docs: add README.md and getting-started.md

Signed-off-by: Gaius <[email protected]>

Author: gaius-qi

README.md

@@ -1,102 +1,15 @@
 # modctl
 
-Modctl is a user-friendly CLI tool for managing OCI model artifacts, which are bundled based on [Model Spec V1](https://github.com/CloudNativeAI/model-spec).
-It offers commands such as `build`, `pull`, `push`, and more, making it easy for users to convert their AI models into OCI artifacts.
-
-## Usage
-
-### Installation
-
-#### Build from source
-
-```shell
-$ git clone https://github.com/CloudNativeAI/modctl.git
-$ make
-$ ./output/modctl -h
-```
-
-### Build
-
-Build the model artifact you need to prepare a Modelfile describe your expected layout of the model artifact in your model repo.
-
-Example of Modelfile:
-
-```shell
-# Model name (string), such as llama3-8b-instruct, gpt2-xl, qwen2-vl-72b-instruct, etc.
-NAME gemma-2b
-
-# Model architecture (string), such as transformer, cnn, rnn, etc.
-ARCH transformer
-
-# Model family (string), such as llama3, gpt2, qwen2, etc.
-FAMILY gemma
-
-# Model format (string), such as onnx, tensorflow, pytorch, etc.
-FORMAT safetensors
-
-# Number of parameters in the model (integer).
-PARAMSIZE 16
-
-# Model precision (string), such as bf16, fp16, int8, etc.
-PRECISION bf16
-
-# Model quantization (string), such as awq, gptq, etc.
-QUANTIZATION awq
-
-# Specify model configuration file.
-CONFIG config.json
+[![GitHub release](https://img.shields.io/github/release/CloudNativeAI/modctl.svg)](https://github.com/CloudNativeAI/modctl/releases)
+[![LICENSE](https://img.shields.io/github/license/CloudNativeAI/modctl.svg?style=flat-square)](https://github.com/CloudNativeAI/modctl/blob/main/LICENSE)
 
-# Specify model configuration file.
-CONFIG generation_config.json
-
-# Model weight.
-MODEL \.safetensors$
-```
-
-Then run the following command to build the model artifact:
-
-```shell
-$ modctl build -t 127.0.0.1/library/test-models:v1 -f Modelfile .
-```
-
-### Pull & Push
-
-Before the `pull` or `push` command, you need to login the registry:
-
-```shell
-$ modctl login -u username -p password example.registry.com
-```
-
-Pull the model artifact from the registry:
-
-```shell
-$ modctl pull 127.0.0.1/library/test-models:v1
-```
-
-Push the model artifact to the registry:
-
-```shell
-$ modctl push 127.0.0.1/library/test-models:v1
-```
-
-### List
-
-List the model artifacts in the local storage:
-
-```shell
-$ modctl ls
-```
-
-### Cleanup
+Modctl is a user-friendly CLI tool for managing OCI model artifacts, which are bundled based on [Model Spec](https://github.com/CloudNativeAI/model-spec).
+It offers commands such as `build`, `pull`, `push`, and more, making it easy for users to convert their AI models into OCI artifacts.
 
-Delete the model artifact in the local storage:
+## Documentation
 
-```shell
-$ modctl rm 127.0.0.1/library/test-models:v1
-```
+You can find the full documentation on the [getting started](./docs/getting-started.md).
 
-Finally, you can use `purge` command to to remove all unnecessary blobs to free up the storage space:
+## LICENSE
 
-```shell
-$ modctl purge
-```
+Apache 2.0 License. Please see [LICENSE](LICENSE) for more information.

docs/getting-started.md

@@ -0,0 +1,105 @@
+# Getting started with modctl
+
+## Installation
+
+### Binary
+
+```shell
+$ go install github.com/CloudNativeAI/modctl@latest
+```
+
+### Build from source
+
+```shell
+$ git clone https://github.com/CloudNativeAI/modctl.git
+$ make
+$ ./output/modctl -h
+```
+
+## Usage
+
+### Build
+
+Build the model artifact you need to prepare a Modelfile describe your expected layout of the model artifact in your model repo.
+
+Example of Modelfile:
+
+```shell
+# Model name (string), such as llama3-8b-instruct, gpt2-xl, qwen2-vl-72b-instruct, etc.
+NAME gemma-2b
+
+# Model architecture (string), such as transformer, cnn, rnn, etc.
+ARCH transformer
+
+# Model family (string), such as llama3, gpt2, qwen2, etc.
+FAMILY gemma
+
+# Model format (string), such as onnx, tensorflow, pytorch, etc.
+FORMAT safetensors
+
+# Number of parameters in the model (integer).
+PARAMSIZE 16
+
+# Model precision (string), such as bf16, fp16, int8, etc.
+PRECISION bf16
+
+# Model quantization (string), such as awq, gptq, etc.
+QUANTIZATION awq
+
+# Specify model configuration file.
+CONFIG config.json
+
+# Specify model configuration file.
+CONFIG generation_config.json
+
+# Model weight.
+MODEL \.safetensors$
+```
+
+Then run the following command to build the model artifact:
+
+```shell
+$ modctl build -t 127.0.0.1/library/test-models:v1 -f Modelfile .
+```
+
+### Pull & Push
+
+Before the `pull` or `push` command, you need to login the registry:
+
+```shell
+$ modctl login -u username -p password example.registry.com
+```
+
+Pull the model artifact from the registry:
+
+```shell
+$ modctl pull 127.0.0.1/library/test-models:v1
+```
+
+Push the model artifact to the registry:
+
+```shell
+$ modctl push 127.0.0.1/library/test-models:v1
+```
+
+### List
+
+List the model artifacts in the local storage:
+
+```shell
+$ modctl ls
+```
+
+### Cleanup
+
+Delete the model artifact in the local storage:
+
+```shell
+$ modctl rm 127.0.0.1/library/test-models:v1
+```
+
+Finally, you can use `purge` command to to remove all unnecessary blobs to free up the storage space:
+
+```shell
+$ modctl purge
+```