Text Generation
Transformers
PyTorch
Safetensors
English
maba
maba-v1
maba-v1.1
recurrent
gated-deltanet
gdn
gdn-2
linear-attention
linear-recurrence
state-space-model
ssm
gqa
grouped-query-attention
swiglu
rmsnorm
rope
speculative-decoding
mtp
multi-token-prediction
tinystories
100m
nlp
casual-lm
transformer
qwen
minicpm
benchmark
Instructions to use AndrewThompson1233/maba-101m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AndrewThompson1233/maba-101m with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="AndrewThompson1233/maba-101m")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("AndrewThompson1233/maba-101m", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use AndrewThompson1233/maba-101m with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "AndrewThompson1233/maba-101m" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "AndrewThompson1233/maba-101m", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/AndrewThompson1233/maba-101m
- SGLang
How to use AndrewThompson1233/maba-101m 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 "AndrewThompson1233/maba-101m" \ --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": "AndrewThompson1233/maba-101m", "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 "AndrewThompson1233/maba-101m" \ --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": "AndrewThompson1233/maba-101m", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use AndrewThompson1233/maba-101m with Docker Model Runner:
docker model run hf.co/AndrewThompson1233/maba-101m
| { | |
| "architectures": [ | |
| "MabaModel" | |
| ], | |
| "model_type": "maba", | |
| "vocab_size": 32768, | |
| "d_emb": 128, | |
| "dim": 640, | |
| "hidden_size": 640, | |
| "n_layers": 20, | |
| "num_hidden_layers": 20, | |
| "n_passes": 2, | |
| "layer_types": [ | |
| 0, | |
| 0, | |
| 0, | |
| 1, | |
| 0, | |
| 0, | |
| 0, | |
| 1, | |
| 0, | |
| 0, | |
| 0, | |
| 1, | |
| 0, | |
| 0, | |
| 0, | |
| 1, | |
| 0, | |
| 0, | |
| 0, | |
| 1 | |
| ], | |
| "n_heads": 10, | |
| "num_attention_heads": 10, | |
| "d_head": 64, | |
| "n_kv_heads": 2, | |
| "num_key_value_heads": 2, | |
| "kernel_size": 4, | |
| "d_ffn": 1728, | |
| "intermediate_size": 1728, | |
| "gate_bias": 2.0, | |
| "rope_theta": 500000.0, | |
| "max_len": 4096, | |
| "max_position_embeddings": 4096, | |
| "mtp_k": 2, | |
| "mtp_weight": 0.3, | |
| "eps": 1e-06, | |
| "rms_norm_eps": 1e-06, | |
| "init_std": 0.02, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.49.0", | |
| "model_version": "1.1.0" | |
| } |