v11: JOB_ID for multi-VPS parallel training, auto-detect NVMe

job_ml.sh

@@ -18,12 +18,13 @@ set -eo pipefail
 
 export HF_TOKEN="${HF_TOKEN:?Set HF_TOKEN}"
 export TRAIN_STAGE="${TRAIN_STAGE:-flux}"
+export JOB_ID="${JOB_ID:-$(hostname | cut -c1-8)}"
 export HF_USER="memoryai"
 export PROJECT_DIR="/home/adminuser/chungcat"
 export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
 export PYTHONPATH="$PROJECT_DIR:${PYTHONPATH:-}"
 
-log() { echo "[$(date +%H:%M:%S)] $1"; }
+log() { echo "[$(date +%H:%M:%S)][$JOB_ID] $1"; }
 
 ###############################################################################
 backup_checkpoint() {
@@ -38,7 +39,7 @@ user = os.environ.get("HF_USER", "memoryai")
 repo_id = f"{user}/4k-image-model-checkpoints"
 
 if stage == "flux":
-    ckpt_dir = Path("/data0/checkpoints/flux_lora")
+    ckpt_dir = Path(f"/data0/checkpoints/flux_lora_{os.environ.get('JOB_ID', 'default')}")
 elif stage == "sr2":
     ckpt_dir = Path("/home/adminuser/chungcat/checkpoints/sr_stage2")
 else:
@@ -62,7 +63,7 @@ to_upload = final if final.exists() else (checkpoints[-1] if checkpoints else No
 
 if to_upload:
     path_in_repo = f"{ckpt_dir.name}/{to_upload.name}"
-    print(f"  Uploading {to_upload.name} -> {repo_id}/{path_in_repo}")
+    print(f"  Uploading {to_upload.name} -> {repo_id}/{path_in_repo} (job: {os.environ.get('JOB_ID', 'default')})")
     upload_folder(folder_path=str(to_upload), repo_id=repo_id, path_in_repo=path_in_repo, repo_type="model")
     print("  Done!")
 else:
@@ -322,11 +323,11 @@ log "  Auto-backup started (PID $BACKUP_PID, every 30min)"
 
 # Train using train_flux_lora.py (encodes on-the-fly, no precompute needed)
 # GPU 0: VAE + text encoders, GPU 1: transformer training
-log "  Starting Flux LoRA training (2 GPU split, encode on-the-fly)..."
+log "  Starting Flux LoRA training (job: $JOB_ID)..."
 python3 -u "$PROJECT_DIR/scripts/training/train_flux_lora.py" \
     --model-name "black-forest-labs/FLUX.1-schnell" \
     --data-dir "$SHARD_DIR" \
-    --output-dir "$DATA0/checkpoints/flux_lora" \
+    --output-dir "$DATA0/checkpoints/flux_lora_$JOB_ID" \
     --batch-size 1 \
     --gradient-accumulation 8 \
     --learning-rate 1e-4 \