Instructions to use sail/scaling-vocab-3b-32k-overtrain with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sail/scaling-vocab-3b-32k-overtrain with Transformers:

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

pipe = pipeline("text-generation", model="sail/scaling-vocab-3b-32k-overtrain")

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("sail/scaling-vocab-3b-32k-overtrain")
model = AutoModelForMultimodalLM.from_pretrained("sail/scaling-vocab-3b-32k-overtrain")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use sail/scaling-vocab-3b-32k-overtrain with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sail/scaling-vocab-3b-32k-overtrain"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sail/scaling-vocab-3b-32k-overtrain",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/sail/scaling-vocab-3b-32k-overtrain

SGLang

How to use sail/scaling-vocab-3b-32k-overtrain 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 "sail/scaling-vocab-3b-32k-overtrain" \
    --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": "sail/scaling-vocab-3b-32k-overtrain",
		"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 "sail/scaling-vocab-3b-32k-overtrain" \
        --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": "sail/scaling-vocab-3b-32k-overtrain",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use sail/scaling-vocab-3b-32k-overtrain with Docker Model Runner:
```
docker model run hf.co/sail/scaling-vocab-3b-32k-overtrain
```

The pre-trained 3B model with the vocabulary size 43K in the paper Scaling Laws with Vocabulary: Larger Models Deserve Larger Vocabularies. We investigate how vocabulary size impacts language model scaling law in this paper.

Based on our approach, we predict the optimal vocabulary size for 3B model is about 43K. Then, we train a Llama-based 3B model on a sampled version Slimpajama datasets. The model with 43K vocabulary outperforms the model with the common vocabulary size, 32K, despite using fewer training tokens. It is noteworthy that the proposed approach can be used for different model sizes.

Downloads last month: 5

Safetensors

Model size

3B params

Tensor type

BF16

Space using sail/scaling-vocab-3b-32k-overtrain 1

Collection including sail/scaling-vocab-3b-32k-overtrain

📈 Scaling Laws with Vocabulary

Collection

Increase your vocabulary size when you scale up your language model • 5 items • Updated Aug 11, 2024 • 7

Paper for sail/scaling-vocab-3b-32k-overtrain

Scaling Laws with Vocabulary: Larger Models Deserve Larger Vocabularies

Paper • 2407.13623 • Published Jul 18, 2024 • 56