Comments (5)
Another way to go about doing this is to sync file artifacts across nodes/locally, and then the user can either launch tensorboard locally or from a cpunode/gpunode. Services like tensorboard/jupyter should be launched interactively rather than as a part of the training task, imo.
from skypilot.
ParTask should be the best solution, if the base tasks are all single-node:
train = sky.Task(run='python train.py').set_resources(sky.Resources('V100'))
tensorboard = sky.Task(run='tensorboard --logdir .')
par_task = sky.ParTask([train, tensorboard]).set_resources(sky.Resources('V100'))
Any reason you wouldn't want to use this @Michaelvll?
Another way to go about doing this is to sync file artifacts across nodes/locally, and then the user can either launch tensorboard locally or from a cpunode/gpunode.
The default mode should be serving those remotely. Sync-to-local should be the optional mode. Event files/logs can be large and syncing to local is not always desirable.
Services like tensorboard/jupyter should be launched interactively rather than as a part of the training task, imo.
Agreed.
from skypilot.
BTW: This is a very good example to have!
from skypilot.
ParTask should be the best solution, if the base tasks are all single-node:
train = sky.Task(run='python train.py').set_resources(sky.Resources('V100')) tensorboard = sky.Task(run='tensorboard --logdir .') par_task = sky.ParTask([train, tensorboard]).set_resources(sky.Resources('V100'))Any reason you wouldn't want to use this @Michaelvll?
Another way to go about doing this is to sync file artifacts across nodes/locally, and then the user can either launch tensorboard locally or from a cpunode/gpunode.
The default mode should be serving those remotely. Sync-to-local should be the optional mode. Event files/logs can be large and syncing to local is not always desirable.
Services like tensorboard/jupyter should be launched interactively rather than as a part of the training task, imo.
Agreed.
In my understanding, the ParTask may implicitly run the two tasks sequentially as it shows in multi-echo example when the resources are not enough. Will it guarantee the two task runs at the same time? Also, I found it not possible to specify the tensorboard to be run on the head node (not the workers) when we have num_node > 1 with ParTask.
from skypilot.
Will it guarantee the two task runs at the same time?
Yes, if the resource demands of both tasks are smaller than the VM resources. It is the same with Ray's semantics.
For example, train
will use {V100:1, CPU:1}
and tensorboard
will use {CPU:1}
(Ray semantics), while the ParTask grabbed a VM with n CPUs and {V100:1}
. So they definitely should be run in parallel.
Also, I found it not possible to specify the tensorboard to be run on the head node (not the workers) when we have num_node > 1 with ParTask.
Actually, the ParTask backend doesn't support any base task with num_nodes > 1 at the moment. You should have run into an assert for that.
from skypilot.
Related Issues (20)
- [k8s] SkyPilot overrides `GIT_SSH_COMMAND` environment variable HOT 2
- [docker] Custom environment variables defined in docker image are not populated
- [Serve] Runpod resources.ports not supported
- [Core] User changing default env will make job submission fail
- [Serve] GCP crendential path error with docker image and replica HOT 16
- [GCP] Instance creation fail when auto reservations exist
- [Docs/Provisioner] Guide to minimize setup overheads by the use of Docker image
- [core] `sky launch` fails on k8s with `bash: /opt/conda/bin/ray: No such file or directory`
- [Tests] GCP Image tests failed on latest master HOT 4
- [Core] Support image id when using docker as runtime environment HOT 1
- Central coordination for multiple skypilot cli users HOT 1
- Examples: add examples for Triton, TensorRT-LLM
- [k8s example] Add example for putting `~/.sky` on a persistent volume HOT 1
- [Storage] Investigate `rclone mount` with VFS caching HOT 1
- `file_mount` with `mode: COPY` slower than expected on Google Cloud Storage
- [Serve] Service update scales to zero unexpectedly? HOT 1
- [SkyServe] : API Authentication Options, HTTPS, More Stable Web Server that http serve HOT 1
- [Core][Controller] Respect region/zone settings in controller resources when creating controller
- [Tests] Add smoke tests for AI gallery
- AWS Serving models insufficient permissions of skypilot role 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 skypilot.