You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fix distributed processing and various fixes (#89)
* simplified process_distributed.sh
* removed useless parameter --mmore-folder
* fixed distributed processing
* by default, processing should still not be distributed
* Fixed documentation of distributed processing
* by default, processing should not be using the fast mode
* removed the acknowledgements line
* small logic fixes
* reformatting with black and deleted leftover TODOs
* sorting imports
* changed docs with pyment
Copy file name to clipboardExpand all lines: README.md
-4Lines changed: 0 additions & 4 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -129,7 +129,3 @@ Don't hesitate to star the project :star: if you find it interesting! (you would
129
129
## License
130
130
131
131
This project is licensed under the Apache 2.0 License, see the [LICENSE :mortar_board:](LICENSE) file for details.
132
-
133
-
## Acknowledgements
134
-
135
-
This project is part of the [**OpenMeditron**](https://huggingface.co/OpenMeditron) initiative developed in [LiGHT](https://www.light-laboratory.org/) lab at EPFL/Yale/CMU Africa in collaboration with the [**SwissAI**](https://www.swiss-ai.org/) initiative. Thank you Scott Mahoney, Mary-Anne Hartley
Replace `rank 1` with a unique rank number for each node (1, 2, 3, etc.).
70
+
Replace `rank 1` with a unique rank number for each node (1, 2, 3, etc.). The node should be ready in a matter of 5 seconds.
71
71
72
72
#### Step 3: Begin Processing
73
73
74
-
Once all nodes are running, return to the master node and type `go` when prompted to start the processing.
74
+
Once all nodes are running, return to the master node and type `go`. The master node proceeds to crawl the input folder, split the workload among connected nodes and make them start their work.
75
+
76
+
The dask server will be automatically shut down by the master node at the end of the processing. This will also shut down the dask workers on all the connected nodes.
0 commit comments