Instructions to use Jiayi-Pan/Tiny-Vicuna-1B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Jiayi-Pan/Tiny-Vicuna-1B with Transformers:

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

pipe = pipeline("text-generation", model="Jiayi-Pan/Tiny-Vicuna-1B")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Jiayi-Pan/Tiny-Vicuna-1B")
model = AutoModelForCausalLM.from_pretrained("Jiayi-Pan/Tiny-Vicuna-1B", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Jiayi-Pan/Tiny-Vicuna-1B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Jiayi-Pan/Tiny-Vicuna-1B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jiayi-Pan/Tiny-Vicuna-1B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Jiayi-Pan/Tiny-Vicuna-1B

SGLang

How to use Jiayi-Pan/Tiny-Vicuna-1B 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 "Jiayi-Pan/Tiny-Vicuna-1B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jiayi-Pan/Tiny-Vicuna-1B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Jiayi-Pan/Tiny-Vicuna-1B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Jiayi-Pan/Tiny-Vicuna-1B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Jiayi-Pan/Tiny-Vicuna-1B with Docker Model Runner:
```
docker model run hf.co/Jiayi-Pan/Tiny-Vicuna-1B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Tiny Vicuna 1B

This model is a fine-tuned version of TinyLlama on WizardVicuna Dataset. It should be fully compatible with Vicuna-v1.5 series.

This model is easy to iterate on for early experiments!

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	34.76
AI2 Reasoning Challenge (25-Shot)	33.45
HellaSwag (10-Shot)	55.92
MMLU (5-Shot)	25.45
TruthfulQA (0-shot)	33.82
Winogrande (5-shot)	58.41
GSM8k (5-shot)	1.52

Downloads last month: 1,922

Safetensors

Model size

1B params

Tensor type

F32

Model tree for Jiayi-Pan/Tiny-Vicuna-1B

Finetunes

1 model

Merges

1 model

Quantizations

3 models

Spaces using Jiayi-Pan/Tiny-Vicuna-1B 15

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

33.450
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

55.920
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

25.450
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

33.820
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

58.410
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

1.520