Instructions to use cosmo3769/starcoderbase-3b-GPTQ with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cosmo3769/starcoderbase-3b-GPTQ with Transformers:

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

pipe = pipeline("text-generation", model="cosmo3769/starcoderbase-3b-GPTQ")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("cosmo3769/starcoderbase-3b-GPTQ")
model = AutoModelForCausalLM.from_pretrained("cosmo3769/starcoderbase-3b-GPTQ")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use cosmo3769/starcoderbase-3b-GPTQ with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cosmo3769/starcoderbase-3b-GPTQ"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cosmo3769/starcoderbase-3b-GPTQ",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/cosmo3769/starcoderbase-3b-GPTQ

SGLang

How to use cosmo3769/starcoderbase-3b-GPTQ 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 "cosmo3769/starcoderbase-3b-GPTQ" \
    --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": "cosmo3769/starcoderbase-3b-GPTQ",
		"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 "cosmo3769/starcoderbase-3b-GPTQ" \
        --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": "cosmo3769/starcoderbase-3b-GPTQ",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use cosmo3769/starcoderbase-3b-GPTQ with Docker Model Runner:
```
docker model run hf.co/cosmo3769/starcoderbase-3b-GPTQ
```

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

starcoderbase-3b-GPTQ

Quantized starcoderbase-3b model to GPTQ format (4-bit precision) using Auto-GPTQ.

Quantization script

Benchmark

Benchmarking script

Baseline starcoderbase-3b model (non-quantized)

Tasks	Version	Filter	n-shot	Metric	Value		Stderr
codexglue_code2text	N/A	none	None	smoothed_bleu_4	1.3519	±	0.3067
- code2text_go	1	none	None	smoothed_bleu_4	1.5781	±	0.3734
- code2text_java	1	none	None	smoothed_bleu_4	1.2778	±	0.1991
- code2text_javascript	1	none	None	smoothed_bleu_4	1.1443	±	0.1181
- code2text_php	1	none	None	smoothed_bleu_4	0.5171	±	0.5171
- code2text_python	1	none	None	smoothed_bleu_4	2.8338	±	1.5323
- code2text_ruby	3	none	None	smoothed_bleu_4	0.7601	±	0.7601

Groups	Version	Filter	n-shot	Metric	Value		Stderr
codexglue_code2text	N/A	none	None	smoothed_bleu_4	1.3519	±	0.3067

Tasks	Version	Filter	n-shot	Metric	Value		Stderr
bigbench_code_line_description_generate_until	1	none	None	exact_match	0	±	0

Tasks	Version	Filter	n-shot	Metric	Value		Stderr
bigbench_code_line_description_multiple_choice	0	none	None	acc	0.25	±	0.0564

Quantized starcoderbase-3b model to GPTQ format

Tasks	Version	Filter	n-shot	Metric	Value		Stderr
codexglue_code2text	N/A	none	None	smoothed_bleu_4	0.9254	±	0.2109
- code2text_go	1	none	None	smoothed_bleu_4	1.4702	±	0.4813
- code2text_java	1	none	None	smoothed_bleu_4	0.6907	±	0.6907
- code2text_javascript	1	none	None	smoothed_bleu_4	0.9469	±	0.0339
- code2text_php	1	none	None	smoothed_bleu_4	0.5171	±	0.5171
- code2text_python	1	none	None	smoothed_bleu_4	1.1676	±	0.2156
- code2text_ruby	3	none	None	smoothed_bleu_4	0.7601	±	0.7601

Groups	Version	Filter	n-shot	Metric	Value		Stderr
codexglue_code2text	N/A	none	None	smoothed_bleu_4	0.9254	±	0.2109

Tasks	Version	Filter	n-shot	Metric	Value		Stderr
bigbench_code_line_description_generate_until	1	none	None	exact_match	0	±	0

Tasks	Version	Filter	n-shot	Metric	Value		Stderr
bigbench_code_line_description_multiple_choice	0	none	None	acc	0.1	±	0.1

Downloads last month: 3