Instructions to use AlumiK/LingLong-317M-Chat with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use AlumiK/LingLong-317M-Chat with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="AlumiK/LingLong-317M-Chat", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("AlumiK/LingLong-317M-Chat", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use AlumiK/LingLong-317M-Chat with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "AlumiK/LingLong-317M-Chat"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AlumiK/LingLong-317M-Chat",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/AlumiK/LingLong-317M-Chat

SGLang

How to use AlumiK/LingLong-317M-Chat with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "AlumiK/LingLong-317M-Chat" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AlumiK/LingLong-317M-Chat",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "AlumiK/LingLong-317M-Chat" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "AlumiK/LingLong-317M-Chat",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use AlumiK/LingLong-317M-Chat with Docker Model Runner:
```
docker model run hf.co/AlumiK/LingLong-317M-Chat
```

AlumiK commited on May 15, 2024

Commit

02ccbed

verified ·

1 Parent(s): 93e9af8

update docs

Browse files

Files changed (1) hide show

README.md +26 -3

README.md CHANGED Viewed

@@ -1,3 +1,26 @@
----
-license: gpl-3.0
----

+---
+language:
+- en
+- zh
+license: gpl-3.0
+pipeline_tag: text-generation
+---
+# LingLong (玲珑): A Small-Scale Chinese PreTrained Language Model
+This is LingLong (玲珑), a Chinese pretrained language model trained by the College of Software at Nankai University.
+Built on the foundation of the GPT-3 architecture, it has been meticulously trained on a vast collection of private
+datasets.
+With a modest parameter count of only 317 million, LingLong is significantly smaller than the original GPT-3 model, yet
+it delivers impressive results across various NLP tasks.
+In Chinese, "玲珑" translates to "exquisite" or "delicate," which perfectly embodies the small yet mighty nature of this
+model.
+Therefore, we chose to name it "LingLong" in honor of its exceptional precision and finesse.
+Although it's true that this model's performance isn't on par with the large pretrained language models boasting
+hundreds of billions of parameters, its relatively low parameter count makes it accessible to researchers with limited
+computing resources.
+As a result, this model serves as an excellent foundation for conducting follow-up research, such as fine-tuning.
+By utilizing this model, researchers can begin to delve into the intricacies of pretrained language models and to
+unravel the mysteries of language processing without the need for excessive computational resources.
+This is the chat version of [LingLong-317M](https://huggingface.co/AlumiK/LingLong-317M)