modify readme and news - #49
Merged
Merged
Conversation
Remove the LingBot-specific RoboTwin MessagePack server, validator, dependency file, and dead action scheduler compatibility alias. Keep direct inference, native HTTP, generic VLA session WebSocket, and MuJoCo entrypoints as the supported integration paths. Update VLA documentation and the MuJoCo example to use the generic service defaults. Verification: scripts/run_ci_tests.sh --skip-install (lint, format, import, CPU unit, and server tests); local MuJoCo smoke; focused LingBot VLA and VLA session tests.
Remove the unreferenced ActionChunkScheduler now that session admission and latest-wins handling are owned by the generic VLA WebSocket runtime. Keep the active VLA contracts, replica sessions, simulator execution path, and all LingBot VLA v2 inference modes unchanged. Update VLA documentation to describe the inference-only verification scope, MuJoCo boundary, and retained parity and performance validation. Verification: 130 focused VLA, LingBot, service, replica, and validation tests passed; 237 non-LiveKit service tests passed; runtime import and compile checks passed; git diff --check passed. Full service collection remains blocked only by the optional av dependency, and model-core tests require the pinned Transformers 5.14.1 environment.
Document the general VLA session-serving abstraction and the LingBot-VLA v2 integration milestone in the English and Chinese homepage News sections.\n\nHighlight semantic observation/action contracts, extensible embodiment adapters, replica-affine sessions, WebSocket continuous-control serving, MuJoCo validation, structured HTTP, BF16 eager, CUDA Graph, and quantized inference support.\n\nVerification:\n- git diff --check
Link the LingBot-VLA v2 News entry to its detailed guide and document CUDA Graph and quantization profile performance comparisons.\n\nAdd H100 BF16 eager versus denoising Graph A/B measurements, quantization latency and VRAM tables, benchmark scope, and release-validation caveats. Keep quantized profiles clearly described as capacity and memory trade-offs where exact replay validation is incomplete.\n\nVerification:\n- git diff --check
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
1 participant
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
Motivation
Type of Change
Changes Made
Testing
pytest tests/)Test commands:
# Add the commands used to test your changes pytest tests/Checklist
ruff)pre-commit run --all-files)pytest tests/)[TYPE] Brief descriptionRelated Issues
Fixes #
Additional Notes
GPU Architecture Support
Performance Impact