diff --git a/README.md b/README.md
index 6d59255..e1394b0 100644
--- a/README.md
+++ b/README.md
@@ -19,7 +19,12 @@ Copy the notebook to your Drive and start running cells.
 ```bash
 git clone https://github.com/tracebloc/start-training.git
 cd start-training
-pip install tracebloc_package>=0.6.32
+
+# Pick the extra that matches your ML framework:
+pip install "tracebloc_package[pytorch]>=0.6.33"      # most common
+# pip install "tracebloc_package[tensorflow]>=0.6.33" # TensorFlow
+# pip install "tracebloc_package[all]>=0.6.33"        # everything
+
 jupyter notebook notebooks/traceblocTrainingGuide.ipynb
 ```
 
diff --git a/notebooks/traceblocTrainingGuide.ipynb b/notebooks/traceblocTrainingGuide.ipynb
index ad150c6..c215396 100644
--- a/notebooks/traceblocTrainingGuide.ipynb
+++ b/notebooks/traceblocTrainingGuide.ipynb
@@ -4,9 +4,9 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Guide to Train Machine Learning Models on tracebloc \ud83d\ude80\n",
+    "# Guide to Train Machine Learning Models on tracebloc 🚀\n",
     "\n",
-    "This notebook walks you through training an ML model on the tracebloc platform \u2014 from connecting your account to launching a training run.\n",
+    "This notebook walks you through training an ML model on the tracebloc platform — from connecting your account to launching a training run.\n",
     "\n",
     "**What you'll do:**\n",
     "1. Connect to your tracebloc account\n",
@@ -15,7 +15,7 @@
     "4. Configure a training plan\n",
     "5. Start training\n",
     "\n",
-    "This guide takes about **10\u201315 minutes** to complete."
+    "This guide takes about **10–15 minutes** to complete."
    ]
   },
   {
@@ -26,13 +26,13 @@
     "\n",
     "Before you begin, make sure you have:\n",
     "\n",
-    "- \u2705 A **tracebloc account** \u2014 [Sign up here](https://ai.tracebloc.io/signup) if you don't have one\n",
-    "- \u2705 **Joined a use case** \u2014 you need an active use case with a dataset. [How to join a use case \u2192](https://docs.tracebloc.io/join-use-case/explore-use-case)\n",
-    "- \u2705 A **model file** (`.py`) compatible with the dataset \u2014 you can use one from the [tracebloc model zoo](https://github.com/tracebloc/model-zoo) or bring your own. [Model structure requirements \u2192](https://docs.tracebloc.io/join-use-case/model-optimization)\n",
+    "- ✅ A **tracebloc account** — [Sign up here](https://ai.tracebloc.io/signup) if you don't have one\n",
+    "- ✅ **Joined a use case** — you need an active use case with a dataset. [How to join a use case →](https://docs.tracebloc.io/join-use-case/explore-use-case)\n",
+    "- ✅ A **model file** (`.py`) compatible with the dataset — you can use one from the [tracebloc model zoo](https://github.com/tracebloc/model-zoo) or bring your own. [Model structure requirements →](https://docs.tracebloc.io/join-use-case/model-optimization)\n",
     "\n",
-    "\ud83d\udcd6 **Full documentation:** [docs.tracebloc.io](https://docs.tracebloc.io)\n",
+    "📖 **Full documentation:** [docs.tracebloc.io](https://docs.tracebloc.io)\n",
     "\n",
-    "\ud83d\udca1 **Prefer Google Colab?** [Open this guide in Colab](https://colab.research.google.com/drive/1N00idtpoaq1lk9OJE6g4bMqd8o-Qex2C) \u2014 runs entirely in your browser, no local setup needed."
+    "💡 **Prefer Google Colab?** [Open this guide in Colab](https://colab.research.google.com/drive/1N00idtpoaq1lk9OJE6g4bMqd8o-Qex2C) — runs entirely in your browser, no local setup needed."
    ]
   },
   {
@@ -50,10 +50,7 @@
    "execution_count": null,
    "metadata": {},
    "outputs": [],
-   "source": [
-    "# Install the tracebloc package (pinned version for compatibility)\n",
-    "!pip install tracebloc_package>=0.6.32 -q"
-   ]
+   "source": "# Install the tracebloc package — the [pytorch] extra covers the most common setup.\n# Using TensorFlow? Replace [pytorch] with [tensorflow]. See the README for all extras.\n!pip install \"tracebloc_package[pytorch]>=0.6.33\" -q"
   },
   {
    "cell_type": "code",
@@ -73,7 +70,7 @@
    "source": [
     "**Expected output:** You'll see a prompt asking for your email and password. After entering them, you should see a confirmation that you're logged in.\n",
     "\n",
-    "\u26a0\ufe0f **If login fails:**\n",
+    "⚠️ **If login fails:**\n",
     "- Double-check your email and password at [ai.tracebloc.io](https://ai.tracebloc.io)\n",
     "- Make sure you've verified your email address\n",
     "- If you don't have an account yet, [sign up here](https://ai.tracebloc.io/signup)"
@@ -162,7 +159,7 @@
    "source": [
     "**Expected output:** A confirmation message showing the model was uploaded successfully.\n",
     "\n",
-    "\ud83d\udca1 **Loading weights?** Follow this naming convention:\n",
+    "💡 **Loading weights?** Follow this naming convention:\n",
     "- Model file: `mymodel.py`\n",
     "- Weights file: `mymodel_weights.pkl`\n",
     "\n",
@@ -209,7 +206,7 @@
    "source": [
     "**Expected output:** A confirmation that the model and dataset are linked.\n",
     "\n",
-    "\u26a0\ufe0f **If this fails:**\n",
+    "⚠️ **If this fails:**\n",
     "- Make sure the dataset ID is correct (check your use case page)\n",
     "- Your model must be compatible with the dataset (e.g., an image classification model for an image dataset)"
    ]
@@ -257,11 +254,11 @@
    "metadata": {},
    "source": [
     "**Expected output:** A summary showing all training parameters including:\n",
-    "- **Training Description** \u2014 experiment name, model name, objective\n",
-    "- **Dataset Parameters** \u2014 dataset ID, size, classes\n",
-    "- **Training Parameters** \u2014 epochs, cycles, batch size, validation split\n",
-    "- **Hyperparameters** \u2014 optimizer, loss function, learning rate, callbacks\n",
-    "- **Augmentation Parameters** \u2014 data augmentation settings\n",
+    "- **Training Description** — experiment name, model name, objective\n",
+    "- **Dataset Parameters** — dataset ID, size, classes\n",
+    "- **Training Parameters** — epochs, cycles, batch size, validation split\n",
+    "- **Hyperparameters** — optimizer, loss function, learning rate, callbacks\n",
+    "- **Augmentation Parameters** — data augmentation settings\n",
     "\n",
     "Review these carefully before starting. Adjust any values using the commands in the table above."
    ]
@@ -293,14 +290,14 @@
     "\n",
     "Your model is now being trained on the tracebloc infrastructure. Here's what to expect:\n",
     "\n",
-    "1. **Training starts** \u2014 the model will begin training on the linked dataset inside a secure environment\n",
-    "2. **Monitor progress** \u2014 go to your use case on [ai.tracebloc.io](https://ai.tracebloc.io) to see training status and logs\n",
-    "3. **View results** \u2014 once training completes, check the leaderboard in your use case to see how your model performed\n",
-    "4. **Compare models** \u2014 if other team members or vendors have submitted models, you can compare performance metrics side by side\n",
+    "1. **Training starts** — the model will begin training on the linked dataset inside a secure environment\n",
+    "2. **Monitor progress** — go to your use case on [ai.tracebloc.io](https://ai.tracebloc.io) to see training status and logs\n",
+    "3. **View results** — once training completes, check the leaderboard in your use case to see how your model performed\n",
+    "4. **Compare models** — if other team members or vendors have submitted models, you can compare performance metrics side by side\n",
     "\n",
     "Training time depends on your dataset size, model complexity, and the number of epochs. A typical training run takes a few minutes to a few hours.\n",
     "\n",
-    "\ud83d\udcd6 **Learn more:** [How to evaluate models \u2192](https://docs.tracebloc.io/join-use-case/model-evaluation)"
+    "📖 **Learn more:** [How to evaluate models →](https://docs.tracebloc.io/join-use-case/model-evaluation)"
    ]
   },
   {
@@ -329,10 +326,10 @@
     "---\n",
     "## Need help?\n",
     "\n",
-    "- \ud83d\udcd6 [Documentation](https://docs.tracebloc.io)\n",
-    "- \ud83d\udce7 [support@tracebloc.io](mailto:support@tracebloc.io)\n",
-    "- \ud83d\udc1b [Open an issue](https://github.com/tracebloc/start-training/issues)\n",
-    "- \ud83d\udcac [Discord](https://discord.gg/tracebloc)"
+    "- 📖 [Documentation](https://docs.tracebloc.io)\n",
+    "- 📧 [support@tracebloc.io](mailto:support@tracebloc.io)\n",
+    "- 🐛 [Open an issue](https://github.com/tracebloc/start-training/issues)\n",
+    "- 💬 [Discord](https://discord.gg/tracebloc)"
    ]
   }
  ],
@@ -349,4 +346,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 4
-}
+}
\ No newline at end of file