fix: term vs trunc #983
Draft
fix: term vs trunc #983
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
sash-a
requested review from
arnupretorius,
DriesSmit,
RuanJohn,
jcformanek,
siddarthsingh1,
OmaymaMahjoub,
ulricharmel,
callumtilbury and
WiemKhlifi
as code owners
OmaymaMahjoub
added
the
benchmark required
There was a problem hiding this comment.
Thanks @sash-a! I will continue the review once we verify the recurrent systems performance :)
Comment on lines
+160
to
+162
| trunc = jnp.repeat(timestep.last(), config.system.num_agents) | ||
| trunc = trunc.reshape(config.arch.num_envs, -1) | ||
| term = 1 - timestep.discount |
There was a problem hiding this comment.
If you can add documentation to timestep.last() and timestep.discount
| info = { | ||
| "episode_return": env_state.episode_return_info, | ||
| "episode_length": env_state.episode_length_info, | ||
| } | ||
|
|
||
| transition = PPOTransition( | ||
| done, action, value, timestep.reward, log_prob, last_timestep.observation, info | ||
| terminal=term, | ||
| truncated=trunc, |
There was a problem hiding this comment.
Random question, why do we include truncation if we don't use it 😅
There was a problem hiding this comment.
I really like the changes of removing last_done from the learner_state
| hstates, | ||
| ) = learner_state | ||
|
|
||
| rng, policy_rng = jax.random.split(rng) | ||
|
|
||
| last_trunc = jnp.repeat(last_timestep.last(), n_agents).reshape(n_envs, -1) |
There was a problem hiding this comment.
if you can add documentation # Add last_trunc to the input to the network.
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Outdated suggestions cannot be applied.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.
Note
This is quite confusing so please check this carefully and make sure I haven't mixed up where to use each one!
What?
Mostly explained in #951. We are mixing up termination and truncation. This is especially prevalent in recurrent systems, but all systems have this issue.
Now very explicit about termination vs truncation, removed all mentions of done.
Done attribute remove from
RnnLearnerStateand replaced withtruncatedin thePpoTransition.How?
Explicitly name variables
terminationandtrunctationinstead ofdone