Merge pull request #10 from tencent-quantum-lab/master

local sync
tencent-quantum-lab · Aug 10, 2023 · 63cf700 · 63cf700
2 parents b18c046 + 1eef036
commit 63cf700
Show file tree

Hide file tree

Showing 2 changed files with 4,019 additions and 6 deletions.
diff --git a/docs/source/tutorials/nnvqe.ipynb b/docs/source/tutorials/nnvqe.ipynb
@@ -6,7 +6,7 @@
    "id": "64ba95d6",
    "metadata": {},
    "source": [
-    "# <center> NN-VQE"
+    "# <center> Neural Network encoded Variational Quantum Eigensolver (NN-VQE)"
    ]
   },
   {
@@ -105,7 +105,7 @@
    "source": [
     "## Ansatz circuit\n",
     "\n",
-    "Now we design the circuit. We choose multi-scale entangled renormalization ansatz (MERA) as the ansatz here, $d$ is the circuit depth. (see tutorial of MERA [here](https://tensorcircuit.readthedocs.io/en/latest/tutorials/mera.html))"
+    "Now we design the circuit. We choose multi-scale entangled renormalization ansatz (MERA) as the ansatz here, $d$ is the circuit depth. (see [MERA tutorial](https://tensorcircuit.readthedocs.io/en/latest/tutorials/mera.html))"
    ]
   },
   {
@@ -2909,7 +2909,7 @@
    "source": [
     "## NN-VQE\n",
     "\n",
-    "Design the NN-VQE. We use a neural network to transform the Hamiltonian parameters to the optimized parameters in the PQC for VQE."
+    "Design the NN-VQE. We use a neural network to transform the Hamiltonian parameters to the optimized parameters in the parameterized quantum circuit (PQC) for VQE."
    ]
   },
   {
@@ -3062,7 +3062,7 @@
     "test_delta = np.linspace(-4.0, 4.0, 201)  # test set\n",
     "test_energies = tf.zeros_like(test_delta).numpy()\n",
     "m = NN_MERA(n, d, lamb, NN_shape, stddev)\n",
-    "m.load_weights(\"DNN-MERA_2[20](-3.0,3.0,20)_drop05.weights.h5\")\n",
+    "m.load_weights(\"NN-VQE.weights.h5\")\n",
     "for i, de in tqdm(enumerate(test_delta)):\n",
     "    test_energies[i] = m(K.reshape(de, [1]))"
    ]
@@ -3074,7 +3074,7 @@
    "source": [
     "## Compare\n",
     "\n",
-    "We compare the results of NN-VQE with the analytical ones to calculate the ground-state relative error. From the figure, we can see that NN-VQE is able to estimate the ground-state energies of parameterized Hamiltonians with high precision without fine-tuning and has a favorable generalization capability."
+    "We compare the results of NN-VQE with the analytical ones to calculate the ground-state energy relative error. From the figure, we can see that NN-VQE is able to estimate the ground-state energies of parameterized Hamiltonians with high precision without fine-tuning and has a favorable generalization capability."
    ]
   },
   {
@@ -3982,7 +3982,7 @@
    "id": "5f9bda8a",
    "metadata": {},
    "source": [
-    "To get more detailed information or further study, please refer to [our paper](https://arxiv.org/abs/2308.01068) and [GitHub](https://github.com/JachyMeow/NN-VQA)."
+    "To get more detailed information or further study, please refer to our [paper](https://arxiv.org/abs/2308.01068) and [GitHub](https://github.com/JachyMeow/NN-VQA)."
    ]
   }
  ],