Ashley Wright
commited on
Commit
•
91b4a85
1
Parent(s):
928edd4
Add ready marker, add README.md
Browse files- README.md +10 -0
- src/main.py +2 -0
README.md
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# edge-maxxing-newdream-sdxl
|
2 |
+
|
3 |
+
This holds the baseline for the SDXL Nvidia GeForce RTX 4090 contest, which can be forked freely and optimized
|
4 |
+
|
5 |
+
Some recommendations are as follows:
|
6 |
+
- Installing dependencies should be done in pyproject.toml, including git dependencies
|
7 |
+
- Compiled models should be included directly in the repository(rather than compiling during loading), loading time matters far more than file sizes
|
8 |
+
- Avoid changing `src/main.py`, as that includes mostly protocol logic. Most changes should be in `models` and `src/pipeline.py`
|
9 |
+
|
10 |
+
For testing, you need a docker container with pytorch and ubuntu 22.04, you can download your listed dependencies with `pip install -e .`, and then running `src/main.py`
|
src/main.py
CHANGED
@@ -19,6 +19,8 @@ def main():
|
|
19 |
connection, _ = inference_socket.accept()
|
20 |
|
21 |
with connection:
|
|
|
|
|
22 |
while True:
|
23 |
size = int.from_bytes(connection.recv(2), byteorder)
|
24 |
|
|
|
19 |
connection, _ = inference_socket.accept()
|
20 |
|
21 |
with connection:
|
22 |
+
connection.send(b'\xFF') # Ready marker
|
23 |
+
|
24 |
while True:
|
25 |
size = int.from_bytes(connection.recv(2), byteorder)
|
26 |
|