Title: Request: PyTorch 2.8+ compatibility patches · Issue #588 · oracle/graalpython · GitHub
Open Graph Title: Request: PyTorch 2.8+ compatibility patches · Issue #588 · oracle/graalpython
X Title: Request: PyTorch 2.8+ compatibility patches · Issue #588 · oracle/graalpython
Description: Summary We're using GraalPy to build PyTorch from source for a project that traces PyTorch models using torch.fx.symbolic_trace and converts them to StableHLO MLIR. The current patches support up to PyTorch 2.7.0, but PyTorch mainline is...
Open Graph Description: Summary We're using GraalPy to build PyTorch from source for a project that traces PyTorch models using torch.fx.symbolic_trace and converts them to StableHLO MLIR. The current patches support up t...
X Description: Summary We're using GraalPy to build PyTorch from source for a project that traces PyTorch models using torch.fx.symbolic_trace and converts them to StableHLO MLIR. The current patches support ...
Opengraph URL: https://github.com/oracle/graalpython/issues/588
X: @github
Domain: github.com
{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Request: PyTorch 2.8+ compatibility patches","articleBody":"## Summary\n\nWe're using GraalPy to build PyTorch from source for a project that traces PyTorch models using `torch.fx.symbolic_trace` and converts them to StableHLO MLIR. The current patches support up to PyTorch 2.7.0, but PyTorch mainline is now at 2.11+.\n\n## Request\n\nPlease add compatibility patches for PyTorch 2.8, 2.9, 2.10, and 2.11 (or at least the latest stable release).\n\n## Context\n\n- **Use case**: We embed GraalPy in a native-image to trace PyTorch models and export them to StableHLO for GPU kernel generation\n- **Current setup**: GraalPy 25.0.1 + PyTorch 2.7.0 (working great!)\n- **Tracing APIs used**: `torch.fx.symbolic_trace`, `torch.jit.trace`, `torch.jit.script`\n\n## Current patch coverage\n\nThe `graalpython/lib-graalpython/patches/` directory contains:\n- `torch-1.13.1.patch`\n- `torch-2.2.1.patch`\n- `torch-2.4.1.patch`\n- `torch-2.7.0.patch`\n\n## Why this matters\n\nPyTorch 2.8+ includes improvements to the FX tracing infrastructure and better support for modern model architectures. Staying current with PyTorch versions helps us support newer models and take advantage of tracing improvements.\n\n## Workaround\n\nWe've bundled the `torch-2.7.0.patch` in our repo and it works well. We're happy to help test patches for newer versions if that would be useful.\n\nThank you for the excellent work on GraalPy - the PyTorch integration is impressive!","author":{"url":"https://github.com/morrismeyer","@type":"Person","name":"morrismeyer"},"datePublished":"2026-01-10T18:19:42.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":0},"url":"https://github.com/588/graalpython/issues/588"}
| route-pattern | /_view_fragments/issues/show/:user_id/:repository/:id/issue_layout(.:format) |
| route-controller | voltron_issues_fragments |
| route-action | issue_layout |
| fetch-nonce | v2:3fc34450-c547-6ba2-38cf-a8e4736a827f |
| current-catalog-service-hash | 81bb79d38c15960b92d99bca9288a9108c7a47b18f2423d0f6438c5b7bcd2114 |
| request-id | A9B4:31D824:75086:9E354:696970BE |
| html-safe-nonce | 3b73feb641c325c8300a3b1d5deb9a9de24683da72407d0affda89d445d3a7e9 |
| visitor-payload | eyJyZWZlcnJlciI6IiIsInJlcXVlc3RfaWQiOiJBOUI0OjMxRDgyNDo3NTA4Njo5RTM1NDo2OTY5NzBCRSIsInZpc2l0b3JfaWQiOiIyMjQzNDYxODc1MTgyOTU2NzM1IiwicmVnaW9uX2VkZ2UiOiJpYWQiLCJyZWdpb25fcmVuZGVyIjoiaWFkIn0= |
| visitor-hmac | 251c87cc17bd79658fcf7f027d8a122ea01344af596a3de3c73f655900861f0a |
| hovercard-subject-tag | issue:3800174527 |
| github-keyboard-shortcuts | repository,issues,copilot |
| google-site-verification | Apib7-x98H0j5cPqHWwSMm6dNU4GmODRoqxLiDzdx9I |
| octolytics-url | https://collector.github.com/github/collect |
| analytics-location | / |
| fb:app_id | 1401488693436528 |
| apple-itunes-app | app-id=1477376905, app-argument=https://github.com/_view_fragments/issues/show/oracle/graalpython/588/issue_layout |
| twitter:image | https://opengraph.githubassets.com/5085ee88f4b775f89a638cf21aa97af6a38e861bcdfa6d618e7b589e6177da4f/oracle/graalpython/issues/588 |
| twitter:card | summary_large_image |
| og:image | https://opengraph.githubassets.com/5085ee88f4b775f89a638cf21aa97af6a38e861bcdfa6d618e7b589e6177da4f/oracle/graalpython/issues/588 |
| og:image:alt | Summary We're using GraalPy to build PyTorch from source for a project that traces PyTorch models using torch.fx.symbolic_trace and converts them to StableHLO MLIR. The current patches support up t... |
| og:image:width | 1200 |
| og:image:height | 600 |
| og:site_name | GitHub |
| og:type | object |
| og:author:username | morrismeyer |
| hostname | github.com |
| expected-hostname | github.com |
| None | c6f193beb8ff08443adc07685d75302ab8aaf0a135f6e251c3ff3112c8deb881 |
| turbo-cache-control | no-preview |
| go-import | github.com/oracle/graalpython git https://github.com/oracle/graalpython.git |
| octolytics-dimension-user_id | 4430336 |
| octolytics-dimension-user_login | oracle |
| octolytics-dimension-repository_id | 129883600 |
| octolytics-dimension-repository_nwo | oracle/graalpython |
| octolytics-dimension-repository_public | true |
| octolytics-dimension-repository_is_fork | false |
| octolytics-dimension-repository_network_root_id | 129883600 |
| octolytics-dimension-repository_network_root_nwo | oracle/graalpython |
| turbo-body-classes | logged-out env-production page-responsive |
| disable-turbo | false |
| browser-stats-url | https://api.github.com/_private/browser/stats |
| browser-errors-url | https://api.github.com/_private/browser/errors |
| release | ec4d88fbe7ed62446d90b0a6d8e6db4248a8f8b8 |
| ui-target | full |
| theme-color | #1e2327 |
| color-scheme | light dark |
Links:
Viewport: width=device-width