From cf40702e70ca5296516db5ac6e2a31cd3cf5d6f9 Mon Sep 17 00:00:00 2001
From: junhsss <junhsssr@gmail.com>
Date: Tue, 28 Mar 2023 07:08:43 +0900
Subject: [PATCH] docs: update README.md

---
 README.md | 52 +++++++++++++++++++++++++++++++++-------------------
 1 file changed, 33 insertions(+), 19 deletions(-)

diff --git a/README.md b/README.md
index ad15722..34b09e9 100644
--- a/README.md
+++ b/README.md
@@ -10,29 +10,36 @@
 
 **Consistency Models** are a new family of generative models that achieve high sample quality without adversarial training. They support _fast one-step generation_ by design, while still allowing for few-step sampling to trade compute for sample quality.
 
+<br />
+
+## Installation
+
+```sh
+$ pip install consistency
+```
+
 ### Note
 
-If you just want to try things out, just do:
+You **don't need to install** `consistency` for just trying things out:
 
 ```python
 from diffusers import DiffusionPipeline
 
-pipeline = DiffusionPipeline("consistency/cifar10-32-demo", custom_pipeline="consistency/pipeline")
+pipeline = DiffusionPipeline.from_pretrained(
+    "consistency/cifar10-32-demo",
+    custom_pipeline="consistency/pipeline",
+)
 
 pipeline().images[0]  # Super Fast Generation! 🤯
 ```
 
-## Installation
-
-```sh
-$ pip install consistency
-```
+<br />
 
 ## Quickstart
 
 ### Basic
 
-Just wrap your favorite _U-Net_ with `Consistency`. 😊
+Just wrap your favorite _U-Net_ with `Consistency`.
 
 ```python
 import torch
@@ -53,7 +60,7 @@ samples = consistency.sample(16, steps=5, use_ema=True)
 
 `Consistency` is self-contained with the training logic and all necessary schedules.
 
-You can train `Consistency` with **PyTorch Lightning**'s `Trainer` 🚀
+You can train it with **PyTorch Lightning**'s `Trainer` 🚀
 
 ```python
 from pytorch_lightning import Trainer
@@ -64,10 +71,6 @@ trainer.fit(consistency, some_dataloader)
 
 <br />
 
-A complete example can be found in [this **script**](https://github.com/junhsss/consistency-models/blob/main/examples/train.py) or in [this **colab notebook**](https://colab.research.google.com/github/junhsss/consistency-models/blob/main/examples/consistency_models.ipynb).
-
-Checkout [this **Wandb workspace**](https://wandb.ai/junhsss/consistency?workspace=user-junhsss) for some experiment results.
-
 ### Push to HF Hub
 
 Just provide your `model_id` and `token`!
@@ -82,16 +85,27 @@ consistency = Consistency(
 )
 ```
 
-You can safely uninstall `consistency` afterwards. Good luck! 🤞:
+You can safely drop `consistency` afterwards. Good luck! 🤞
 
 ```python
 from diffusers import DiffusionPipeline
 
-pipeline = DiffusionPipeline("your_model_id", custom_pipeline="consistency/pipeline")
+pipeline = DiffusionPipeline.from_pretrained(
+    "your_model_id",
+    custom_pipeline="consistency/pipeline",
+)
 
 pipeline().images[0]
 ```
 
+<br />
+
+A complete example can be found in [this **script**](https://github.com/junhsss/consistency-models/blob/main/examples/train.py) or in [this **colab notebook**](https://colab.research.google.com/github/junhsss/consistency-models/blob/main/examples/consistency_models.ipynb).
+
+Checkout [this **Wandb workspace**](https://wandb.ai/junhsss/consistency?workspace=user-junhsss) for some experiment results.
+
+<br />
+
 ## Available Models
 
 | model_id                    | sample_size |
@@ -100,10 +114,14 @@ pipeline().images[0]
 
 If you've trained some checkpoints. **Share with us! 🤗**
 
+<br />
+
 ## Documentation
 
 In progress... 🛠
 
+<br />
+
 ## Reference
 
 ```bibtex
@@ -118,7 +136,3 @@ In progress... 🛠
   copyright = {arXiv.org perpetual, non-exclusive license}
 }
 ```
-
-## Todo
-
-- [ ] `diffusers` integration. (`ConsistencyPipeline` + `ConsistencyScheduler`)