Update

Dockerfile

@@ -21,7 +21,7 @@ RUN pip install --no-cache-dir --upgrade pip && \
 # CMD ["TRANSFORMERS_CACHE=./", "infinity_emb", "--model-name-or-path", "sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2", "--port", "7860"]
 
 # CMD ["sh", "-c", "HF_HOME=/tmp/cache", "python", "m3_server.py"]
-# ["sh", "-c", "'FOO=BAR python app.py'"]
+# CMD ["python", "m3_server.py"]
 
 # CMD ["sh", "start-m3-server.sh"]  # OK
-CMD ["python", "m3_server.py"]
+["sh", "-c", "'FOO=BAR python app.py'"]

m3_server.py

@@ -17,7 +17,7 @@ from starlette.status import HTTP_504_GATEWAY_TIMEOUT
 Path("/tmp/cache").mkdir(exist_ok=True)
 os.environ["HF_HOME"] = "/tmp/cache"
 os.environ["TRANSFORMERS_CACHE"] = "/tmp/cache"
-# does not quite work, need Path("/tmp/cache").mkdir(exist_ok=True)?
+# does not quite work
 
 batch_size = 2  # gpu batch_size in order of your available vram
 max_request = 10  # max request for future improvements on api calls / gpu batches (for now is pretty basic)