Skip to content

fix: OnPolicyAlgorithm doesnot have the parameter: create_eval_env #36

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
hscspring wants to merge 2 commits into
base: main
Choose a base branch
from
Open

fix: OnPolicyAlgorithm doesnot have the parameter: create_eval_env #36

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
856b7d2
fix: OnPolicyAlgorithm doesnot have the parameter: create_eval_env
hscspring Feb 1, 2023
58b0632
fix: It's better to assign dtype in DictSpace of TextGenEnv explicitly
hscspring Feb 1, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion rl4lms/algorithms/nlpo/nlpo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -114,7 +114,7 @@ def __init__(
use_sde=False,
sde_sample_freq=-1,
tensorboard_log=tensorboard_log,
create_eval_env=create_eval_env,
# create_eval_env=create_eval_env,
policy_kwargs=policy_kwargs,
verbose=verbose,
seed=seed,
Expand Down
13 changes: 7 additions & 6 deletions rl4lms/envs/text_generation/env.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,7 @@
from cmath import inf
from typing import Dict, Tuple, Optional, List

import numpy as np
import torch
from gym import Env, spaces
from gym.spaces.dict import Dict as DictSpace
Expand Down Expand Up @@ -57,24 +58,24 @@ def __init__(
# we have to provide fixed sized inputs (padded) because sb3 support for DictObsersevation is limited
# while creating rollout buffers, observations are concatenated for each key
"prompt_or_input_encoded_pt": spaces.Box(
low=0, high=self._vocab_size, shape=(self._max_text_length,)
low=0, high=self._vocab_size, shape=(self._max_text_length,), dtype=np.int64
),
"prompt_or_input_attention_mask_pt": spaces.Box(
low=0, high=1, shape=(self._max_text_length,)
low=0, high=1, shape=(self._max_text_length,), dtype=np.int64
),
"context_encoded_pt": spaces.Box(
low=0, high=self._vocab_size, shape=(self.max_steps,)
low=0, high=self._vocab_size, shape=(self.max_steps,), dtype=np.int64
),
"context_attention_mask_pt": spaces.Box(
low=0, high=1, shape=(self.max_steps,)
low=0, high=1, shape=(self.max_steps,), dtype=np.int64
),
"input_encoded_pt": spaces.Box(
low=0,
high=self._vocab_size,
shape=(self._max_text_length + self.max_steps,),
shape=(self._max_text_length + self.max_steps,), dtype=np.int64
),
"input_attention_mask_pt": spaces.Box(
low=0, high=1, shape=(self._max_text_length + self.max_steps,)
low=0, high=1, shape=(self._max_text_length + self.max_steps,), dtype=np.int64
),
}
)
Expand Down