fix(lisong): fix icm/rnd+onppo config bugs and app_key env bugs #564
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
song2181
changed the title
3 tasks
PaParaZz1
requested changes
Dec 29, 2022
| @@ -185,7 +186,7 @@ def step(self, action): | |||
|
|
|||
| obs = self.gen_obs() | |||
|
|
|||
| return obs, reward, done, {} | |||
| return obs, reward, done, done, {} | |||
There was a problem hiding this comment.
bacause in file minigrid_env.py , we use the step method of Gymnasium env, return is (observation, reward, terminated, truncated, info)
obs, rew, done, _, info = self._env.step(action)
…into dev-icm-onppo
… into dev-icm-onppo
PaParaZz1
reviewed
Mar 1, 2023
| self.tb_logger.add_scalar('icm_reward/icm_reward_mean', icm_reward.mean(), self.estimate_cnt_icm) | ||
| self.tb_logger.add_scalar('icm_reward/icm_reward_min', icm_reward.min(), self.estimate_cnt_icm) | ||
| self.tb_logger.add_scalar('icm_reward/icm_reward_std', icm_reward.std(), self.estimate_cnt_icm) | ||
| icm_reward = (raw_icm_reward - raw_icm_reward.min()) / (raw_icm_reward.max() - raw_icm_reward.min() + 1e-8) |
PaParaZz1
requested changes
Mar 1, 2023
| @@ -171,8 +174,12 @@ def __init__(self, config: EasyDict, device: str, tb_logger: 'SummaryWriter') -> | |||
| self.ce = nn.CrossEntropyLoss(reduction="mean") | |||
| self.forward_mse = nn.MSELoss(reduction='none') | |||
| self.reverse_scale = config.reverse_scale | |||
| self.res = nn.Softmax(dim=-1) | |||
There was a problem hiding this comment.
why use softmax here if we only need to sample action by argmax operation
| value_weight=0.5, | ||
| entropy_weight=0.001, | ||
| clip_ratio=0.2, | ||
| adv_norm=False, | ||
| adv_norm=True, | ||
| value_norm=True, |
There was a problem hiding this comment.
offppo doesn't have value norm
| @@ -151,6 +151,9 @@ class ICMRewardModel(BaseRewardModel): | |||
| update_per_collect=100, | |||
| # (float) the importance weight of the forward and reverse loss | |||
| reverse_scale=1, | |||
| intrinsic_reward_weight=0.003, # 1/300 | |||
There was a problem hiding this comment.
add comments for each fields in default config
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
fix config max_envstep. fix MiniGrid-AKTDT-7x7-1-v0 bug
Related Issue
TODO
Check List