Tanmoy Sarkar

Make `save_hyperparamters` not save another class's args when `LightningModule` instantiated within another class's constructor

10

## Duplicate args conflict with `.save_hyperparameters()` Simple hack to fix the issue https://github.com/PyTorchLightning/pytorch-lightning/issues/13181 by @[serena-ruan](https://github.com/serena-ruan) quick test ```python3 class Mod1(LightningModule): def __init__(self, same_arg): super().__init__() self.same_arg = same_arg self.save_hyperparameters() class Mod2(LightningModule):...

Continued ViT support

6

## 🚀 Feature Lightning-transformers is still nlp specific, I want to work on ViT support. @SeanNaren

Multi worker inference in Databricks

1

Multi worker training in horovod is great. But I am facing problem with inference specifically in databricks. Is there any way to do inference in databricks. And how to accumulate...

Only one gpu is getting utilized for Multi gpu search

1

# Summary While running multigpu search its only utilizing one gpu compute, but I can see the VRAM usage for other gpus Running on: - [ ] CPU - [x]...

Cucat Featurization base

8

#### Starter script ```python3 import pandas as pd import cudf import graphistry df = pd.read_csv('https://gist.githubusercontent.com/silkspace/c7b50d0c03dc59f63c48d68d696958ff/raw/31d918267f86f8252d42d2e9597ba6fc03fcdac2/redteam_50k.csv', index_col=0) red_team = pd.read_csv('https://gist.githubusercontent.com/silkspace/5cf5a94b9ac4b4ffe38904f20d93edb1/raw/888dabd86f88ea747cf9ff5f6c44725e21536465/redteam_labels.csv', index_col=0) df['feats'] = df.src_computer + ' ' + df.dst_computer + '...

[FEA] Smart dependency manager to replace lazy imports

3

#### Motivations: - Sometimes modules are getting imported several times through lazy import - Not easy to manage dependencies from file to file This is how we are managing dependencies...