Clarify last round behavior of SNPE-A

Author: michaeldeistler

tutorials/16_implemented_methods.ipynb

@@ -66,10 +66,13 @@
     "\n",
     "inference = NPE_A(prior)\n",
     "proposal = prior\n",
-    "for _ in range(num_rounds):\n",
+    "for r in range(num_rounds):\n",
     "    theta = proposal.sample((num_sims,))\n",
     "    x = simulator(theta)\n",
-    "    _ = inference.append_simulations(theta, x, proposal=proposal).train()\n",
+    "    # NPE-A trains a Gaussian density estimator in all but the last round. In the last round,\n",
+    "    # it trains a mixture of Gaussians, which is why we have to pass the `final_round` flag.\n",
+    "    final_round = True if r == num_rounds - 1 else False\n",
+    "    _ = inference.append_simulations(theta, x, proposal=proposal).train(final_round=final_round)\n",
     "    posterior = inference.build_posterior().set_default_x(x_o)\n",
     "    proposal = posterior"
    ]
@@ -598,7 +601,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.14"
+   "version": "3.12.4"
   },
   "toc": {
    "base_numbering": 1,