fix: widen dependency version ranges to resolve install failures#78
Open
SuperMarioYL wants to merge 1 commit intoNVIDIA:mainfrom
Open
fix: widen dependency version ranges to resolve install failures#78SuperMarioYL wants to merge 1 commit intoNVIDIA:mainfrom
SuperMarioYL wants to merge 1 commit intoNVIDIA:mainfrom
Conversation
Remove restrictive upper bounds from all dependencies in pyproject.toml. The current pins (e.g., huggingface-hub < 0.25, safetensors < 0.5, einops == 0.7) prevent installation with modern Python environments where newer versions of these packages are already installed. The codebase uses only stable, well-established APIs from each library (hf_hub_download, safetensors.torch.load_model, einops.rearrange, numpy array ops, sentencepiece.SentencePieceProcessor) that have remained backward-compatible across major versions. Signed-off-by: supermario_leo <[email protected]>
Author
|
Gentle ping on this — happy to rebase onto the latest main or address any feedback. Let me know if there's anything I can adjust to help this move forward. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
Multiple users report installation failures because dependency upper bounds are too restrictive. For example:
huggingface-hub >= 0.24, < 0.25— latest is 0.30+, any modern environment failssafetensors >= 0.4.0, < 0.5— latest is 0.5+einops == 0.7— latest is 0.8+aiohttp >= 3.10.5, < 3.11— latest is 3.11+This blocks pip installs unless users manually downgrade everything, which often conflicts with other packages (especially torch ecosystem).
Related: #22, #55, #29, #10, #11
Fix
Remove all restrictive upper bounds, keeping only minimum version floors. The codebase uses only stable, well-established APIs:
huggingface-hubhf_hub_download()safetensorsload_model(),load_file(),save_file()einopsrearrange()numpysentencepieceSentencePieceProcessor()aiohttpNone of these APIs have breaking changes in their newer versions.
Scope
Note: PR #63 addresses the
torchupper bound separately; PR #76 addressestorchandaiohttp. This PR covers the remaining dependencies not addressed by those PRs.Changes
moshi/pyproject.toml