Add an example to export a model with a test #1
Conversation
This file contains 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
Downloading pip dependencies takes 2-4 min, but downloading the new docker image takes +1min from the pytorch one Not sure how downloading all dependencies to a new image takes in comparison to pulling docker image Docker image scripts are copied from pytorch, I tried to remove stuff but there's definitely a lot left over
Follows https://github.com/pytorch/pytorch/pull/105759/files Old commits: 22min This PR: 15min
The https://pytorch.org/tutorials/beginner/deep_learning_nlp_tutorial.html is not showing up the main tutorial which has been broken down into small tutorials. The README provides links to those tutorials. This PR makes the connection. All the broken down tutorials has original author's name mentioned for providing credit. Signed-off-by: Sahdev Zala <[email protected]>
…orch#2828) The current RNN module does not implement the canonical "vanilla RNN" which consists of 3 linear layers, and a tanh operated on the sum of the projected x_t and h_{t-1}
As in title. We are now using gha to build tutorials The windows scripts aren't being used and I'm not sure if people are still interested in them. If there is interest, I can add them back to a different folder (probably the ci folder)
* Update FSDP_tutorial.rst This fixes the install instruction and link to docs to point to latest stable. * Update intermediate_source/FSDP_tutorial.rst Co-authored-by: Svetlana Karslioglu <[email protected]> --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
Add back in an explicitly defined user in the docker build because I can't figure out which user is being used by the docker image. If a user other than ec2-user is being used, then it is not possibly for the next job to clean up files from previous jobs because it lacks permission. Should fix the problems like https://github.com/pytorch/tutorials/actions/runs/8635938916/job/23674817847 ``` Cleaning the repository Warning: Unable to clean or reset the repository. The repository will be recreated instead. Deleting the contents of '/home/ec2-user/actions-runner/_work/tutorials/tutorials' Error: File was unable to be removed Error: EACCES: permission denied, unlink '/home/ec2-user/actions-runner/_work/tutorials/tutorials/.jenkins/__pycache__/get_files_to_run.cpython-310.pyc' ```
…2821) * Update message Co-authored-by: intel-hm <[email protected]> * Improve tensor dimensionality description and functional layers Co-author-by: albanD --------- Co-authored-by: jmarin <[email protected]> Co-authored-by: intel-hm <[email protected]>
* Pull 2.3 RC binaries * Turn off failing torchtext tutorials * Disable usb_semisup_learn * [export] Remove interactive aoti compilation (pytorch#2835) --------- Co-authored-by: Svetlana Karslioglu <[email protected]> Co-authored-by: Angela Yi <[email protected]>
* Add tutorial for swap_tensors in nn.Module * Add ref in load_state_dict_tips to address outdated constraint --------- Co-authored-by: Svetlana Karslioglu <[email protected]> Co-authored-by: albanD <[email protected]>
…or (pytorch#2819) * Add dynamic quant to the tutorial of PT2E quantization with X86Inductor --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
* Add Tensor Parallel Tutorial --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
* [export] Add non-strict and derived dims tutorials * add another derived dim example --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
…amples (pytorch#2831) * update the sdpa docs to use new context manager and attention_bias example --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
… and distributed state dict (pytorch#2832) * updates dcp tutorial with recent updates to api including save, load Co-authored-by: Svetlana Karslioglu <[email protected]> --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
* Disable intermediate_source/torchvision_tutorial.py Due to `RuntimeError: DataLoader worker (pid(s) 20092) exited unexpectedly`
* Add tutorial for user defined triton kernels --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
Convert to Python and update for 2.3 --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
Co-authored-by: jmarin <[email protected]> Co-authored-by: Svetlana Karslioglu <[email protected]>
* Save/read Checkpoint to file --------- Co-authored-by: Kai Fricke <[email protected]> Co-authored-by: Svetlana Karslioglu <[email protected]>
* Change Mobile to Edge * Update layout.html to collapse sections
Add parentheses to method. Co-authored-by: Svetlana Karslioglu <[email protected]>
…ytorch#2813) * Update nested tensor + MHA tutorial to include SDPA + torch.compile * Editorial update --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
…h#2855) `c1` outputs a Tensor with size (N, 6, 28, 28), and `s2` should not change the 6 into 16.
functionalizes functionalization
Co-authored-by: Alanna Burke <[email protected]> Co-authored-by: Svetlana Karslioglu <[email protected]>
Co-authored-by: Svetlana Karslioglu <[email protected]>
Co-authored-by: Alanna Burke <[email protected]> Co-authored-by: sekyondaMeta <[email protected]> Co-authored-by: Svetlana Karslioglu <[email protected]>
* Fix dead link of compiler troubleshooting --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
Signed-off-by: zzb610 <[email protected]> Co-authored-by: Svetlana Karslioglu <[email protected]>
* Update torch_export_tutorial.py --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
The prefactor m was incorrectly added to the vector-Jacobian product formula in the autograd tutorial. This change was based on the mistaken assumption that m-scaling was necessary to account for multiple terms in the summation. However, the chain rule and vector-Jacobian product formula already correctly aggregate contributions from all intermediate variables without requiring explicit scaling by m.
* make_distribute_tutorial work in google_colab --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
Co-authored-by: Svetlana Karslioglu <[email protected]>
…ytorch#3198) * Update C++ Custom Operators tutorial with note about AMD GPU. * Put in a NOTE call out box and change to AMD ROCm. --------- Co-authored-by: Svetlana Karslioglu <[email protected]>
…ue challenges and solutions (pytorch#3180) * Added a recipe for showcasing torch.export flow for 4 models. --------- Co-authored-by: Svetlana Karslioglu <[email protected]> Co-authored-by: Angela Yi <[email protected]>
Update recipe template link Co-authored-by: Svetlana Karslioglu <[email protected]>
Fix broken links in cpp_autograd.rst
Fixing broken links in RPC Tutorial
Fix link in pytorch-with-examples Co-authored-by: Svetlana Karslioglu <[email protected]>
* Update .lycheeignore Adding a link to ignore * Update .lycheeignore
Edit author link
Update links in dynamic quantization bert tutorial
Adding https://www.uber.com/blog/deep-neuroevolution/ to ignore, link valid but not passing
Update broken links
Update to the tuning guide and launch script links Co-authored-by: Mark Saroufim <[email protected]>
Link no longer exists so giving credit to creator instead
Update to utility link
* Pull 2.6 binaries * Update
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.
No description provided.