Yangyangt/try mapping#1
Draft
yy-code-nv wants to merge 11 commits into
Draft
Conversation
Apply the cosmos-framework-release pipeline output over the public tree: - rewrite import paths (imaginaire.* / projects.cosmos3.* -> cosmos_framework.*) - rewrite internal module/file references in comments, docstrings, and path-form string literals via the file mapping - redact sensitive infra identifiers (S3 buckets, NFS/user paths, internal cluster config classes) - normalize license headers to OpenMDW (stamp missing, swap Apache, bump year) - exclude the one_logger / training_telemetry READMEs (broken relative links) - strip all TODO comments and any name-attributed marker comments (e.g. `# NOTE (alice): ...`); keep un-attributed NOTE/HACK/etc. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…aked proprietary headers Re-run of the release pipeline with two new transforms: - Strip COSMOS-RELEASE-BEGIN-IGNORE..END-IGNORE blocks (one_logger, cuDNN, internal notes). - Replace leaked NVIDIA proprietary/confidential headers with the OpenMDW SPDX header. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Re-run of the release pipeline: source added COSMOS-RELEASE-BEGIN-IGNORE markers around the dataloader_weighted_url imports and registration calls, which the pipeline now removes. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…+ data_registration deps - New: cosmos_framework/data/imaginaire/webdataset/augmentors/image/ (6 modules + flip.py + __init__). - configs/base/defaults/callbacks.py: new COSMOS-RELEASE-IGNORE block strips the dataloading_monitor import + DetailedDataLoadingSpeedMonitor callback, removing the dataloading_monitor / webdataset.utils.stream dependency. - data/vfm/augmentors/text_transforms_for_image.py: inlines _CAPTION_EMBEDDING_KEY_MAPPING_IMAGES, removing the data_sources.data_registration import. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…n nvlm pair - configs/base/vlm/defaults/callbacks.py: new COSMOS-RELEASE-IGNORE block removes the dataloading_monitor import + speed-monitor callback. - Drop nvlm_data_unify.py and nvlm_sample_loaders_and_part_filters.py: unreachable in the release (no CF code imports nvlm_data_unify), now excluded from the mapping. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
multiview_dataloader is unreachable: nothing imports it, and its only effect would be to register the (unmapped) multiview_data_source / multiview_dataset. Now excluded from the mapping. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Source added a COSMOS-RELEASE-IGNORE block around the dataloader.register_data_debug import + call site; pipeline now strips it. This leaves dataloader.py with zero CF importers — orphan candidate. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
After vlm/config.py was IGNORE-stripped of its register_data_debug import,
dataloader.py has zero CF importers. It was also the only consumer of
the unmapped vlm.{collate_fn,debug_data_qwen,dummy_data_qwen} modules,
so dropping it resolves those three dangling imports as well.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Pulled in by the vfm augmentors (bytes_to_media, tokenize_data). Only project-internal imports are imaginaire.utils.log and the already-shipped qwen3vl_processor; no new external deps. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Required by model/tokenizer/models/__init__.py (calculate_psnr re-export). FVDMetric is commented out in source (depended on the unmapped FVD + paths modules). A residual lazy import of tokenizer.evaluation.metric inside TokenizerMetric's codebook-usage branch remains dangling; only that code path is affected. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…ocess, drop broken tests) Ship nemotron3densevl/nemotronvl processors, tokenizer evaluation/metric, and vfm/video_preprocess to satisfy previously-dangling imports. Drop test files that import unshipped helpers (helper_test/unittest_utils, scripts.eval). CF now has no dangling cosmos_framework module imports. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
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.
No description provided.