Bionemo demos (#2968)

* updated bionemo demos to v1.8

* cleaned demos outputs for clarity

* added linces and fixed naming README

* fixed license headers and readme hyperlink

* black fixing code

* isort and flake8 fixes

* addressing PR changes

* removed unrequired infer copy file

* updated other runs files/configs and fixed path in downstream notebook

* fixed fedavg max_epochs setting to 1, removed extra data in taps yamls, fixed column used for each site

* changed fedavg* and local* yamls to have site specific data for tap. for sabdab fedavg changed to original ???

* also sabdab local changed dataset.train config to ???

* tap fix configurations

* update nb, nvflare version

* use strict false

* use full data for central training of sabdab

---------

Co-authored-by: Chester Chen <512707+chesterxgchen@users.noreply.github.com>
Co-authored-by: Holger Roth <6304754+holgerroth@users.noreply.github.com>
Co-authored-by: Holger Roth <hroth@nvidia.com>

examples/advanced/bionemo/README.md

@@ -4,39 +4,25 @@
 
 This directory contains examples of running BioNeMo in a federated learning environment using [NVFlare](https://github.com/NVIDIA/NVFlare).
 
-1. The [task_fitting](./task_fitting/README.md) example includes a notebook that shows how to obtain protein learned representations in the form of embeddings using the ESM-1nv pre-trained model. 
+## Notebooks
+
+In this repo you will find two notebooks under the `task_fitting` and `downstream` folders respectively: 
+1. The [task_fitting](./task_fitting/task_fitting.ipynb) notebook example includes a notebook that shows how to obtain protein learned representations in the form of embeddings using the ESM-1nv pre-trained model. 
 The model is trained with NVIDIA's BioNeMo framework for Large Language Model training and inference.
-2. The [downstream](./downstream/README.md) example shows three different downstream tasks for fine-tuning a BioNeMo ESM-style model.
+2. The [downstream](./downstream/downstream_nvflare.ipynb) notebook example shows three different downstream tasks for fine-tuning a BioNeMo ESM-style model.
 
 ## Requirements
 
 Download and run the [BioNeMo docker container](https://catalog.ngc.nvidia.com/orgs/nvidia/teams/clara/containers/bionemo-framework).
-> **Note:** The examples here were tested with `nvcr.io/nvidia/clara/bionemo-framework:1.0`
+> **Note:** The examples here were tested with `nvcr.io/nvidia/clara/bionemo-framework:1.8`
 
-We recommend following the [Quickstart Guide](https://docs.nvidia.com/bionemo-framework/latest/quickstart-fw.html#docker-container-access) 
+We recommend following the [Quickstart Guide](https://docs.nvidia.com/bionemo-framework/latest/access-startup.html?highlight=docker) 
 on how to get the BioNeMo container. 
 
-First, copy the NeMo code to a local directory and configure the launch script so that downloaded models can be reused 
-```commandline
-CONTAINER="nvcr.io/nvidia/clara/bionemo-framework:1.0"
-DEST_PATH="."
-CONTAINER_NAME=bionemo
-docker run --name $CONTAINER_NAME -itd --rm $CONTAINER bash
-docker cp $CONTAINER_NAME:/opt/nvidia/bionemo $DEST_PATH
-docker kill $CONTAINER_NAME
-```
-
-Next, download the pre-trained models using
-```commandline
-cd ./bionemo
-./launch.sh download
-cd ..
-```
-
-Then, start the container and Jupyter Lab to run the NVFlare experiments with NVFlare using
+Start the container and Jupyter Lab to run the NVFlare experiments with NVFlare using
 ```commandline
 ./start_bionemo.sh
 ```
 It will start Jupyter Lab at `http://hostname:8888`.
 
-For information about how to get started with BioNeMo refer to the [documentation](https://docs.nvidia.com/bionemo-framework/latest).
+For information about how to get started with BioNeMo refer to the [documentation](https://docs.nvidia.com/bionemo-framework/latest).