Fix: Correct paths in dataset TOML file and git lfs in start

start.sh

@@ -1,5 +1,5 @@
 #!/bin/bash
-# --- Final Definitive Startup Script (v23.7 - Re-adds necessary LoRA arguments) ---
+# --- Final Definitive Startup Script (v24 - Adds git lfs pull to fix num_samples=0) ---
 
 set -e
 echo "--- Startup Script Initialized ---"
@@ -15,28 +15,30 @@ VAE_PATH="$MODELS_DIR/Wan2.1_VAE.pth"
 CLIP_PATH="$MODELS_DIR/models_clip_open-clip-xlm-roberta-large-vit-huge-14.pth"
 T5_PATH="$MODELS_DIR/models_t5_umt5-xxl-enc-bf16.pth"
 
-echo "--- Checking for model files in persistent storage... ---"
+echo "--- Checking for model files... ---"
 if [ ! -f "$DIT_PATH" ]; then
-    echo "Downloading DiT model..."
     huggingface-cli download jujutechnology/WANfortraining wan2.1_i2v_720p_14B_fp8_e4m3fn.safetensors --local-dir $MODELS_DIR --local-dir-use-symlinks False
 fi
 if [ ! -f "$VAE_PATH" ]; then
-    echo "Downloading VAE model..."
     huggingface-cli download jujutechnology/WANfortraining Wan2.1_VAE.pth --local-dir $MODELS_DIR --local-dir-use-symlinks False
 fi
 if [ ! -f "$CLIP_PATH" ]; then
-    echo "Downloading CLIP model..."
     huggingface-cli download jujutechnology/WANfortraining models_clip_open-clip-xlm-roberta-large-vit-huge-14.pth --local-dir $MODELS_DIR --local-dir-use-symlinks False
 fi
 if [ ! -f "$T5_PATH" ]; then
-    echo "Downloading T5 model..."
     huggingface-cli download jujutechnology/WANfortraining models_t5_umt5-xxl-enc-bf16.pth --local-dir $MODELS_DIR --local-dir-use-symlinks False
 fi
+echo "--- Models are present. ---"
 
-echo "--- All models are present. Starting training... ---"
-ls -lh $MODELS_DIR
+# --- CRITICAL STEP: Force checkout of Git LFS files in the repo ---
+echo "--- Ensuring all dataset images are fully downloaded (git lfs pull)... ---"
+git lfs pull
+echo "--- LFS checkout complete. Verifying file sizes: ---"
+ls -lh /code/dataset/ebPhotos-001/ # This will now show megabyte-sized files
 
-# --- Run the training command with CORRECT LoRA arguments ---
+echo "--- Starting training... ---"
+
+# --- Run the training command ---
 accelerate launch wan_train_network.py \
     --task="i2v-14B" \
     --dit="$DIT_PATH" \