Instructions to use mjschock/mamba-1.4b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mjschock/mamba-1.4b with Transformers:

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

pipe = pipeline("text-generation", model="mjschock/mamba-1.4b", trust_remote_code=True)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mjschock/mamba-1.4b", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("mjschock/mamba-1.4b", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use mjschock/mamba-1.4b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mjschock/mamba-1.4b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mjschock/mamba-1.4b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/mjschock/mamba-1.4b

SGLang

How to use mjschock/mamba-1.4b 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 "mjschock/mamba-1.4b" \
    --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": "mjschock/mamba-1.4b",
		"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 "mjschock/mamba-1.4b" \
        --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": "mjschock/mamba-1.4b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use mjschock/mamba-1.4b with Docker Model Runner:
```
docker model run hf.co/mjschock/mamba-1.4b
```

mamba-1.4b / configuration_mamba.py

mjschock

Upload config

e2a3c23 verified over 2 years ago

Raw

History Blame Contribute Delete

1.11 kB

	import mamba_ssm
	from transformers import PretrainedConfig

	mamba_config_defaults = mamba_ssm.models.config_mamba.MambaConfig()

	class MambaConfig(PretrainedConfig):
	model_type = "mamba"

	def __init__(
	self,
	d_model: int = mamba_config_defaults.d_model,
	fused_add_norm: bool = mamba_config_defaults.fused_add_norm,
	n_layer: int = mamba_config_defaults.n_layer,
	pad_vocab_size_multiple: int = mamba_config_defaults.pad_vocab_size_multiple,
	residual_in_fp32: bool = mamba_config_defaults.residual_in_fp32,
	rms_norm: bool = mamba_config_defaults.rms_norm,
	ssm_cfg: dict = mamba_config_defaults.ssm_cfg,
	vocab_size: int = mamba_config_defaults.vocab_size,
	**kwargs,
	):
	self.d_model = d_model
	self.fused_add_norm = fused_add_norm
	self.n_layer = n_layer
	self.pad_vocab_size_multiple = pad_vocab_size_multiple
	self.residual_in_fp32 = residual_in_fp32
	self.rms_norm = rms_norm
	self.ssm_cfg = ssm_cfg
	self.vocab_size = vocab_size

	super().__init__(**kwargs)