Inference Move API #178
Merged
Inference Move API #178
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
nishadgothoskar
force-pushed
the
fiddle_with_inference
branch
from
0d13266 to
b41ede4
Compare
|
|
||
| ### Run inference ### | ||
| for T in tqdm(range(len(all_data))): | ||
| key = b3d.split_key(key) | ||
| trace = inference_step(trace, key, all_data[T]["rgbd"]) | ||
| trace = inference.advance_time(key, trace, all_data[T]["rgbd"]) | ||
| trace = inference_old.inference_step(trace, key, inference_hyperparams)[ |
There was a problem hiding this comment.
@nishadgothoskar for tonight, can we aim to launch at least 1 run that is using the new inference algorithm? This will be through a call to inference.inference_step_c2f. I can add this to the script if helpful.
There was a problem hiding this comment.
yes definitely. that is the plan.
| @@ -154,38 +160,43 @@ def propose_a_points_attributes( | |||
| return _propose_a_points_attributes( | |||
| key, | |||
| observed_rgbd_for_point=observed_rgbd_for_point, | |||
| latent_depth=new_state["pose"].apply(hyperparams["vertices"][vertex_index])[2], | |||
| latent_rgbd_for_point=jnp.array( | |||
There was a problem hiding this comment.
@nishadgothoskar isn't this going to break inference?
There was a problem hiding this comment.
What I mean is -- don't we need to set the rgb correctly here?
There was a problem hiding this comment.
i think the rgb is ignored
|
Got it - in that case, why did we change the signature to have RGB? Should
we revert?
…On Mon, Sep 16, 2024 at 3:19 PM nishadgothoskar ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In src/b3d/chisight/gen3d/inference_moves.py
<#178 (comment)>:
> @@ -154,38 +160,43 @@ def propose_a_points_attributes(
return _propose_a_points_attributes(
key,
observed_rgbd_for_point=observed_rgbd_for_point,
- latent_depth=new_state["pose"].apply(hyperparams["vertices"][vertex_index])[2],
+ latent_rgbd_for_point=jnp.array(
i think the rgb is ignored
—
Reply to this email directly, view it on GitHub
<#178 (comment)>, or
unsubscribe
<https://github.com/notifications/unsubscribe-auth/AD7OROLEKRYDAXSFRABLU4TZW4VNTAVCNFSM6AAAAABOJYCOCGVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDGMBXGU4DANJRG4>
.
You are receiving this because you commented.Message ID:
***@***.***>
|
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.
No description provided.