Tutorials: fix headers and remove emojis (#969)

## Description

Fixes header levels and removes emojis from headers in Tutorials

## Ticket


https://wandb.atlassian.net/browse/DOCS-1176?atlOrigin=eyJpIjoiNDhhOTJmZDJjMDViNDg5ZmE2ZjljN2UzMjQ0M2RhOTciLCJwIjoiamlyYS1zbGFjay1pbnQifQ

content/tutorials/_index.md

@@ -28,7 +28,7 @@ See the following tutorials for step by step information on how to use popular M
 
 - [PyTorch](/tutorials/pytorch)
 - [PyTorch Lightning](/tutorials/lightning)
-- [HuggingFace 🤗 Transformers](/tutorials/huggingface)
+- [HuggingFace Transformers](/tutorials/huggingface)
 - Tensorflow
     - [Track experiments](/tutorials/tensorflow)
     - [Tune hyperparameters](/tutorials/tensorflow_sweeps)

content/tutorials/artifacts.md

@@ -13,7 +13,7 @@ Follow along with a [video tutorial](http://tiny.cc/wb-artifacts-video).
 
 ## About artifacts
 
-An artifact, like a Greek [amphora 🏺](https://en.wikipedia.org/wiki/Amphora),
+An artifact, like a Greek [amphora](https://en.wikipedia.org/wiki/Amphora),
 is a produced object -- the output of a process.
 In ML, the most important artifacts are _datasets_ and _models_.
 
@@ -30,7 +30,9 @@ where a training run takes in a dataset and produces a model.
 Since one run can use another run's output as an input, `Artifact`s and `Run`s together form a directed graph (a bipartite [DAG](https://en.wikipedia.org/wiki/Directed_acyclic_graph), with nodes for `Artifact`s and `Run`s
 and arrows that connect a `Run` to the `Artifact`s it consumes or produces.
 
-# 0️⃣ Install and Import
+## Use artifacts to track models and datatsets
+
+### Install and Import
 
 Artifacts are part of our Python library, starting with version `0.9.2`.
 
@@ -49,7 +51,7 @@ import os
 import wandb
 ```
 
-# 1️⃣ Log a Dataset
+### Log a Dataset
 
 First, let's define some Artifacts.
 
@@ -160,7 +162,7 @@ def load_and_log():
 load_and_log()
 ```
 
-### 🚀 `wandb.init`
+#### `wandb.init`
 
 
 When we make the `Run` that's going to produce the `Artifact`s,
@@ -180,7 +182,7 @@ This keeps the graph of your Artifacts nice and tidy.
 
 > **Rule of 👍**: the `job_type` should be descriptive and correspond to a single step of your pipeline. Here, we separate out `load`ing data from `preprocess`ing data.
 
-### 🏺 `wandb.Artifact`
+#### `wandb.Artifact`
 
 
 To log something as an `Artifact`, we have to first make an `Artifact` object.
@@ -202,7 +204,7 @@ The `metadata` just needs to be serializable to JSON.
 
 > **Rule of 👍**: the `metadata` should be as descriptive as possible.
 
-### 🐣 `artifact.new_file` and ✍️ `run.log_artifact`
+#### `artifact.new_file` and `run.log_artifact`
 
 Once we've made an `Artifact` object, we need to add files to it.
 
@@ -227,7 +229,7 @@ including any `Artifact`s that got logged.
 
 We'll see some examples that make better use of the other components of the Run page below.
 
-# 2️⃣ Use a Logged Dataset Artifact
+### Use a Logged Dataset Artifact
 
 `Artifact`s in W&B, unlike artifacts in museums,
 are designed to be _used_, not just stored.
@@ -325,7 +327,7 @@ steps = {"normalize": True,
 preprocess_and_log(steps)
 ```
 
-### ✔️ `run.use_artifact`
+#### `run.use_artifact`
 
 These steps are simpler. The consumer just needs to know the `name` of the `Artifact`, plus a bit more.
 
@@ -342,8 +344,7 @@ so the `Artifact` we want is `mnist-raw:latest`.
 Use custom `alias`es like `latest` or `best` when you want an `Artifact`
 that satisifies some property
 
-### 📥 `artifact.download`
-
+#### `artifact.download`
 
 Now, you may be worrying about the `download` call.
 If we download another copy, won't that double the burden on memory?
@@ -363,10 +364,10 @@ Check out its contents with `!tree artifacts`:
 !tree artifacts
 ```
 
-### 🌐 The Artifacts page on [wandb.ai](https://wandb.ai)
+#### The Artifacts page 
 
 Now that we've logged and used an `Artifact`,
-let's check out the Artifacts tab on the Run page.
+let's check out the Artifacts tab on the Run page. 
 
 Navigate to the Run page URL from the `wandb` output
 and select the "Artifacts" tab from the left sidebar
@@ -384,7 +385,7 @@ with the `type`s of `Artifact`s
 and the `job_type`s of `Run` as the two types of nodes,
 with arrows to represent consumption and production.
 
-# 3️⃣ Log a Model
+### Log a Model
 
 That's enough to see how the API for `Artifact`s works,
 but let's follow this example through to the end of the pipeline
@@ -478,7 +479,7 @@ model_config = {"hidden_layer_sizes": [32, 64],
 build_model_and_log(model_config)
 ```
 
-### ➕ `artifact.add_file`
+#### `artifact.add_file`
 
 
 Instead of simultaneously writing a `new_file` and adding it to the `Artifact`,
@@ -489,7 +490,7 @@ and then `add` them to the `Artifact` in another.
 
 > **Rule of 👍**: use `new_file` when you can, to prevent duplication.
 
-# 4️⃣ Use a Logged Model Artifact
+#### Use a Logged Model Artifact
 
 Just like we could call `use_artifact` on a `dataset`,
 we can call it on our `initialized_model`

content/tutorials/experiments.md

@@ -90,7 +90,7 @@ The following image shows what a dashboard can look like:
 
 Now that we know how to integrate W&B into a psuedo machine learning training loop, let's track a machine learning experiment using a basic PyTorch neural network. The following code will also upload model checkpoints to W&B that you can then share with other teams in your organization.
 
-##  Track a machine learning experiment using Pytorch
+## Track a machine learning experiment using Pytorch
 
 The following code cell defines and trains a simple MNIST classifier. During training, you will see W&B prints out URLs. Click on the project page link to see your results stream in live to a W&B project.
 

content/tutorials/integration-tutorials/huggingface.md

@@ -74,7 +74,7 @@ Optionally, we can set environment variables to customize W&B logging. See [docu
 %env WANDB_WATCH=all
 ```
 
-# 👟 Train the model
+## Train the model
 Next, call the downloaded training script [run_glue.py](https://huggingface.co/transformers/examples.html#glue) and see training automatically get tracked to the Weights & Biases dashboard. This script fine-tunes BERT on the Microsoft Research Paraphrase Corpus— pairs of sentences with human annotations indicating whether they are semantically equivalent.
 
 
@@ -109,7 +109,7 @@ Here's an example comparing [BERT vs DistilBERT](https://app.wandb.ai/jack-morri
 
 {{< img src="/images/tutorials/huggingface-comparearchitectures.gif" alt="" >}}
 
-### Track key information effortlessly by default
+## Track key information effortlessly by default
 Weights & Biases saves a new run for each experiment. Here's the information that gets saved by default:
 - **Hyperparameters**: Settings for your model are saved in Config
 - **Model Metrics**: Time series data of metrics streaming in are saved in Log

content/tutorials/integration-tutorials/keras.md

@@ -14,7 +14,7 @@ Use Weights & Biases for machine learning experiment tracking, dataset versionin
 This Colab notebook introduces the `WandbMetricsLogger` callback. Use this callback for [Experiment Tracking](/guides/track). It will log your training and validation metrics along with system metrics to Weights and Biases.
 
 
-## 🌴 Setup and Installation
+## Setup and Installation
 
 First, let us install the latest version of Weights and Biases. We will then authenticate this colab instance to use W&B.
 
@@ -43,31 +43,31 @@ If this is your first time using W&B or you are not logged in, the link that app
 wandb.login()
 ```
 
-## 🌳 Hyperparameters
+## Hyperparameters
 
 Use of proper config system is a recommended best practice for reproducible machine learning. We can track the hyperparameters for every experiment using W&B. In this colab we will be using simple Python `dict` as our config system.
 
 
 ```python
 configs = dict(
-    num_classes = 10,
-    shuffle_buffer = 1024,
-    batch_size = 64,
-    image_size = 28,
-    image_channels = 1,
-    earlystopping_patience = 3,
-    learning_rate = 1e-3,
-    epochs = 10
+    num_classes=10,
+    shuffle_buffer=1024,
+    batch_size=64,
+    image_size=28,
+    image_channels=1,
+    earlystopping_patience=3,
+    learning_rate=1e-3,
+    epochs=10,
 )
 ```
 
-## 🍁 Dataset
+## Dataset
 
 In this colab, we will be using [CIFAR100](https://www.tensorflow.org/datasets/catalog/cifar100) dataset from TensorFlow Dataset catalog. We aim to build a simple image classification pipeline using TensorFlow/Keras.
 
 
 ```python
-train_ds, valid_ds = tfds.load('fashion_mnist', split=['train', 'test'])
+train_ds, valid_ds = tfds.load("fashion_mnist", split=["train", "test"])
 ```
 
 
@@ -90,14 +90,10 @@ def parse_data(example):
 def get_dataloader(ds, configs, dataloader_type="train"):
     dataloader = ds.map(parse_data, num_parallel_calls=AUTOTUNE)
 
-    if dataloader_type=="train":
+    if dataloader_type == "train":
         dataloader = dataloader.shuffle(configs["shuffle_buffer"])
-
-    dataloader = (
-        dataloader
-        .batch(configs["batch_size"])
-        .prefetch(AUTOTUNE)
-    )
+
+    dataloader = dataloader.batch(configs["batch_size"]).prefetch(AUTOTUNE)
 
     return dataloader
 ```
@@ -108,17 +104,21 @@ trainloader = get_dataloader(train_ds, configs)
 validloader = get_dataloader(valid_ds, configs, dataloader_type="valid")
 ```
 
-# 🎄 Model
+## Model
 
 
 ```python
 def get_model(configs):
-    backbone = tf.keras.applications.mobilenet_v2.MobileNetV2(weights='imagenet', include_top=False)
+    backbone = tf.keras.applications.mobilenet_v2.MobileNetV2(
+        weights="imagenet", include_top=False
+    )
     backbone.trainable = False
 
-    inputs = layers.Input(shape=(configs["image_size"], configs["image_size"], configs["image_channels"]))
+    inputs = layers.Input(
+        shape=(configs["image_size"], configs["image_size"], configs["image_channels"])
+    )
     resize = layers.Resizing(32, 32)(inputs)
-    neck = layers.Conv2D(3, (3,3), padding="same")(resize)
+    neck = layers.Conv2D(3, (3, 3), padding="same")(resize)
     preprocess_input = tf.keras.applications.mobilenet.preprocess_input(neck)
     x = backbone(preprocess_input)
     x = layers.GlobalAveragePooling2D()(x)
@@ -134,33 +134,35 @@ model = get_model(configs)
 model.summary()
 ```
 
-## 🌿 Compile Model
+## Compile Model
 
 
 ```python
 model.compile(
-    optimizer = "adam",
-    loss = "categorical_crossentropy",
-    metrics = ["accuracy", tf.keras.metrics.TopKCategoricalAccuracy(k=5, name='top@5_accuracy')]
+    optimizer="adam",
+    loss="categorical_crossentropy",
+    metrics=[
+        "accuracy",
+        tf.keras.metrics.TopKCategoricalAccuracy(k=5, name="top@5_accuracy"),
+    ],
 )
 ```
 
-## 🌻 Train
+## Train
 
 
 ```python
 # Initialize a W&B run
-run = wandb.init(
-    project = "intro-keras",
-    config = configs
-)
+run = wandb.init(project="intro-keras", config=configs)
 
 # Train your model
 model.fit(
     trainloader,
-    epochs = configs["epochs"],
-    validation_data = validloader,
-    callbacks = [WandbMetricsLogger(log_freq=10)] # Notice the use of WandbMetricsLogger here
+    epochs=configs["epochs"],
+    validation_data=validloader,
+    callbacks=[
+        WandbMetricsLogger(log_freq=10)
+    ],  # Notice the use of WandbMetricsLogger here
 )
 
 # Close the W&B run

content/tutorials/integration-tutorials/keras_models.md

@@ -13,7 +13,7 @@ Use Weights & Biases for machine learning experiment tracking, dataset versionin
 
 This Colab notebook introduces the `WandbModelCheckpoint` callback. Use this callback to log your model checkpoints to Weight and Biases [Artifacts](/guides/artifacts).
 
-## 🌴 Setup and Installation
+## Setup and Installation
 
 First, let us install the latest version of Weights and Biases. We will then authenticate this colab instance to use W&B.
 
@@ -43,7 +43,7 @@ If this is your first time using W&B or you are not logged in, the link that app
 wandb.login()
 ```
 
-## 🌳 Hyperparameters
+## Hyperparameters
 
 Use of proper config system is a recommended best practice for reproducible machine learning. We can track the hyperparameters for every experiment using W&B. In this colab we will be using simple Python `dict` as our config system.
 
@@ -61,7 +61,7 @@ configs = dict(
 )
 ```
 
-## 🍁 Dataset
+## Dataset
 
 In this colab, we will be using [CIFAR100](https://www.tensorflow.org/datasets/catalog/cifar100) dataset from TensorFlow Dataset catalog. We aim to build a simple image classification pipeline using TensorFlow/Keras.
 
@@ -108,7 +108,7 @@ trainloader = get_dataloader(train_ds, configs)
 validloader = get_dataloader(valid_ds, configs, dataloader_type="valid")
 ```
 
-## 🎄 Model
+## Model
 
 
 ```python
@@ -134,7 +134,7 @@ model = get_model(configs)
 model.summary()
 ```
 
-## 🌿 Compile Model
+## Compile Model
 
 
 ```python
@@ -145,7 +145,7 @@ model.compile(
 )
 ```
 
-## 🌻 Train
+## Train
 
 
 ```python