Update documentation from main repository

Author: future-xy

docs/api/sllm-store-cli.md

@@ -33,7 +33,7 @@ pip install serverless-llm-store
 ```
 
 ## Example Workflow
-1. Firstly, start the ServerlessLLM Store server. By default, it uses ./models as the storage path.
+1. Firstly, start the ServerlessLLM Store server. By default, it uses ~/models as the storage path.
 Launch the checkpoint store server in a separate process:
 ``` bash
 # 'mem_pool_size' is the maximum size of the memory pool in GB. It should be larger than the model size.

docs/deployment/single_machine.md

@@ -132,7 +132,7 @@ ray start --address=0.0.0.0:6379 --num-cpus=4 --num-gpus=1 \
 
 ### 2. Start the ServerlessLLM Store Server
 
-Next, start the ServerlessLLM Store server. By default, it uses `./models` as the storage path.
+Next, start the ServerlessLLM Store server. By default, it uses `~/models` as the storage path.
 
 Open a new terminal and run:
 

docs/store/quickstart.md

@@ -50,7 +50,7 @@ We highly recommend using a fast storage device (e.g., NVMe SSD) to store the mo
 For example, create a directory `models` on the NVMe SSD and link it to the local path.
 ```bash
 mkdir -p /mnt/nvme/models   # Replace '/mnt/nvme' with your NVMe SSD path.
-ln -s /mnt/nvme/models ./models
+ln -s /mnt/nvme/models ~/models
 ```
 :::
 
@@ -63,8 +63,8 @@ import torch
 from transformers import AutoModelForCausalLM
 model = AutoModelForCausalLM.from_pretrained('facebook/opt-1.3b', torch_dtype=torch.float16)
 
-# Replace './models' with your local path.
-save_model(model, './models/facebook/opt-1.3b')
+# Replace '~/models' with your local path.
+save_model(model, '~/models/facebook/opt-1.3b')
 ```
 
 2. Launch the checkpoint store server in a separate process:
@@ -94,7 +94,7 @@ for i in range(num_gpus):
     torch.cuda.synchronize()
 
 start = time.time()
-model = load_model("facebook/opt-1.3b", device_map="auto", torch_dtype=torch.float16, storage_path="./models/", fully_parallel=True)
+model = load_model("facebook/opt-1.3b", device_map="auto", torch_dtype=torch.float16, storage_path=os.path.expanduser("~/models"), fully_parallel=True)
 # Please note the loading time depends on the model size and the hardware bandwidth.
 print(f"Model loading time: {time.time() - start:.2f}s")
 
@@ -164,7 +164,7 @@ from vllm import LLM, SamplingParams
 
 import os
 
-storage_path = os.getenv("STORAGE_PATH", "./models")
+storage_path = os.getenv("STORAGE_PATH", os.path.expanduser("~/models"))
 model_name = "facebook/opt-1.3b"
 model_path = os.path.join(storage_path, model_name)
 
@@ -211,8 +211,8 @@ from sllm_store.transformers import save_lora
 from transformers import AutoModelForCausalLM
 model = AutoModelForCausalLM.from_pretrained('facebook/opt-1.3b', torch_dtype=torch.float16) -->
 
-# Replace './models' with your local path.
-save_lora(adapter, './models/facebook/opt-1.3b')
+# Replace '~/models' with your local path.
+save_lora(adapter, '~/models/facebook/opt-1.3b')
 ```
 
 2. Launch the checkpoint store server in a separate process:
@@ -227,9 +227,9 @@ import time
 import torch
 from sllm_store.transformers import load_model, load_lora
 
-model = load_model("facebook/opt-1.3b", device_map="auto", torch_dtype=torch.float16, storage_path="./models/", fully_parallel=True)
+model = load_model("facebook/opt-1.3b", device_map="auto", torch_dtype=torch.float16, storage_path=os.path.expanduser("~/models"), fully_parallel=True)
 
-model = load_lora("facebook/opt-1.3b", adapter_name="demo_lora", adapter_path="ft_facebook/opt-1.3b_adapter1", device_map="auto", torch_dtype=torch.float16, storage_path="./models/")
+model = load_lora("facebook/opt-1.3b", adapter_name="demo_lora", adapter_path="ft_facebook/opt-1.3b_adapter1", device_map="auto", torch_dtype=torch.float16, storage_path=os.path.expanduser("~/models"))
 
 # Please note the loading time depends on the base model size and the hardware bandwidth.
 print(f"Model loading time: {time.time() - start:.2f}s")