diff --git a/README.md b/README.md
index d5c9d42..5b002e0 100644
--- a/README.md
+++ b/README.md
@@ -138,7 +138,12 @@ import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM
 tokenizer = AutoTokenizer.from_pretrained("internlm/internlm2-chat-7b", trust_remote_code=True)
 # Set `torch_dtype=torch.float16` to load model in float16, otherwise it will be loaded as float32 and might cause OOM Error.
-model = AutoModelForCausalLM.from_pretrained("internlm/internlm2-chat-7b", trust_remote_code=True, torch_dtype=torch.float16).cuda()
+model = AutoModelForCausalLM.from_pretrained("internlm/internlm2-chat-7b", device_map="auto", trust_remote_code=True, torch_dtype=torch.float16)
+# (Optional) If on low resource devices, you can load model in 4-bit or 8-bit to further save GPU memory via bitsandbytes.
+  # InternLM 7B in 4bit will cost nearly 8GB GPU memory. 
+  # pip install -U bitsandbytes
+  # 8-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_8bit=True)
+  # 4-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_4bit=True)
 model = model.eval()
 response, history = model.chat(tokenizer, "hello", history=[])
 print(response)
@@ -158,6 +163,11 @@ model_dir = snapshot_download('Shanghai_AI_Laboratory/internlm2-chat-7b')
 tokenizer = AutoTokenizer.from_pretrained(model_dir, device_map="auto", trust_remote_code=True)
 # Set `torch_dtype=torch.float16` to load model in float16, otherwise it will be loaded as float32 and might cause OOM Error.
 model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, torch_dtype=torch.float16)
+# (Optional) If on low resource devices, you can load model in 4-bit or 8-bit to further save GPU memory via bitsandbytes.
+  # InternLM 7B in 4bit will cost nearly 8GB GPU memory. 
+  # pip install -U bitsandbytes
+  # 8-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_8bit=True)
+  # 4-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_4bit=True)
 model = model.eval()
 response, history = model.chat(tokenizer, "hello", history=[])
 print(response)
diff --git a/README_zh-CN.md b/README_zh-CN.md
index f682cd2..2892a1a 100644
--- a/README_zh-CN.md
+++ b/README_zh-CN.md
@@ -136,7 +136,12 @@ import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM
 tokenizer = AutoTokenizer.from_pretrained("internlm/internlm2-chat-7b", trust_remote_code=True)
 # 设置`torch_dtype=torch.float16`来将模型精度指定为torch.float16，否则可能会因为您的硬件原因造成显存不足的问题。
-model = AutoModelForCausalLM.from_pretrained("internlm/internlm2-chat-7b", trust_remote_code=True, torch_dtype=torch.float16).cuda()
+model = AutoModelForCausalLM.from_pretrained("internlm/internlm2-chat-7b", device_map="auto",trust_remote_code=True, torch_dtype=torch.float16)
+# (可选) 如果在低资源设备上，可以通过bitsandbytes加载4-bit或8-bit量化的模型，进一步节省GPU显存.
+  # 4-bit 量化的 InternLM 7B 大约会消耗 8GB 显存. 
+  # pip install -U bitsandbytes
+  # 8-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_8bit=True)
+  # 4-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_4bit=True)
 model = model.eval()
 response, history = model.chat(tokenizer, "你好", history=[])
 print(response)
@@ -155,6 +160,11 @@ from modelscope import snapshot_download, AutoTokenizer, AutoModelForCausalLM
 model_dir = snapshot_download('Shanghai_AI_Laboratory/internlm2-chat-7b')
 tokenizer = AutoTokenizer.from_pretrained(model_dir, device_map="auto", trust_remote_code=True)
 model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, torch_dtype=torch.float16)
+# (可选) 如果在低资源设备上，可以通过bitsandbytes加载4-bit或8-bit量化的模型，进一步节省GPU显存.
+  # 4-bit 量化的 InternLM 7B 大约会消耗 8GB 显存. 
+  # pip install -U bitsandbytes
+  # 8-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_8bit=True)
+  # 4-bit: model = AutoModelForCausalLM.from_pretrained(model_dir, device_map="auto", trust_remote_code=True, load_in_4bit=True)
 model = model.eval()
 response, history = model.chat(tokenizer, "hello", history=[])
 print(response)