Update README.md

README.md

@@ -1,15 +1,17 @@
----
-license: apache-2.0
----
-
 <p align="center">
- <img src="./lumina-logo.png" width="30%"/> 
+ <img src="../assets/lumina-logo.png" width="30%"/> 
  <br>
 </p>
 
 # Lumina-T2I
 
-Lumina-T2I is a model that generates images base on text condition, supporting various text encoders and models of different parameter sizes. With minimal training costs, it achieves high-quality image generation by training from scratch. Additionally, it offers usage through CLI console programs and Web Demo displays.
+Lumina-T2I is a model that generates images based on text conditions, supporting various text encoders and models of different parameter sizes. With minimal training costs, it achieves high-quality image generation by training from scratch. Additionally, it offers usage through CLI console programs and Web Demo displays.
+
+Our generative model has `LargeDiT` as the backbone, the text encoder is the `LLaMa` 7B model, and the VAE uses a version of `sdxl` fine-tuned by stabilityai.
+
+- Generation Model: Large-DiT
+- Text Encoder: LLaMa-7B
+- VAE: stabilityai/sd-vae-ft-sdxl
 
 ## 📰 News
 
@@ -21,13 +23,7 @@ More checkpoints of our model will be released soon~
 
 | Resolution | Flag-DiT Parameter| Text Encoder | Prediction | Download URL  |
 | ---------- | ----------------------- | ------------ | -----------|-------------- |
-| 1024       | 5B             |    LLaMa-7B  |   Rectified Flow | [hugging face](https://huggingface.co/Alpha-VLLM/Lumina-T2X/tree/main/Lumina-T2I/5B/1024px) |
-
-Using git for cloning the model you want to use:
-
-```bash
-git clone https://huggingface.co/Alpha-VLLM/Lumina-T2I
-```
+| 1024       | 5B             |    LLaMa-7B  |   Rectified Flow | [hugging face](https://huggingface.co/Alpha-VLLM/Lumina-T2I) |
 
 ## Installation
 
@@ -47,6 +43,12 @@ On some outdated distros (e.g., CentOS 7), you may also want to check that a lat
 gcc --version
 ```
 
+Downloading Lumina-T2X repo from github:
+
+```bash
+git clone https://github.com/Alpha-VLLM/Lumina-T2X
+```
+
 ### 1. Create a conda environment and install PyTorch
 
 Note: You may want to adjust the CUDA version [according to your driver version](https://docs.nvidia.com/deploy/cuda-compatibility/#default-to-minor-version).
@@ -66,7 +68,7 @@ Note: You may want to adjust the CUDA version [according to your driver version]
   or you can use
 
   ```bash
-  cd Lumina-T2I
+  cd lumina-t2i
   pip install -r requirements.txt
   ```
 
@@ -105,7 +107,7 @@ pip install -v --disable-pip-version-check --no-cache-dir --no-build-isolation -
 
 ## Inference
 
-To ensure that our generative model is ready to use right out of the box, we provide a user-friendly CLI program and a locally deployable Web Demo site.
+To ensure that our generative model is ready to use immediately, we provide a user-friendly CLI program and a locally deployable Web Demo site.
 
 ### CLI
 
@@ -115,17 +117,33 @@ To ensure that our generative model is ready to use right out of the box, we pro
 pip install -e .
 ```
 
-2. Setting your personal inference configuration
+2. Prepare the pre-trained model
+
+⭐⭐ (Recommended) you can use huggingface_cli to download our model:
+
+```bash
+huggingface-cli download --resume-download Alpha-VLLM/Lumina-T2I --local-dir /path/to/ckpt
+```
+
+or using git for cloning the model you want to use:
+
+```bash
+git clone https://huggingface.co/Alpha-VLLM/Lumina-T2I
+``` 
+
+1. Setting your personal inference configuration
 
 Update your own personal inference settings to generate different styles of images, checking `config/infer/config.yaml` for detailed settings. Detailed config structure:
 
+> `/path/to/ckpt` should be a directory containing `consolidated*.pth` and `model_args.pth`
+
 ```yaml
 - settings:
 
   model:
-    ckpt: ""
-    ckpt_lm: ""
-    token: ""
+    ckpt: "/path/to/ckpt"           # if ckpt is "", you should use `--ckpt` for passing model path when using `lumina` cli.
+    ckpt_lm: ""                     # if ckpt is "", you should use `--ckpt_lm` for passing model path when using `lumina` cli.
+    token: ""                       # if LLM is a huggingface gated repo, you should input your access token from huggingface and when token is "", you should `--token` for accessing the model.
 
   transport:
     path_type: "Linear"             # option: ["Linear", "GVP", "VP"]