Comments (8)
Glad you found it useful! I'll put a SLURM tutorial together tomorrow.
In the meantime, check out the multi-node example:
from lightning.
Thanks. Looking forward to see SLURM tutorial.
from lightning.
Put the tutorial together here.
Try it out and lmk what you think.
https://link.medium.com/Bs6NxJyjRY
from lightning.
Closing this because the tutorial has been created :)
But, feel free to add a comment if the tutorial is missing something or needs clarification.
from lightning.
Closing this because the tutorial has been created :)
But, feel free to add a comment if the tutorial is missing something or needs clarification.
In the blog, there is a comment:
Disclaimer: This tutorial assumes your cluster is managed by SLURM.
In fact, I have no idea about how to manage cluster using slurm...
When cluster is managed by slurm, I can try PTL in my cluster.
from lightning.
SLURM is installed by your cluster administrator. I guess here I’m using cluster to refer to an academic or corporate cluster.
If it’s literally your own cluster with your own machines then PTL may not support that case because it uses NCCL to communicate and a few flags SLURM sets.
But if it’s an academic or corporate cluster, it most likely has SLURM installed.
If you use srun or sbatch to run jobs, then you’re using SLURM.
from lightning.
Yes, now I want to try slurm by myself...
from lightning.
I did not feel SLURM convenient than Horovod.
from lightning.
Related Issues (20)
- Setting a compression on `optimize` fails on data deserialization HOT 7
- The streaming S3Client is unable to access credentials in ~/.aws/config
- Allow flexible and easy to configure HSDP HOT 8
- Emulating multiple devices with a single GPU HOT 3
- Integrating a tool to debug neural networks more efficiently
- Don't save pretrained submodules in checkpoint HOT 3
- MisconfigurationException HOT 1
- Linear layer has no grad when I forward twice and the first time under `with torch.no_grad()` scope HOT 1
- Append data to pre-optimized dataset HOT 1
- Model stuck after saving a checkpoing when using the FSDPStrategy HOT 1
- Custom learning rate scheduler does not work HOT 2
- DeepSpeedStrategy integration HOT 2
- Model weights are being reset to zero on Windows HOT 8
- The prefetch_batches in _PrefetchDataFetcher always is 1 HOT 1
- Infinite recursion error in graveyard imports of MixedPrecision plugin
- NCCL when trying to train on 2 nodes HOT 1
- Multiple Event Files Generated in Version Folder Using TensorBoardLogger HOT 2
- how to set find_unused_parameters in Fabric? HOT 1
- Validation runs only for one iteration when restarting from checkpoint mid-epoch, wrongly reporting validation loss HOT 2
- Permission denied: '/usr/share/lmod/lmod/init/ksh_funcs/libcudart.so' HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from lightning.