alibaba / alibaba/BladeDISC

Survey DISC compilation on TorchScript graphs that are dumped from LazyTensor.

Open
#222 2 comments 0 reactions 0 assignees View on GitHub
TorchDisc training
Dominant language
C++
Stars
933
Forks
169
PR merge metrics
No merged PRs in 30d

Description

We are capable to dump the TorchScript computation graph tracing via LazyTensor now, including backward and gradient updating subgraphs. With the full graph dumped we can reuse the DISC compilation flow used in inference.

This issue is the placeholder for all the related tasks. Actions:
- [ ] Add utilities to help dump TorchScript computation graph traced by LazyTensor https://github.com/alibaba/BladeDISC/pull/286
- [ ] Dump TSBackend graphs and inputs for debugging and survey

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reviewing the existing LazyTensor TorchScript graph-dumping work in PR #286 and the issue's two action items. Inspect how TSBackend graphs and inputs are currently produced for debugging; done means the required dump utilities exist and TSBackend graphs and inputs can be surveyed for compilation.

Written by the indexing model from the issue text.

Assessment

Tech stack
pytorch
Domain
compilers, machine-learning
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.