We need to return a
single scalar for each episode's result to monitor training in the
multi-agent RL setting. This function specifies what is the desired metric,
e.g., the reward of agent 1 or the average reward over all agents.
:param logger: A logger that logs statistics during
training/testing/updating. To not log anything, keep the default logger.
:param verbose: whether to print status information to stdout.
If set to False, status information will still be logged (provided that
logging is enabled via the `logging` module).
:param show_progress: whether to display a progress bar when training.
:param test_in_train: whether to test in the training phase.
"""
__doc__: str
@staticmethod
def gen_doc(learning_type: str) -> str:
"""Document string for subclass trainer."""
step_means = f'The "step" in {learning_type} trainer means '
if learning_type != "offline":
step_means += "an environment step (a.k.a. transition)."
else: # offline
step_means += "a gradient step."
trainer_name = learning_type.capitalize() + "Trainer"
return f"""An iterator class for {learning_type} trainer procedure.
Returns an iterator that yields a 3-tuple (epoch, stats, info) of
train results on every epoch.
{step_means}
Example usage:
::
trainer = {trainer_name}(...)
for epoch, epoch_stat, info in trainer:
print("Epoch:", epoch)
print(epoch_stat)
print(info)
do_something_with_policy()
query_something_about_policy()
make_a_plot_with(epoch_stat)
display(info)
- epoch int: the epoch number
- epoch_stat dict: a large collection of metrics of the current epoch
- info dict: result returned from :func:`~tianshou.trainer.gather_info`
You can even iterate on several trainers at the same time:
::
trainer1 = {trainer_name}(...)
trainer2 = {trainer_name}(...)
for result1, result2, ... in zip(trainer1, trainer2, ...):
compare_results(result1, result2, ...)
"""
def __init__(
self,
policy: BasePolicy,
max_epoch: int,
batch_size: int,
train_collector: Collector | None = None,
test_collector: Collector | None = None,
buffer: ReplayBuffer | None = None,
step_per_epoch: int | None = None,
repeat_per_collect: int | None = None,
episode_per_test: int | None = None,
update_per_step: float = 1.0,
step_per_collect: int | None = None,
episode_per_collect: int | None = None,
train_fn: Callable[[int, int], None] | None = None,
test_fn: Callable[[int, int | None], None] | None = None,
stop_fn: Callable[[float], bool] | None = None,
save_best_fn: Callable[[BasePolicy], None] | None = None,
save_checkpoint_fn: Callable[[int, int, int], str] | None = None,
resume_from_log: bool = False,
reward_metric: Callable[[np.ndarray], np.ndarray] | None = None,
logger: BaseLogger = LazyLogger(),
verbose: bool = True,
show_progress: bool = True,
test_in_train: bool = True,
save_fn: Callable[[BasePolicy], None] | None = None,
):
if save_fn:
deprecation(
"save_fn in trainer is marked as deprecated and will be "
"removed in the future. Please use save_best_fn instead.",
)
assert save_best_fn is None
save_best_fn = save_fn
self.policy = policy
if buffer is not None:
buffer = policy.process_buffer(buffer)
self.buffer = buffer
self.train_collector = train_collector
self.test_collector = test_collector
self.logger = logger
self.start_time = time.time()
self.stat: defaultdict[str, MovAvg] = defaultdict(MovAvg)
self.best_reward = 0.0
self.best_reward_std = 0.0
self.start_epoch = 0
# This is only used for logging but creeps into the implementations
# of the trainers. I believe it would be better to remove
self.gradient_step = 0
self.env_step = 0
self.max_epoch = max_epoch
self.step_per_epoch = step_per_epoch
# either on of these two
self.step_per_collect = step_per_collect
self.episode_per_collect = episode_per_collect
self.update_per_step = update_per_step
self.repeat_per_collect = repeat_per_collect
self.episode_per_test = episode_per_test
self.batch_size = batch_size
self.train_fn = train_fn
self.test_fn = test_fn
self.stop_fn = stop_fn
self.save_best_fn = save_best_fn
self.save_checkpoint_fn = save_checkpoint_fn
self.reward_metric = reward_metric
self.verbose = verbose
self.show_progress = show_progress
self.test_in_train = test_in_train
self.resume_from_log = resume_from_log
self.is_run = False
self.last_rew, self.last_len = 0.0, 0
self.epoch = self.start_epoch
self.best_epoch = self.start_epoch
self.stop_fn_flag = False
self.iter_num = 0
def reset(self) -> None:
"""Initialize or reset the instance to yield a new iterator from zero."""
self.is_run = False
self.env_step = 0
if self.resume_from_log:
(
self.start_epoch,
self.env_step,
self.gradient_step,
) = self.logger.restore_data()
self.last_rew, self.last_len = 0.0, 0
self.start_time = time.time()
if self.train_collector is not None:
self.train_collector.reset_stat()
if self.train_collector.policy != self.policy or self.test_collector is None:
self.test_in_train = False
if self.test_collector is not None:
assert self.episode_per_test is not None
assert not isinstance(self.test_collector, AsyncCollector) # Issue 700
self.test_collector.reset_stat()
test_result = test_episode(
self.policy,
self.test_collector,
self.test_fn,
self.start_epoch,
self.episode_per_test,
self.logger,
self.env_step,
self.reward_metric,
)
self.best_epoch = self.start_epoch
self.best_reward, self.best_reward_std = (
test_result["rew"],
test_result["rew_std"],
)
if self.save_best_fn:
self.save_best_fn(self.policy)
self.epoch = self.start_epoch
self.stop_fn_flag = False
self.iter_num = 0
def __iter__(self): # type: ignore
self.reset()
return self
def __next__(self) -> None | tuple[int, dict[str, Any], dict[str, Any]]:
"""Perform one epoch (both train and eval)."""
self.epoch += 1
self.iter_num += 1
if self.iter_num > 1:
# iterator exhaustion check
if self.epoch > self.max_epoch:
raise StopIteration
# exit flag 1, when stop_fn succeeds in train_step or test_step
if self.stop_fn_flag:
raise StopIteration
# set policy in train mode
self.policy.train()
epoch_stat: dict[str, Any] = {}
progress = tqdm.tqdm if self.show_progress else DummyTqdm
# perform n step_per_epoch
with progress(total=self.step_per_epoch, desc=f"Epoch #{self.epoch}", **tqdm_config) as t:
while t.n < t.total and not self.stop_fn_flag:
data: dict[str, Any] = {}
result: dict[str, Any] = {}
if self.train_collector is not None:
data, result, self.stop_fn_flag = self.train_step()
t.update(result["n/st"])
if self.stop_fn_flag:
t.set_postfix(**data)
break
else:
assert self.buffer, "No train_collector or buffer specified"
result["n/ep"] = len(self.buffer)
result["n/st"] = int(self.gradient_step)
t.update()
self.policy_update_fn(data, result)
t.set_postfix(**data)
if t.n <= t.total and not self.stop_fn_flag:
t.update()
# for offline RL
if self.train_collector is None:
self.env_step = self.gradient_step * self.batch_size
if not self.stop_fn_flag:
self.logger.save_data(
self.epoch,
self.env_step,
self.gradient_step,
self.save_checkpoint_fn,
)
# test
if self.test_collector is not None:
test_stat, self.stop_fn_flag = self.test_step()
if not self.is_run:
epoch_stat.update(test_stat)
if not self.is_run:
epoch_stat.update({k: v.get() for k, v in self.stat.items()})
epoch_stat["gradient_step"] = self.gradient_step
epoch_stat.update(
{
"env_step": self.env_step,
"rew": self.last_rew,
"len": int(self.last_len),
"n/ep": int(result["n/ep"]),
"n/st": int(result["n/st"]),
},
)
info = gather_info(
self.start_time,
self.train_collector,
self.test_collector,
self.best_reward,
self.best_reward_std,
)
return self.epoch, epoch_stat, info
return None
def test_step(self) -> tuple[dict[str, Any], bool]:
"""Perform one testing step."""
assert self.episode_per_test is not None
assert self.test_collector is not None
stop_fn_flag = False
test_result = test_episode(
self.policy,
self.test_collector,
self.test_fn,
self.epoch,
self.episode_per_test,
self.logger,
self.env_step,
self.reward_metric,
)
rew, rew_std = test_result["rew"], test_result["rew_std"]
if self.best_epoch < 0 or self.best_reward < rew:
self.best_epoch = self.epoch
self.best_reward = float(rew)
self.best_reward_std = rew_std
if self.save_best_fn:
self.save_best_fn(self.policy)
log_msg = (
f"Epoch #{self.epoch}: test_reward: {rew:.6f} ± {rew_std:.6f},"
f" best_reward: {self.best_reward:.6f} ± "
f"{self.best_reward_std:.6f} in #{self.best_epoch}"
)
log.info(log_msg)
if self.verbose:
print(log_msg, flush=True)
if not self.is_run:
test_stat = {
"test_reward": rew,
"test_reward_std": rew_std,
"best_reward": self.best_reward,
"best_reward_std": self.best_reward_std,
"best_epoch": self.best_epoch,
}
else:
test_stat = {}
if self.stop_fn and self.stop_fn(self.best_reward):
stop_fn_flag = True
return test_stat, stop_fn_flag
def train_step(self) -> tuple[dict[str, Any], dict[str, Any], bool]:
"""Perform one training step."""
assert self.episode_per_test is not None
assert self.train_collector is not None
stop_fn_flag = False
if self.train_fn:
self.train_fn(self.epoch, self.env_step)
result = self.train_collector.collect(
n_step=self.step_per_collect,
n_episode=self.episode_per_collect,
)
if result["n/ep"] > 0 and self.reward_metric:
rew = self.reward_metric(result["rews"])
result.update(rews=rew, rew=rew.mean(), rew_std=rew.std())
self.env_step += int(result["n/st"])
self.logger.log_train_data(result, self.env_step)
self.last_rew = result["rew"] if result["n/ep"] > 0 else self.last_rew
self.last_len = result["len"] if result["n/ep"] > 0 else self.last_len
data = {
"env_step": str(self.env_step),
"rew": f"{self.last_rew:.2f}",
"len": str(int(self.last_len)),
"n/ep": str(int(result["n/ep"])),
"n/st": str(int(result["n/st"])),
}
if (
result["n/ep"] > 0
and self.test_in_train
and self.stop_fn
and self.stop_fn(result["rew"])
):
assert self.test_collector is not None
test_result = test_episode(
self.policy,
self.test_collector,
self.test_fn,
self.epoch,
self.episode_per_test,
self.logger,
self.env_step,
)
if self.stop_fn(test_result["rew"]):
stop_fn_flag = True
self.best_reward = test_result["rew"]
self.best_reward_std = test_result["rew_std"]
else:
self.policy.train()
return data, result, stop_fn_flag
def log_update_data(self, data: dict[str, Any], losses: dict[str, Any]) -> None:
"""Log losses to current logger."""
for k in losses:
self.stat[k].add(losses[k])
losses[k] = self.stat[k].get()
data[k] = f"{losses[k]:.3f}"
self.logger.log_update_data(losses, self.gradient_step)
@abstractmethod
def policy_update_fn(self, data: dict[str, Any], result: dict[str, Any]) -> None:
"""Policy update function for different trainer implementation.
:param data: information in progress bar.
:param result: collector's return value.
"""
def run(self) -> dict[str, float | str]:
"""Consume iterator.
See itertools - recipes. Use functions that consume iterators at C speed
(feed the entire iterator into a zero-length deque).
"""
try:
self.is_run = True
deque(self, maxlen=0) # feed the entire iterator into a zero-length deque
info = gather_info(
self.start_time,
self.train_collector,
self.test_collector,
self.best_reward,
self.best_reward_std,
)
finally:
self.is_run = False
return info
def _sample_and_update(self, buffer: ReplayBuffer, data: dict[str, Any]) -> None:
self.gradient_step += 1
# Note: since sample_size=batch_size, this will perform
# exactly one gradient step. This is why we don't need to calculate the
# number of gradient steps, like in the on-policy case.
losses = self.policy.update(sample_size=self.batch_size, buffer=buffer)
data.update({"gradient_step": str(self.gradient_step)})
self.log_update_data(data, losses)
class OfflineTrainer(BaseTrainer):
"""Offline trainer, samples mini-batches from buffer and passes them to update.
Uses a buffer directly and usually does not have a collector.
"""
# for mypy
assert isinstance(BaseTrainer.__doc__, str)
__doc__ += BaseTrainer.gen_doc("offline") + "\n".join(BaseTrainer.__doc__.split("\n")[1:])
def policy_update_fn(
self,
data: dict[str, Any],
result: dict[str, Any] | None = None,
) -> None:
"""Perform one off-line policy update."""
assert self.buffer
self._sample_and_update(self.buffer, data)
class OffpolicyTrainer(BaseTrainer):
"""Offpolicy trainer, samples mini-batches from buffer and passes them to update.
Note that with this trainer, it is expected that the policy's `learn` method
does not perform additional mini-batching but just updates params from the received
mini-batch.
"""
# for mypy
assert isinstance(BaseTrainer.__doc__, str)
__doc__ += BaseTrainer.gen_doc("offpolicy") + "\n".join(BaseTrainer.__doc__.split("\n")[1:])
def policy_update_fn(self, data: dict[str, Any], result: dict[str, Any]) -> None:
"""Perform off-policy updates.
:param data:
:param result: must contain `n/st` key, see documentation of
`:meth:~tianshou.data.collector.Collector.collect` for the kind of
data returned there. `n/st` stands for `step_count`
"""
assert self.train_collector is not None
n_collected_steps = result["n/st"]
# Same as training intensity, right?
num_updates = round(self.update_per_step * n_collected_steps)
for _ in range(num_updates):
self._sample_and_update(self.train_collector.buffer, data)
class OnpolicyTrainer(BaseTrainer):
"""On-policy trainer, passes the entire buffer to .update and resets it after.
Note that it is expected that the learn method of a policy will perform
batching when using this trainer.
"""
# for mypy
assert isinstance(BaseTrainer.__doc__, str)
__doc__ = BaseTrainer.gen_doc("onpolicy") + "\n".join(BaseTrainer.__doc__.split("\n")[1:])
def policy_update_fn(
self,
data: dict[str, Any],
result: dict[str, Any] | None = None,
) -> None:
"""Perform one on-policy update."""
assert self.train_collector is not None
losses = self.policy.update(
0,
self.train_collector.buffer,
# Note: sample_size is 0, so the whole buffer is used for the update.
# The kwargs are in the end passed to the .learn method, which uses
# batch_size to iterate through the buffer in mini-batches
# Off-policy algos typically don't use the batch_size kwarg at all
batch_size=self.batch_size,
repeat=self.repeat_per_collect,
)
# just for logging, no functional role
# TODO: remove the gradient step counting in trainers? Doesn't seem like
# it's important and it adds complexity
self.gradient_step += 1
if self.batch_size > 0:
self.gradient_step += int((len(self.train_collector.buffer) - 0.1) // self.batch_size)
# Note: this is the main difference to the off-policy trainer!
# The second difference is that batches of data are sampled without replacement
# during training, whereas in off-policy or offline training, the batches are
# sampled with replacement (and potentially custom prioritization).
self.train_collector.reset_buffer(keep_statistics=True)
# The step is the number of mini-batches used for the update, so essentially
self.log_update_data(data, losses)