Instructions to use pvlabs/PingVortexLM1-20M-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use pvlabs/PingVortexLM1-20M-Base with Transformers:

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

pipe = pipeline("text-generation", model="pvlabs/PingVortexLM1-20M-Base")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("pvlabs/PingVortexLM1-20M-Base")
model = AutoModelForCausalLM.from_pretrained("pvlabs/PingVortexLM1-20M-Base")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use pvlabs/PingVortexLM1-20M-Base with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "pvlabs/PingVortexLM1-20M-Base"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "pvlabs/PingVortexLM1-20M-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/pvlabs/PingVortexLM1-20M-Base

SGLang

How to use pvlabs/PingVortexLM1-20M-Base 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 "pvlabs/PingVortexLM1-20M-Base" \
    --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": "pvlabs/PingVortexLM1-20M-Base",
		"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 "pvlabs/PingVortexLM1-20M-Base" \
        --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": "pvlabs/PingVortexLM1-20M-Base",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use pvlabs/PingVortexLM1-20M-Base with Docker Model Runner:
```
docker model run hf.co/pvlabs/PingVortexLM1-20M-Base
```

PingVortexLM1-20M-Base

A small experimental language model based on LLaMA architecture trained on custom high-quality English dataset with around 200M tokens. This model is just an experiment, it is not designed for coherent text generation or logical reasoning and may produce repetitive or nonsensical outputs.

Built by PingVortex Labs.

Model Details

Parameters: 20M
Context length: 8192 tokens
Language: English only
License: Apache 2.0

Usage

from transformers import LlamaForCausalLM, PreTrainedTokenizerFast

model = LlamaForCausalLM.from_pretrained("pvlabs/PingVortexLM1-20M-Base")
tokenizer = PreTrainedTokenizerFast.from_pretrained("pvlabs/PingVortexLM1-20M-Base")

# don't expect a coherent response
prompt = "The capital of France is"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=50, repetition_penalty=1.3)
print(tokenizer.decode(outputs[0]))

Made by PingVortex.

Downloads last month: 52

Safetensors

Model size

19.2M params

Tensor type

BF16

Collection including pvlabs/PingVortexLM1-20M-Base

PingVortexLM

Collection

Series of our PingVortexLM models • 1 item • Updated Apr 11 • 1