Repository navigation
Conversation
TurboVLA keeps F32 weights by default for fidelity (bf16 is ~4e-2 from fp32). On SYCL, F16 weights with flash attention stay within 3.4e-4 of the F32 graph, so that is the default there unless --weight-dtype / --flash-attn are given. Core Ultra 7 356H, 2 views at 256 px, vla-bench p50: 117-128 -> 66-74 ms. LIBERO-object x100: 100/100 in both (median 140 -> 89 ms per request).
This file contains hidden or 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
turbovla: f16 weights and flash attention by default on SYCL
TurboVLA keeps F32 weights by default for fidelity; with BF16 weights its actions are ~4e-2 from fp32. On SYCL, F16-resident weights stay within 3.4e-4 of the F32 graph, flash attention included (oneDNN's fused SDPA). So this makes F16 + FA the SYCL default, unless
--weight-dtypeor--flash-attnis given. Other backends are unchanged. Independent of the other PRs.Results
Core Ultra 7 356H (Panther Lake), 2 views at 256 px:
vla-benchp50LIBERO-object: 10 tasks x 10 episodes per arm, same binary, every task 10/10 in both arms. Ranges are two measurements at different GPU clocks.
Verified
-Wall -Wextra(first-party code)ctestpassesvla_predict_checkdiff), or the change ismeant to move it and a LIBERO sweep is below
Archs and backends tested:
Linux x86 / PTL