Instructions to use RedHatAI/starcoder2-3b-quantized.w8a16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RedHatAI/starcoder2-3b-quantized.w8a16 with Transformers:

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

pipe = pipeline("text-generation", model="RedHatAI/starcoder2-3b-quantized.w8a16")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("RedHatAI/starcoder2-3b-quantized.w8a16")
model = AutoModelForCausalLM.from_pretrained("RedHatAI/starcoder2-3b-quantized.w8a16")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use RedHatAI/starcoder2-3b-quantized.w8a16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "RedHatAI/starcoder2-3b-quantized.w8a16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "RedHatAI/starcoder2-3b-quantized.w8a16",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/RedHatAI/starcoder2-3b-quantized.w8a16

SGLang

How to use RedHatAI/starcoder2-3b-quantized.w8a16 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 "RedHatAI/starcoder2-3b-quantized.w8a16" \
    --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": "RedHatAI/starcoder2-3b-quantized.w8a16",
		"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 "RedHatAI/starcoder2-3b-quantized.w8a16" \
        --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": "RedHatAI/starcoder2-3b-quantized.w8a16",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use RedHatAI/starcoder2-3b-quantized.w8a16 with Docker Model Runner:
```
docker model run hf.co/RedHatAI/starcoder2-3b-quantized.w8a16
```

starcoder2-3b-quantized.w8a16

File size: 1,782 Bytes

{
  "_name_or_path": "/root/.cache/huggingface/hub/models--bigcode--starcoder2-3b/snapshots/733247c55e3f73af49ce8e9c7949bf14af205928",
  "architectures": [
    "Starcoder2ForCausalLM"
  ],
  "attention_dropout": 0.1,
  "bos_token_id": 0,
  "embedding_dropout": 0.1,
  "eos_token_id": 0,
  "hidden_act": "gelu_pytorch_tanh",
  "hidden_size": 3072,
  "initializer_range": 0.018042,
  "intermediate_size": 12288,
  "max_position_embeddings": 16384,
  "mlp_type": "default",
  "model_type": "starcoder2",
  "norm_epsilon": 1e-05,
  "norm_type": "layer_norm",
  "num_attention_heads": 24,
  "num_hidden_layers": 30,
  "num_key_value_heads": 2,
  "residual_dropout": 0.1,
  "rope_theta": 999999.4420358813,
  "sliding_window": 4096,
  "torch_dtype": "float32",
  "transformers_version": "4.43.3",
  "use_bias": true,
  "use_cache": true,
  "vocab_size": 49152,
  "quantization_config": {
    "config_groups": {
      "group_0": {
        "input_activations": null,
        "output_activations": null,
        "targets": [
          "Linear"
        ],
        "weights": {
          "block_structure": null,
          "dynamic": false,
          "group_size": null,
          "num_bits": 8,
          "observer": "minmax",
          "observer_kwargs": {},
          "strategy": "channel",
          "symmetric": true,
          "type": "int"
        }
      }
    },
    "format": "pack-quantized",
    "global_compression_ratio": 1.803492483621931,
    "ignore": [
      "lm_head"
    ],
    "kv_cache_scheme": null,
    "quant_method": "compressed-tensors",
    "quantization_status": "frozen",
    "sparsity_config": {
      "format": "dense",
      "global_sparsity": 1.333353131567116,
      "registry_requires_subclass": false,
      "sparsity_structure": "unstructured"
    }
  }
}