clarify: Colab T4 runs on GPU via PyTorch (not CPU)

notebooks/show4dstem_colab.ipynb

@@ -9,7 +9,7 @@
     "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/?url=https%3A%2F%2Fhuggingface.co%2Fdatasets%2Fbobleesj%2Fquantem-data%2Fresolve%2Fmain%2Fnotebooks%2Fshow4dstem_colab.ipynb)\n",
     "\n",
     "Open a 4D-STEM dataset in your browser - **no install, no GPU setup**. Runs on Colab's\n",
-    "free runtime. Backend auto-detects; this sample uses the CPU/numpy path so it works on\n",
+    "Open a 4D-STEM dataset in your browser - **no install, no GPU setup**. On Colab's free **Tesla T4**, Show4DSTEM runs on the **GPU via PyTorch (CUDA)** - NumPy is only the input format; it is uploaded to the GPU. (On a CPU-only runtime it falls back to CPU.)\n",
     "any runtime (CPU or GPU).\n"
    ]
   },