Update PyTorch example for Ubuntu 24.04

pytorch/Makefile

@@ -3,6 +3,14 @@
 
 # PyTorch and the pre-trained model must be installed on the system. See README for details.
 
+SHELL := /bin/bash
+
+THIS_DIR := $(shell pwd)
+
+VENV_DIR ?= $(THIS_DIR)/my_venv
+
+ENTRYPOINT := $(VENV_DIR)/bin/python3
+
 ARCH_LIBDIR ?= /lib/$(shell $(CC) -dumpmachine)
 
 ifeq ($(DEBUG),1)
@@ -21,7 +29,8 @@ pytorch.manifest: pytorch.manifest.template
 	gramine-manifest \
 		-Dlog_level=$(GRAMINE_LOG_LEVEL) \
 		-Darch_libdir=$(ARCH_LIBDIR) \
-		-Dentrypoint=$(realpath $(shell sh -c "command -v python3")) \
+		-Dentrypoint=$(ENTRYPOINT) \
+		-Dvenv_dir=$(VENV_DIR) \
 		$< > $@
 
 # Make on Ubuntu <= 20.04 doesn't support "Rules with Grouped Targets" (`&:`),

pytorch/README.md

@@ -25,6 +25,7 @@ installation.
 - `sudo apt install python3-pip lsb-release` to install `pip` and `lsb_release`.
   The former is required to install additional Python packages while the latter
   is used by the Makefile.
+- `python3 -m venv my_venv && source my_venv/bin/activate` to create and activate a virtual environment.
 - `pip3 install --user torchvision pillow` to install the torchvision and pillow
   Python packages and their dependencies (usually in $HOME/.local). WARNING:
   This downloads several hundred megabytes of data!

pytorch/pytorch.manifest.template

@@ -22,6 +22,7 @@ fs.mounts = [
   { path = "/usr/lib", uri = "file:/usr/lib" },
   { path = "{{ arch_libdir }}", uri = "file:{{ arch_libdir }}" },
   { path = "/usr/{{ arch_libdir }}", uri = "file:/usr/{{ arch_libdir }}" },
+  { path = "{{ venv_dir }}", uri = "file:{{ venv_dir }}" },
 {% for path in python.get_sys_path(entrypoint) %}
   { path = "{{ path }}", uri = "file:{{ path }}" },
 {% endfor %}
@@ -43,6 +44,7 @@ sgx.trusted_files = [
   "file:/usr/lib/",
   "file:{{ arch_libdir }}/",
   "file:/usr/{{ arch_libdir }}/",
+  "file:{{ venv_dir }}/",
 {% for path in python.get_sys_path(entrypoint) %}
   "file:{{ path }}{{ '/' if path.is_dir() else '' }}",
 {% endfor %}