diff --git a/README.md b/README.md
index b6383fb1f81f79fce9450646d23d6717a25dd17c..58a4fe20d9df7d7a6c6b9e9343232e37baac6c20 100644
--- a/README.md
+++ b/README.md
@@ -396,6 +396,7 @@ DeepSparkHub甄选上百个应用算法和模型，覆盖AI和通用计算各领
 [CLIP](multimodal/Language-Image_Pre-Training/clip/pytorch/README.md)  | PyTorch | CIFAR100
 [ControlNet](multimodal/diffusion/ControlNet/README.md)  | PyTorch | Fill50K
 [DDPM](multimodal/diffusion/ddpm/README.md)  | PyTorch | CIFAR-10
+[LLaVA](multimodal/llava/pytorch/README.md)  | PyTorch | LLaVA-Pretrain
 [L-Verse](multimodal/Language-Image_Pre-Training/L-Verse/pytorch/README.md)  | PyTorch | ImageNet
 [Stable Diffusion 1.4](multimodal/diffusion/stable-diffusion/training/README.md)  | PyTorch | pokemon-images
 [Stable Diffusion 1.5](multimodal/diffusion/stable-diffusion/sd_1.5/README.md)  | PyTorch | pokemon-images
diff --git a/multimodal/llava/pytorch/README.md b/multimodal/llava/pytorch/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..9659eb432bac133a22d8a112e671218a47987096
--- /dev/null
+++ b/multimodal/llava/pytorch/README.md
@@ -0,0 +1,55 @@
+# LLaVA 1.5
+
+## Model description
+
+LLaVA is an open-source chatbot trained by fine-tuning LLaMA/Vicuna on GPT-generated multimodal
+instruction-following data. It is an auto-regressive language model, based on the transformer
+architecture.
+
+## Step 1: Preparation
+
+```bash
+git clone --depth 1 https://github.com/haotian-liu/LLaVA
+cd LLaVA/
+git checkout c121f0432da27facab705978f83c4ada465e46fd
+mkdir -p checkpoints/
+mkdir -p data/
+```
+
+### Weights
+
+Download the [llava-hf/llava-1.5-7b-hf](https://huggingface.co/llava-hf/llava-1.5-7b-hf) and put it
+at `checkpoints/`.
+
+Download the [openai/clip-vit-large-patch14](https://huggingface.co/openai/clip-vit-large-patch14)
+and put it at `checkpoints/`.
+
+### Datasets
+
+Download the [liuhaotian/LLaVA-Pretrain](https://huggingface.co/datasets/liuhaotian/LLaVA-Pretrain)
+and put it at `data/`.
+
+## Step 2: Installation
+
+```bash
+
+pip3 install -e .
+pip3 install -e ".[train]"
+pip3 install protobuf
+```
+
+## Step 3: Training
+
+```bash
+bash train.sh
+```
+
+## Results
+
+| Model        | GPUs    | seconds per iteration |
+| ------------ | ------- | --------------------- |
+| LLaVA 1.5 7B | BI-V150 | 8.01s                 |
+
+## Reference
+
+- [LLaVA](https://github.com/haotian-liu/LLaVA)
\ No newline at end of file