Merge pull request #373 from particle1331/dev

PDM add libraries / update versions
particle1331 · Jun 16, 2024 · 68f8c5b · 68f8c5b
2 parents 1ccff5b + 736704a
commit 68f8c5b
Show file tree

Hide file tree

Showing 6 changed files with 257 additions and 92 deletions.
diff --git a/.gitattributes b/.gitattributes
diff --git a/README.md b/README.md
@@ -38,8 +38,6 @@ created using [`pdm`](https://github.com/pdm-project/pdm)
 
 ```
 pip install -U pdm
-pdm venv create
-pdm venv activate
 pdm install
 ```
 
@@ -50,6 +48,7 @@ The following libraries (specified in `pdm.lock`) will be installed:
 │ fastapi                            │ 0.111.0        │
 │ Flask                              │ 3.0.3          │
 │ keras                              │ 2.15.0         │
+│ lightning                          │ 2.3.0          │
 │ matplotlib                         │ 3.9.0          │
 │ mlflow                             │ 2.13.2         │
 │ numpy                              │ 1.26.4         │
@@ -61,11 +60,11 @@ The following libraries (specified in `pdm.lock`) will be installed:
 │ tensorflow                         │ 2.15.1         │
 │ tensorflow-datasets                │ 4.9.6          │
 │ tensorflow-estimator               │ 2.15.0         │
-│ torch                              │ 2.3.1          │
-│ torchaudio                         │ 2.3.1          │
+│ torch                              │ 2.2.2          │
+│ torchaudio                         │ 2.2.2          │
 │ torchinfo                          │ 1.8.0          │
 │ torchmetrics                       │ 1.4.0.post0    │
-│ torchvision                        │ 0.18.1         │
+│ torchvision                        │ 0.17.2         │
 │ uvicorn                            │ 0.30.1         │
 │ xgboost                            │ 2.0.3          │
 ╰────────────────────────────────────┴────────────────╯

diff --git a/docs/intro.md b/docs/intro.md
@@ -35,6 +35,7 @@ Effect of batch normalization on the magnitude of preactivation gradients.
 │ fastapi                            │ 0.111.0        │
 │ Flask                              │ 3.0.3          │
 │ keras                              │ 2.15.0         │
+│ lightning                          │ 2.3.0          │
 │ matplotlib                         │ 3.9.0          │
 │ mlflow                             │ 2.13.2         │
 │ numpy                              │ 1.26.4         │
@@ -46,11 +47,11 @@ Effect of batch normalization on the magnitude of preactivation gradients.
 │ tensorflow                         │ 2.15.1         │
 │ tensorflow-datasets                │ 4.9.6          │
 │ tensorflow-estimator               │ 2.15.0         │
-│ torch                              │ 2.3.1          │
-│ torchaudio                         │ 2.3.1          │
+│ torch                              │ 2.2.2          │
+│ torchaudio                         │ 2.2.2          │
 │ torchinfo                          │ 1.8.0          │
 │ torchmetrics                       │ 1.4.0.post0    │
-│ torchvision                        │ 0.18.1         │
+│ torchvision                        │ 0.17.2         │
 │ uvicorn                            │ 0.30.1         │
 │ xgboost                            │ 2.0.3          │
 ╰────────────────────────────────────┴────────────────╯

diff --git a/docs/nb/notes/containers.ipynb b/docs/nb/notes/containers.ipynb
@@ -795,9 +795,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "This section deals with quality of life improvement for developing with Docker. For example, any changes to `main.py` will not affect the running app (which is isolated). This can make development difficult with time-consuming rebuilds. \n",
-    "\n",
-    "Also, if we have a large files or builds that will be used for development, then we will have to copy or rebuild this each time the image is built. Our local IDE also does not have features like autocomplete and will generally compain of missing packages not installed in the local environment."
+    "This section deals with quality of life improvement for developing with Docker. For example, any local changes to `main.py` will not affect the correspondding script in the container. This can make development difficult with time-consuming rebuilds, e.g. with large images or large files. Finally, our local IDE will not have features like autocomplete and will generally compain of missing packages."
    ]
   },
   {
@@ -809,7 +807,7 @@
     "name: vs-code-module-not-found\n",
     "width: 1200px\n",
     "---\n",
-    "Module not found. One solution is to install the requirements in a virtual env. But this is not ideal if you want straightforward reproducibility.\n",
+    "Module not found. One solution is to install the requirements in a local virtual env. But this is not ideal if you want straightforward reproducibility.\n",
     "```"
    ]
   },
@@ -1068,7 +1066,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "**Note:** *Refer to the following section on docker compose application used here.*\n",
+    "**Note:** *Refer to the following section on docker compose for the application used here.*\n",
     "\n",
     "For debugging, we use https://github.com/microsoft/debugpy. Run the application defined in the following compose file. This overrides the Dockerfile entrypoint and simply runs the uvicorn server with debugpy client listening on port 5678."
    ]