Comments (10)
Sounds great. can you send me a pm on twitter? we can coordinate through there _willfalcon
from lightning.
Seems like on twitter both the sides have to follow each other in order to share PMs. Any other platform?
from lightning.
good point haha. didn't know that about twitter. I've added you anyhow, but let's just chat here.
First step to helping is for you to do the following things and document your experience such as what was hard, what was not intuitive, what documentation would be helpful, etc.
- Build and run the demo.
- Build a simple classifier on your own using this.
From there we can see where the gaps are in usability and documentation. Then we can make a dev plan to get those things fixed in the library.
How does that sound? What's a good ETA to get that done?
Thanks!
from lightning.
Yes. I will do these after this week only. (I am too busy this week)
Anyways, the things that majorly concern me is the organisation of the package. You can have a look at my other library (Not related to DL) EinsteinPy, but you can have a look at how extensively its tested etc.
From there we can see where the gaps are in usability and documentation. Then we can make a dev plan to get those things fixed in the library.
For this plan, I will get back to you this Tuesday?
from lightning.
Sounds great. Yeah, this is a research library, so I didn't write tests for it because I've been moving quickly through research for publication. Tests would be great! I have however tested it myself when using/coding it, just didn't write an explicit test out.
Writing code for research has a less rigorous approach than code for a production system haha.
from lightning.
I have however tested it myself when using/coding it, just didn't write an explicit test out.
No issues! I would love to help. I will make some organisational changes in a PR, for making docs and CI easy! Is that fine?
from lightning.
perfect. Looking forward!
from lightning.
One question. Why is the package name pytorch_lightning
somewhere and pytorch-lightning
somewhere else?
from lightning.
Oh! I am sorry, the tuesday thing got slipped off my mind. Will do this shortly :)
from lightning.
Hi! lighting has been updated to include: tensorboard visualization, multi-gpu training, multi-node training, thoroughly documented and updated clean syntax.
Would be auper helpful if you tried version 0.21 and provided feedback!
Thanks!
from lightning.
Related Issues (20)
- torch.hstack() will cause the lose of the gradient flow track
- Flag `use_distributed_sampler` is not work HOT 2
- test_ddp.py hangs at test_ddp_configure_ddp
- UI issue with Documentation HOT 2
- Introduce sharded checkpointing for NO_SHARD FSDP. HOT 6
- LightningModule.train_dataloader not being called HOT 2
- CUDA memory increase (caused CUDA OOM) when saving checkpoint at the `train_epoch_end` HOT 2
- Wrong direction url in example HOT 1
- neptune.ai logger produces lots of errors when logging "training/epoch"
- `self.all_gather` used in `on_training_epoch_end` reports `RuntimeError`
- CLI trainer_defaults logger leads to error in config.yaml HOT 6
- LightningCLI cannot use strategy and accelerator class instances. HOT 1
- Correctly specifying max_steps when accumulate_grad_batches > 1 HOT 1
- Address FSDP + manual optimization HOT 1
- `dirpath` in `OnExceptionCheckpoint` should default to the same directory as `ModelCheckpoint`
- Allow overriding the initial batch/step index, for when resuming training.
- Manual optimisation and AMP HOT 1
- LightningCLI docs example for EarlyStopping missing required args in config file HOT 2
- Dump prof files from AdvancedProfiler
- Modules with `nn.Parameter` not Converted by Lightning Mixed Precision 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.