From 31ca4897bbff517156f7730b9562ac30061b39d5 Mon Sep 17 00:00:00 2001
From: Laurent Mazare <laurent.mazare@gmail.com>
Date: Fri, 20 Oct 2023 09:08:39 +0100
Subject: [PATCH] Readme updates. (#1134)

---
 README.md                              | 15 +++++++--------
 candle-examples/examples/vit/README.md | 20 ++++++++++++++++++++
 2 files changed, 27 insertions(+), 8 deletions(-)
 create mode 100644 candle-examples/examples/vit/README.md

diff --git a/README.md b/README.md
index fd3a9fbf..03c2a1f5 100644
--- a/README.md
+++ b/README.md
@@ -130,8 +130,11 @@ And then head over to
 
 <!--- ANCHOR: useful_libraries --->
 
-## Useful Libraries
-- [`candle-lora`](https://github.com/EricLBuehler/candle-lora) provides a LoRA implementation that conforms to the official `peft` implementation.
+## Useful External Resources
+- [`candle-tutorial`](https://github.com/ToluClassics/candle-tutorial): a
+  very detailed tutorial showing how to convert a PyTorch model to Candle.
+- [`candle-lora`](https://github.com/EricLBuehler/candle-lora): a LoRA implementation
+  that conforms to the official `peft` implementation.
 
 If you have an addition to this list, please submit a pull request.
 
@@ -163,12 +166,8 @@ If you have an addition to this list, please submit a pull request.
     - Stable Diffusion v1.5, v2.1, XL v1.0.
     - Wurstchen v2.
     - Computer Vision Models.
-        - DINOv2.
-        - ConvMixer.
-        - EfficientNet.
-        - ResNet-18/34/50/101/152.
-        - yolo-v3.
-        - yolo-v8.
+        - DINOv2, ConvMixer, EfficientNet, ResNet, ViT.
+        - yolo-v3, yolo-v8.
         - Segment-Anything Model (SAM).
 - File formats: load models from safetensors, npz, ggml, or PyTorch files.
 - Serverless (on CPU), small and fast deployments.
diff --git a/candle-examples/examples/vit/README.md b/candle-examples/examples/vit/README.md
new file mode 100644
index 00000000..42e9a6a7
--- /dev/null
+++ b/candle-examples/examples/vit/README.md
@@ -0,0 +1,20 @@
+# candle-vit
+
+Vision Transformer (ViT) model implementation following the lines of
+[vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224)
+This uses a classification head trained on the ImageNet dataset and returns the
+probabilities for the top-5 classes.
+
+## Running an example
+
+```
+$ cargo run --example vit --release -- --image tiger.jpg
+
+loaded image Tensor[dims 3, 224, 224; f32]
+model built
+tiger, Panthera tigris  : 100.00%
+tiger cat               : 0.00%
+jaguar, panther, Panthera onca, Felis onca: 0.00%
+leopard, Panthera pardus: 0.00%
+lion, king of beasts, Panthera leo: 0.00%
+```