hongshaorou/dreamerv3-torch
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

erased unused lines of code

Browse Source
This commit is contained in:
NM512 2023-07-22 21:20:55 +09:00
parent d1f4d5c709
commit 106317015d
2 changed files with 8 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
dreamer.py
View File

@ -150,7 +150,6 @@ class Dreamer(nn.Module):
return tools.OneHotDist(probs=probs).sample() return tools.OneHotDist(probs=probs).sample()
else: else:
return torch.clip(torchd.normal.Normal(action, amount).sample(), -1, 1) return torch.clip(torchd.normal.Normal(action, amount).sample(), -1, 1)
raise NotImplementedError(self._config.action_noise)
def _train(self, data): def _train(self, data):
metrics = {} metrics = {}
@ -181,7 +180,7 @@ def make_dataset(episodes, config):
return dataset return dataset
def make_env(config, logger, mode, train_eps, eval_eps): def make_env(config, mode):
suite, task = config.task.split("_", 1) suite, task = config.task.split("_", 1)
if suite == "dmc": if suite == "dmc":
import envs.dmc as dmc import envs.dmc as dmc
@ -260,7 +259,7 @@ def main(config):
else: else:
directory = config.evaldir directory = config.evaldir
eval_eps = tools.load_episodes(directory, limit=1) eval_eps = tools.load_episodes(directory, limit=1)
make = lambda mode: make_env(config, logger, mode, train_eps, eval_eps) make = lambda mode: make_env(config, mode)
train_envs = [make("train") for _ in range(config.envs)] train_envs = [make("train") for _ in range(config.envs)]
eval_envs = [make("eval") for _ in range(config.envs)] eval_envs = [make("eval") for _ in range(config.envs)]
acts = train_envs[0].action_space acts = train_envs[0].action_space

12
tools.py
View File

@ -123,7 +123,7 @@ class Logger:
def simulate(agent, envs, cache, directory, logger, is_eval=False, limit=None, steps=0, episodes=0, state=None): def simulate(agent, envs, cache, directory, logger, is_eval=False, limit=None, steps=0, episodes=0, state=None):
# Initialize or unpack simulation state. # initialize or unpack simulation state
if state is None: if state is None:
step, episode = 0, 0 step, episode = 0, 0
done = np.ones(len(envs), bool) done = np.ones(len(envs), bool)
@ -134,7 +134,7 @@ def simulate(agent, envs, cache, directory, logger, is_eval=False, limit=None, s
else: else:
step, episode, done, length, obs, agent_state, reward = state step, episode, done, length, obs, agent_state, reward = state
while (steps and step < steps) or (episodes and episode < episodes): while (steps and step < steps) or (episodes and episode < episodes):
# Reset envs if necessary. # reset envs if necessary
if done.any(): if done.any():
indices = [index for index, d in enumerate(done) if d] indices = [index for index, d in enumerate(done) if d]
results = [envs[i].reset() for i in indices] results = [envs[i].reset() for i in indices]
@ -148,7 +148,7 @@ def simulate(agent, envs, cache, directory, logger, is_eval=False, limit=None, s
add_to_cache(cache, envs[i].id, t) add_to_cache(cache, envs[i].id, t)
for index, result in zip(indices, results): for index, result in zip(indices, results):
obs[index] = result obs[index] = result
# Step agents. # step agents
obs = {k: np.stack([o[k] for o in obs]) for k in obs[0]} obs = {k: np.stack([o[k] for o in obs]) for k in obs[0]}
action, agent_state = agent(obs, done, agent_state) action, agent_state = agent(obs, done, agent_state)
if isinstance(action, dict): if isinstance(action, dict):
@ -159,7 +159,7 @@ def simulate(agent, envs, cache, directory, logger, is_eval=False, limit=None, s
else: else:
action = np.array(action) action = np.array(action)
assert len(action) == len(envs) assert len(action) == len(envs)
# Step envs. # step envs
results = [e.step(a) for e, a in zip(envs, action)] results = [e.step(a) for e, a in zip(envs, action)]
obs, reward, done = zip(*[p[:3] for p in results]) obs, reward, done = zip(*[p[:3] for p in results])
obs = list(obs) obs = list(obs)
@ -169,7 +169,7 @@ def simulate(agent, envs, cache, directory, logger, is_eval=False, limit=None, s
length += 1 length += 1
step += len(envs) step += len(envs)
length *= 1 - done length *= 1 - done
# Add to cache # add to cache
for a, result, env in zip(action, results, envs): for a, result, env in zip(action, results, envs):
o, r, d, info = result o, r, d, info = result
o = {k: convert(v) for k, v in o.items()} o = {k: convert(v) for k, v in o.items()}
@ -267,7 +267,7 @@ class CollectDataset:
def reset(self): def reset(self):
obs = self._env.reset() obs = self._env.reset()
transition = obs.copy() transition = obs.copy()
# Missing keys will be filled with a zeroed out version of the first # missing keys will be filled with a zeroed out version of the first
# transition, because we do not know what action information the agent will # transition, because we do not know what action information the agent will
# pass yet. # pass yet.
transition["reward"] = 0.0 transition["reward"] = 0.0