docs: update demos (#322)

sozercan · Jun 24, 2024 · 70ce939 · 70ce939
1 parent c68ab71
commit 70ce939
Show file tree

Hide file tree

Showing 9 changed files with 57 additions and 9 deletions.
diff --git a/README.md b/README.md
@@ -4,7 +4,7 @@
 <img src="./website/static/img/logo.png" width="200"><br>
 </p>
 
-AIKit is a one-stop shop to quickly get started to host, deploy, build and fine-tune large language models (LLMs).
+AIKit is a comprehensive platform to quickly get started to host, deploy, build and fine-tune large language models (LLMs).
 
 AIKit offers two main capabilities:
 

diff --git a/website/docs/create-images.md b/website/docs/create-images.md
@@ -16,7 +16,9 @@ docker buildx create --use --name aikit-builder
 
 You can easily build an image using the following ways:
 
-## Hugging Face
+### Hugging Face
+
+🎬 Demo: https://www.youtube.com/watch?v=DI5NbdEFLC8
 
 You can use [Hugging Face](https://huggingface.co) models directly by providing the model URL. For example:
 
@@ -48,6 +50,8 @@ Resulting model name will be the file name. In this case, `llama-2-7b-chat.Q4_K_
 
 ### OCI Artifacts
 
+🎬 Demo: https://www.youtube.com/watch?v=G6PrzhEe_p8
+
 You can use OCI artifacts to download models. For example:
 
 ```bash
@@ -95,6 +99,8 @@ Please note that ARM64 support only applies to the `llama.cpp` backend with CPU
 
 ## Advanced Usage
 
+🎬 Demo: https://www.youtube.com/watch?v=5AQfG5VwN2c&list=PLx4Tje2rS923Bkw83GkobOyjIFLfxNrvs&index=2
+
 Create an `aikitfile.yaml` with the following structure:
 
 ```yaml

diff --git a/website/docs/demo.md b/website/docs/demo.md
@@ -2,14 +2,38 @@
 title: Demos
 ---
 
-## Fine Tuning Mistral 7b with OpenHermes dataset and building a model image for inference
+Below are various demos to help you get started with AIKit.
 
-https://www.youtube.com/watch?v=FZuVb-9i-94
+## Inference
 
-## Building an image with a Llama 2 model
+### Running inference with CUDA
 
-[![Building an image with a Llama 2 model](https://asciinema.org/a/J9bitkONKPvedSfU1RkrmVEhD.svg 'Building an image with a Llama 2 model')](https://asciinema.org/a/J9bitkONKPvedSfU1RkrmVEhD)
+https://www.youtube.com/watch?v=yFh_Zfk34PE
 
-## Inference
+### Running inference (on ARM64)
+
+https://www.youtube.com/watch?v=O0AOnxXp-o4
+
+### Using Stable Diffusion to generate images
+
+https://www.youtube.com/watch?v=gh7b-rt70Ug
+
+## Building a model image
+
+### Building a model image from model in Hugging Face and running inference
 
-[![Inference](https://asciinema.org/a/DYh5bCQMNPSis1whhsfPeMOoM.svg 'Inference')](https://asciinema.org/a/DYh5bCQMNPSis1whhsfPeMOoM)
+https://www.youtube.com/watch?v=DI5NbdEFLC8
+
+### Building a model image from model in ollama (OCI artifacts) and running inference
+
+https://www.youtube.com/watch?v=G6PrzhEe_p8
+
+### Building a model image from configuration and running inference
+
+https://www.youtube.com/watch?v=5AQfG5VwN2c
+
+## Fine tuning
+
+### Fine Tuning Mistral 7b with OpenHermes dataset and building a model image for inference
+
+https://www.youtube.com/watch?v=FZuVb-9i-94
diff --git a/website/docs/fine-tune.md b/website/docs/fine-tune.md
@@ -81,6 +81,10 @@ $ ls -al _output
 -rw-r--r--  1 sozercan sozercan 7161089856 Mar  3 00:19 aikit-model-q4_k_m.gguf
 ```
 
+## Demo
+
+https://www.youtube.com/watch?v=FZuVb-9i-94
+
 ## What's next?
 
 👉 Now that you have a fine-tuned model output as a GGUF file, you can refer to [Creating Model Images](./create-images.md) on how to create an image with AIKit to serve your fine-tuned model!

diff --git a/website/docs/gpu.md b/website/docs/gpu.md
@@ -56,3 +56,7 @@ If GPU acceleration is working, you'll see output that is similar to following i
 5:32AM DBG GRPC(llama-2-7b-chat.Q4_K_M.gguf-127.0.0.1:43735): stderr llm_load_tensors: offloaded 35/35 layers to GPU
 5:32AM DBG GRPC(llama-2-7b-chat.Q4_K_M.gguf-127.0.0.1:43735): stderr llm_load_tensors: VRAM used: 5869 MB
 ```
+
+## Demo
+
+https://www.youtube.com/watch?v=yFh_Zfk34PE
diff --git a/website/docs/intro.md b/website/docs/intro.md
@@ -3,7 +3,7 @@ title: Introduction
 slug: /
 ---
 
-AIKit is a one-stop shop to quickly get started to host, deploy, build and fine-tune large language models (LLMs).
+AIKit is a comprehensive platform to quickly get started to host, deploy, build and fine-tune large language models (LLMs).
 
 AIKit offers two main capabilities:
 

diff --git a/website/docs/kubernetes.md b/website/docs/kubernetes.md
@@ -12,6 +12,8 @@ You can use [kind](https://kind.sigs.k8s.io/) to create a local Kubernetes clust
 
 ## Helm Chart
 
+🎬 Demo: https://www.youtube.com/watch?v=ws5AUtLkuuc
+
 For advanced deployments or customization options, you can use the [Helm chart](https://helm.sh/) provided in the `charts` directory.
 
 Please make sure you have Helm installed and configured. If you don't have Helm installed, you can follow the instructions [here](https://helm.sh/docs/intro/install/).

diff --git a/website/docs/quick-start.md b/website/docs/quick-start.md
@@ -45,6 +45,10 @@ Output should be similar to:
 
 That's it! 🎉 API is OpenAI compatible so this is a drop-in replacement for any OpenAI API compatible client.
 
+## Demo
+
+https://www.youtube.com/watch?v=O0AOnxXp-o4
+
 ## What's next?
 
 👉 If you are interested in other pre-made models (such as Gemma, Mixtral or Phi), please refer to [Pre-made models](./premade-models.md).

diff --git a/website/docs/stablediffusion.md b/website/docs/stablediffusion.md
@@ -19,3 +19,7 @@ Please make sure to change syntax to `#syntax=ghcr.io/sozercan/aikit:latest` in
 :::
 
 https://github.com/sozercan/aikit/blob/main/test/aikitfile-stablediffusion.yaml
+
+## Demo
+
+https://www.youtube.com/watch?v=gh7b-rt70Ug
-Original file line number
+Diff line change
@@ Expand Up / @@ -3,7 +3,7 @@ title: Introduction @@
     slug: /
     ---
-    AIKit is a one-stop shop to quickly get started to host, deploy, build and fine-tune large language models (LLMs).
+    AIKit is a comprehensive platform to quickly get started to host, deploy, build and fine-tune large language models (LLMs).
     AIKit offers two main capabilities:
@@ Expand Down @@