From 8839ef0e3af0927701051f96c4bfbaf711fcb713 Mon Sep 17 00:00:00 2001
From: AJ Keller <aj@neurosity.co>
Date: Mon, 1 Jul 2024 00:50:20 +0000
Subject: [PATCH] ADD: final configs for first release

---
 scripts/download.sh       |  1 +
 scripts/preprocess.sh     |  0
 scripts/setup.sh          |  7 ++++++-
 scripts/train_parallel.sh | 11 ++++++-----
 src/eeg/utils.py          |  1 +
 5 files changed, 14 insertions(+), 6 deletions(-)
 create mode 100755 scripts/download.sh
 mode change 100644 => 100755 scripts/preprocess.sh

diff --git a/scripts/download.sh b/scripts/download.sh
new file mode 100755
index 0000000..a6a58b2
--- /dev/null
+++ b/scripts/download.sh
@@ -0,0 +1 @@
+rsync -auxvL --no-owner --no-group nedc-tuh-eeg@www.isip.piconepress.com:data/tuh_eeg/tuh_eeg/v2.0.1/ .
\ No newline at end of file
diff --git a/scripts/preprocess.sh b/scripts/preprocess.sh
old mode 100644
new mode 100755
diff --git a/scripts/setup.sh b/scripts/setup.sh
index 907a4ca..e4ddb8a 100755
--- a/scripts/setup.sh
+++ b/scripts/setup.sh
@@ -3,6 +3,11 @@
 # Print welcome message
 echo "Welcome to the setup script!"
 
+# Make data directory
+echo "Creating data directory..."
+mkdir -p data
+mkdir -p data/tuh_eeg
+
 # Ask user for their name to set git config user.name
 echo "Asking for your name to set git configuration..."
 read -p "Please enter your name: " name
@@ -23,7 +28,7 @@ apt-get update
 
 # Install tmux and htop for terminal multiplexing and process monitoring
 echo "Installing tmux and htop..."
-apt-get install -y tmux htop
+apt-get install -y tmux htop rsync
 
 # Generate a new SSH key using the provided email
 echo "Generating a new SSH key..."
diff --git a/scripts/train_parallel.sh b/scripts/train_parallel.sh
index e2f7442..e278b6b 100755
--- a/scripts/train_parallel.sh
+++ b/scripts/train_parallel.sh
@@ -1,11 +1,11 @@
-python -m torch.distributed.launch --nproc_per_node=2 \
+python -m torch.distributed.launch --nproc_per_node=8 \
     src/train_gpt.py \
     --training-steps=50000 \
-    --eval_every_n_steps=100 \
-    --log-every-n-steps=1 \
+    --eval_every_n_steps=100000 \
+    --log-every-n-steps=100 \
     --per-device-training-batch-size=32 \
     --per-device-validation-batch-size=32 \
-    --num-workers=16 \
+    --num-workers=32 \
     --num_chunks=32 \
     --chunk_len=500 \
     --chunk_ovlp=50 \
@@ -15,4 +15,5 @@ python -m torch.distributed.launch --nproc_per_node=2 \
     --training-style=CSM_causal \
     --embedding-dim=1024 \
     --train-data-path=data/npy_tuh_eeg \
-    --verbose=True
\ No newline at end of file
+    --verbose=True \
+    &> train_parallel.log
diff --git a/src/eeg/utils.py b/src/eeg/utils.py
index 01b4fbc..250167f 100644
--- a/src/eeg/utils.py
+++ b/src/eeg/utils.py
@@ -10,6 +10,7 @@
 
 # EEG_ALL_CHANNELS = sorted(list(set(EEG_10_20_CHANNELS + EEG_10_10_CHANNELS)))
 EEG_ALL_CHANNELS = sorted(list(set(EEG_10_20_CHANNELS)))
+NUM_CHANNELS = len(EEG_ALL_CHANNELS)
 
 def align_data_to_standard_channels(input_data, channel_locations):
     """