last updates/fixes before paper

Author: kmheckel

docs/conf.py

@@ -14,7 +14,7 @@
 project = 'Spyx'
 copyright = '2023, Kade Heckel'
 author = 'Kade Heckel'
-release = 'v0.1.18'
+release = 'v0.1.19'
 
 # -- General configuration ---------------------------------------------------
 # https://www.sphinx-doc.org/en/master/usage/configuration.html#general-configuration

docs/examples/surrogate_gradient/SurrogateGradientTutorial.ipynb

@@ -153,7 +153,7 @@
       },
       "outputs": [],
       "source": [
-        "shd_dl = spyx.data.SHD_loader(256,128,128)"
+        "shd_dl = spyx.loaders.SHD_loader(256,128,128)"
       ]
     },
     {
@@ -259,9 +259,9 @@
         "\n",
         "    # Haiku has the ability to stack multiple layers/recurrent modules as one entity\n",
         "    core = hk.DeepRNN([\n",
-        "        snn.LIF((64,), activation=spyx.axn.Axon(spyx.axn.triangular())), #LIF neuron layer with triangular activation\n",
+        "        snn.LIF((64,), activation=spyx.axn.triangular()), #LIF neuron layer with triangular activation\n",
         "        hk.Linear(64, with_bias=False),\n",
-        "        snn.LIF((64,), activation=spyx.axn.Axon(spyx.axn.triangular())),\n",
+        "        snn.LIF((64,), activation=spyx.axn.triangular()),\n",
         "        hk.Linear(20, with_bias=False),\n",
         "        snn.LI((20,)) # Non-spiking final layer\n",
         "    ])\n",
@@ -312,6 +312,9 @@
         "    # We use optax for our optimizer.\n",
         "    opt = optax.lion(learning_rate=schedule)\n",
         "\n",
+        "    Loss = spyx.fn.integral_crossentropy()\n",
+        "    Acc = spyx.fn.integral_accuracy()\n",
+        "\n",
         "    # create and initialize the optimizer\n",
         "    opt_state = opt.init(params)\n",
         "    grad_params = params\n",
@@ -321,7 +324,7 @@
         "    def net_eval(weights, events, targets):\n",
         "        readout = SNN.apply(weights, events)\n",
         "        traces, V_f = readout\n",
-        "        return spyx.fn.integral_crossentropy(traces, targets)\n",
+        "        return Loss(traces, targets)\n",
         "\n",
         "    # Use JAX to create a function that calculates the loss and the gradient!\n",
         "    surrogate_grad = jax.value_and_grad(net_eval)\n",
@@ -357,8 +360,8 @@
         "        # unpack the final layer outputs and end state of each SNN layer\n",
         "        traces, V_f = readout\n",
         "        # compute accuracy, predictions, and loss\n",
-        "        acc, pred = spyx.fn.integral_accuracy(traces, targets)\n",
-        "        loss = spyx.fn.integral_crossentropy(traces, targets)\n",
+        "        acc, pred = Acc(traces, targets)\n",
+        "        loss = Loss(traces, targets)\n",
         "        # we return the parameters here because of how jax.scan is structured.\n",
         "        return grad_params, jnp.array([acc, loss])\n",
         "\n",
@@ -421,14 +424,17 @@
       "source": [
         "def test_gd(SNN, params, dl):\n",
         "\n",
+        "    Loss = spyx.fn.integral_crossentropy()\n",
+        "    Acc = spyx.fn.integral_accuracy()\n",
+        "\n",
         "    @jax.jit\n",
         "    def test_step(params, data):\n",
         "        events, targets = data\n",
         "        events = jnp.unpackbits(events, axis=1)\n",
         "        readout = SNN.apply(params, events)\n",
         "        traces, V_f = readout\n",
-        "        acc, pred = spyx.fn.integral_accuracy(traces, targets)\n",
-        "        loss = spyx.fn.integral_crossentropy(traces, targets)\n",
+        "        acc, pred = Acc(traces, targets)\n",
+        "        loss = Loss(traces, targets)\n",
         "        return params, [acc, loss, pred, targets]\n",
         "\n",
         "    test_data = dl.test_epoch()\n",