fix: finally made keras interface working

pyproject.toml

@@ -17,7 +17,7 @@ numba = "^0.59.0"
 tensorflow = { version = "^2.16.1", markers = "sys_platform == 'linux' or sys_platform == 'darwin'" }
 # TODO: the marker is a temporary solution due to the lack of the tensorflow-io 0.32.0's wheels for Windows, this package is one of
 # the tensorflow requirements
-qibo = "^0.2.6"
+qibo = {git="https://github.com/qiboteam/qibo", branch="qiboml_models_updates"}
 
 [tool.poetry.group.dev]
 optional = true

src/qiboml/models/encoding_decoding.py

@@ -45,7 +45,7 @@ def __post_init__(self):
             self.circuit.add(gates.RZ(q, theta=0.0))
 
     def forward(self, x: "ndarray") -> Circuit:
-        self.circuit.set_parameters(x)
+        self.circuit.set_parameters(x.ravel())
         return self.circuit
 
 
@@ -92,9 +92,14 @@ def __post_init__(self):
 
     def forward(self, x: Circuit) -> "ndarray":
         return self.observable.expectation_from_samples(
-            super().forward(x).frequencies()
+            super().forward(x).samples(),
+            input_samples=True,
         )
 
+    @property
+    def output_shape(self):
+        return (1,)
+
 
 """
    .

src/qiboml/models/keras.py

@@ -79,10 +79,16 @@ def __init__(self, layers: list[QuantumCircuitLayer]):
             )
 
     def call(self, x: "ndarray"):
-        for layer in self.layers[:-1]:
+        for layer in self.layers:
             x = layer.forward(x)
         return x
 
+    def compute_output_shape(self):
+        return self.layers[-1].output_shape
+
     @property
     def nqubits(self):
         return self.layers[0].circuit.nqubits
+
+    def __hash__(self):
+        return super().__hash__()

tutorials/model_example.ipynb

@@ -56,16 +56,16 @@
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "[Qibo 0.2.6|INFO|2024-06-20 12:12:39]: Using qibojit (numba) backend on /CPU:0\n"
+      "[Qibo 0.2.10|INFO|2024-07-07 13:00:30]: Using qibojit (numba) backend on /CPU:0\n"
      ]
     },
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "tf.Tensor(1.0, shape=(), dtype=float64)\n",
-      "tf.Tensor(1.0, shape=(), dtype=float64)\n",
-      "tf.Tensor(1.0, shape=(), dtype=float64)\n"
+      "0.486\n",
+      "0.528\n",
+      "0.526\n"
      ]
     }
    ],
@@ -93,9 +93,9 @@
       "> Model: Sequential(\n",
       "  (0): Linear(in_features=128, out_features=5, bias=True)\n",
       "  (1): Sigmoid()\n",
-      "  (2): BinaryEncodingLayer(nqubits=5, qubits=[0, 1, 2, 3, 4], circuit=<qibo.models.circuit.Circuit object at 0x35734a490>, initial_state=None, backend=tensorflow)\n",
-      "  (3): ReuploadingLayer(nqubits=5, qubits=(0, 2, 4), circuit=<qibo.models.circuit.Circuit object at 0x354a67990>, initial_state=None, backend=tensorflow)\n",
-      "  (4): QuantumDecodingLayer(nqubits=5, qubits=<range_iterator object at 0x357448d20>, circuit=<qibo.models.circuit.Circuit object at 0x35131c990>, initial_state=None, backend=tensorflow, nshots=1000)\n",
+      "  (2): BinaryEncodingLayer(nqubits=5, qubits=[0, 1, 2, 3, 4], circuit=<qibo.models.circuit.Circuit object at 0x36d7e9f10>, initial_state=None, backend=tensorflow)\n",
+      "  (3): ReuploadingLayer(nqubits=5, qubits=(0, 2, 4), circuit=<qibo.models.circuit.Circuit object at 0x35afcc650>, initial_state=None, backend=tensorflow)\n",
+      "  (4): QuantumDecodingLayer(nqubits=5, qubits=<range_iterator object at 0x36de9f120>, circuit=<qibo.models.circuit.Circuit object at 0x35fa6e5d0>, initial_state=None, backend=tensorflow, nshots=1000)\n",
       ")\n",
       "> Parameters\n",
       "0.weight torch.Size([5, 128])\n",
@@ -150,21 +150,10 @@
      "output_type": "stream",
      "text": [
       "> Model: <Sequential name=sequential, built=False>\n",
-      "> Outputs\n"
-     ]
-    },
-    {
-     "ename": "OperatorNotAllowedInGraphError",
-     "evalue": "Exception encountered when calling Model.call().\n\n\u001b[1mIterating over a symbolic `tf.Tensor` is not allowed. You can attempt the following resolutions to the problem: If you are running in Graph mode, use Eager execution mode or decorate this function with @tf.function. If you are using AutoGraph, you can try decorating this function with @tf.function. If that does not work, then you may be using an unsupported feature or your source code may not be visible to AutoGraph. See https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/autograph/g3doc/reference/limitations.md#access-to-source-code for more information.\u001b[0m\n\nArguments received by Model.call():\n  • args=('<KerasTensor shape=(1, 5), dtype=float32, sparse=False, name=keras_tensor_1>',)\n  • kwargs=<class 'inspect._empty'>",
-     "output_type": "error",
-     "traceback": [
-      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
-      "\u001b[0;31mOperatorNotAllowedInGraphError\u001b[0m            Traceback (most recent call last)",
-      "Cell \u001b[0;32mIn[5], line 31\u001b[0m\n\u001b[1;32m     29\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m x \u001b[38;5;129;01min\u001b[39;00m data:\n\u001b[1;32m     30\u001b[0m     x \u001b[38;5;241m=\u001b[39m tf\u001b[38;5;241m.\u001b[39mexpand_dims(x, axis\u001b[38;5;241m=\u001b[39m\u001b[38;5;241m0\u001b[39m)\n\u001b[0;32m---> 31\u001b[0m     \u001b[38;5;28mprint\u001b[39m(\u001b[43mmodel\u001b[49m\u001b[43m(\u001b[49m\u001b[43mx\u001b[49m\u001b[43m)\u001b[49m\u001b[38;5;241m.\u001b[39mprobabilities())\n",
-      "File \u001b[0;32m~/python_envs/qibo/lib/python3.11/site-packages/keras/src/utils/traceback_utils.py:122\u001b[0m, in \u001b[0;36mfilter_traceback.<locals>.error_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    119\u001b[0m     filtered_tb \u001b[38;5;241m=\u001b[39m _process_traceback_frames(e\u001b[38;5;241m.\u001b[39m__traceback__)\n\u001b[1;32m    120\u001b[0m     \u001b[38;5;66;03m# To get the full stack trace, call:\u001b[39;00m\n\u001b[1;32m    121\u001b[0m     \u001b[38;5;66;03m# `keras.config.disable_traceback_filtering()`\u001b[39;00m\n\u001b[0;32m--> 122\u001b[0m     \u001b[38;5;28;01mraise\u001b[39;00m e\u001b[38;5;241m.\u001b[39mwith_traceback(filtered_tb) \u001b[38;5;28;01mfrom\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m\n\u001b[1;32m    123\u001b[0m \u001b[38;5;28;01mfinally\u001b[39;00m:\n\u001b[1;32m    124\u001b[0m     \u001b[38;5;28;01mdel\u001b[39;00m filtered_tb\n",
-      "File \u001b[0;32m~/git/qiboml/src/qiboml/models/keras.py:77\u001b[0m, in \u001b[0;36mModel.call\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     75\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mcall\u001b[39m(\u001b[38;5;28mself\u001b[39m, x: \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mndarray\u001b[39m\u001b[38;5;124m\"\u001b[39m):\n\u001b[1;32m     76\u001b[0m     \u001b[38;5;28;01mfor\u001b[39;00m layer \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mlayers[:\u001b[38;5;241m-\u001b[39m\u001b[38;5;241m1\u001b[39m]:\n\u001b[0;32m---> 77\u001b[0m         x \u001b[38;5;241m=\u001b[39m \u001b[43mlayer\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mforward\u001b[49m\u001b[43m(\u001b[49m\u001b[43mx\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     78\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m x\n",
-      "File \u001b[0;32m~/git/qiboml/src/qiboml/models/encoding_decoding.py:30\u001b[0m, in \u001b[0;36mBinaryEncodingLayer.forward\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     28\u001b[0m circuit \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mcircuit\u001b[38;5;241m.\u001b[39mcopy()\n\u001b[1;32m     29\u001b[0m ones \u001b[38;5;241m=\u001b[39m x\u001b[38;5;241m.\u001b[39mravel() \u001b[38;5;241m==\u001b[39m \u001b[38;5;241m1\u001b[39m\n\u001b[0;32m---> 30\u001b[0m \u001b[43m\u001b[49m\u001b[38;5;28;43;01mfor\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[43mbit\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;129;43;01min\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[43mones\u001b[49m\u001b[43m:\u001b[49m\n\u001b[1;32m     31\u001b[0m \u001b[43m    \u001b[49m\u001b[43mcircuit\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43madd\u001b[49m\u001b[43m(\u001b[49m\u001b[43mgates\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mX\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mqubits\u001b[49m\u001b[43m[\u001b[49m\u001b[43mbit\u001b[49m\u001b[43m]\u001b[49m\u001b[43m)\u001b[49m\u001b[43m)\u001b[49m\n\u001b[1;32m     32\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m circuit\n",
-      "\u001b[0;31mOperatorNotAllowedInGraphError\u001b[0m: Exception encountered when calling Model.call().\n\n\u001b[1mIterating over a symbolic `tf.Tensor` is not allowed. You can attempt the following resolutions to the problem: If you are running in Graph mode, use Eager execution mode or decorate this function with @tf.function. If you are using AutoGraph, you can try decorating this function with @tf.function. If that does not work, then you may be using an unsupported feature or your source code may not be visible to AutoGraph. See https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/autograph/g3doc/reference/limitations.md#access-to-source-code for more information.\u001b[0m\n\nArguments received by Model.call():\n  • args=('<KerasTensor shape=(1, 5), dtype=float32, sparse=False, name=keras_tensor_1>',)\n  • kwargs=<class 'inspect._empty'>"
+      "> Outputs\n",
+      "0.40480000000000005\n",
+      "0.4228\n",
+      "0.40959999999999996\n"
      ]
     }
    ],
@@ -181,7 +170,7 @@
     "\n",
     "q_model = ks.Model(\n",
     "    layers = [\n",
-    "        ed.BinaryEncodingLayer(5),\n",
+    "        ed.PhaseEncodingLayer(5),\n",
     "        ans.ReuploadingLayer(nqubits=5, qubits=(0,2,4)),\n",
     "        ed.ExpectationLayer(nqubits=5, qubits=reversed(range(5)), observable=observable)\n",
     "    ]\n",
@@ -199,16 +188,14 @@
     "print(\"> Outputs\")\n",
     "for x in data:\n",
     "    x = tf.expand_dims(x, axis=0)\n",
-    "    print(model(x).probabilities())\n",
+    "    print(model(x))\n",
     "\n"
    ]
   },
   {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "71bc58a6-bddd-4397-b402-895574989ebc",
+   "cell_type": "markdown",
+   "id": "b166d1b6-77d2-4858-843d-9cb756184c29",
    "metadata": {},
-   "outputs": [],
    "source": []
   }
  ],