fix typos

sdk/src/account/mod.rs

@@ -111,7 +111,7 @@ impl Account {
 
     fn __hash__(&self) -> u64 {
         let mut hasher = DefaultHasher::new();
-        // because it's enouth to hash the private key only we add a dummy string so that:
+        // because it's enough to hash the private key only we add a dummy string so that:
         // hash(Account) != hash(PrivateKey)
         "account".hash(&mut hasher);
         self.private_key.hash(&mut hasher);

zkml-research/MNIST_preprocessing/6_presentation.ipynb

@@ -167,7 +167,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# conver to pytorch tensors\n",
+    "# convertto pytorch tensors\n",
     "import torch\n",
     "\n",
     "train_images_tensor_initial = torch.from_numpy(train_images_2d).float()\n",

zkml-research/MNIST_preprocessing/7_trees.ipynb

@@ -167,7 +167,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# conver to pytorch tensors\n",
+    "# convertto pytorch tensors\n",
     "import torch\n",
     "\n",
     "train_images_tensor_initial = torch.from_numpy(train_images_2d).float()\n",

zkml-research/MNIST_preprocessing/8_further_preprocessing.ipynb

@@ -167,7 +167,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# conver to pytorch tensors\n",
+    "# convertto pytorch tensors\n",
     "import torch\n",
     "\n",
     "train_images_tensor_initial = torch.from_numpy(train_images_2d).float()\n",

zkml-research/zkml-web-app/mnist_train.ipynb

@@ -13,7 +13,7 @@
    "source": [
     "MLP neural networks are expressive ML models. Here, we use them for the MNIST classification task, which contains images of handwritten digits. We show the inference of MLP neural networks for the MNIST dataset is possible in a zero knowledge environment.\n",
     "\n",
-    "For this, we first download the dataset, and then compute feature representations of the dataset. We then train and test an MLP neural network on the feature dataset using PyTorch (beacuse we can train with L1 regularization which is helpful for pruning later on). Afterward, we iteratively prune the network (meaning we set weights and biases close to 0 to actually 0, this will save cicruit constraints later on) and fine-tune it. Then, we convert the final PyTorch MLP model to a scikit-learn MLP model (since the transpiler supports scikit-learn models) and transpile the final MLP neural network to Leo, evaluate the Leo network and create a zero knowledge proof."
+    "For this, we first download the dataset, and then compute feature representations of the dataset. We then train and test an MLP neural network on the feature dataset using PyTorch (because we can train with L1 regularization which is helpful for pruning later on). Afterward, we iteratively prune the network (meaning we set weights and biases close to 0 to actually 0, this will save circuitconstraints later on) and fine-tune it. Then, we convert the final PyTorch MLP model to a scikit-learn MLP model (since the transpiler supports scikit-learn models) and transpile the final MLP neural network to Leo, evaluate the Leo network and create a zero knowledge proof."
    ]
   },
   {
@@ -235,7 +235,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# conver to pytorch tensors\n",
+    "# convertto pytorch tensors\n",
     "import torch\n",
     "\n",
     "train_images_tensor_initial = torch.from_numpy(train_images).float()\n",

zkml/examples/MLP_neural_networks/mnist_dataset.ipynb

@@ -13,7 +13,7 @@
    "source": [
     "MLP neural networks are expressive ML models. Here, we use them for the MNIST classification task, which contains images of handwritten digits. We show the inference of MLP neural networks for the MNIST dataset is possible in a zero knowledge environment.\n",
     "\n",
-    "For this, we first download the dataset, and then compute feature representations of the dataset. We then train and test an MLP neural network on the feature dataset using PyTorch (beacuse we can train with L1 regularization which is helpful for pruning later on). Afterward, we iteratively prune the network (meaning we set weights and biases close to 0 to actually 0, this will save cicruit constraints later on) and fine-tune it. Then, we convert the final PyTorch MLP model to a scikit-learn MLP model (since the transpiler supports scikit-learn models) and transpile the final MLP neural network to Leo, evaluate the Leo network and create a zero knowledge proof."
+    "For this, we first download the dataset, and then compute feature representations of the dataset. We then train and test an MLP neural network on the feature dataset using PyTorch (because we can train with L1 regularization which is helpful for pruning later on). Afterward, we iteratively prune the network (meaning we set weights and biases close to 0 to actually 0, this will save circuitconstraints later on) and fine-tune it. Then, we convert the final PyTorch MLP model to a scikit-learn MLP model (since the transpiler supports scikit-learn models) and transpile the final MLP neural network to Leo, evaluate the Leo network and create a zero knowledge proof."
    ]
   },
   {
@@ -235,7 +235,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# conver to pytorch tensors\n",
+    "# convertto pytorch tensors\n",
     "import torch\n",
     "\n",
     "train_images_tensor_initial = torch.from_numpy(train_images).float()\n",

zkml/tests/helper.py

@@ -271,7 +271,7 @@ def read_idx1_ubyte_label_file(filename):
 
 
 
-    # conver to pytorch tensors
+    # convertto pytorch tensors
     import torch
 
     train_images_tensor_initial = torch.from_numpy(train_images_2d).float()