Instructions to use reaperdoesntknow/MoA-150M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use reaperdoesntknow/MoA-150M with Transformers:

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

pipe = pipeline("text-generation", model="reaperdoesntknow/MoA-150M")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import MoAMetricLM
model = MoAMetricLM.from_pretrained("reaperdoesntknow/MoA-150M", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use reaperdoesntknow/MoA-150M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "reaperdoesntknow/MoA-150M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "reaperdoesntknow/MoA-150M",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/reaperdoesntknow/MoA-150M

SGLang

How to use reaperdoesntknow/MoA-150M 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 "reaperdoesntknow/MoA-150M" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "reaperdoesntknow/MoA-150M",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "reaperdoesntknow/MoA-150M" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "reaperdoesntknow/MoA-150M",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use reaperdoesntknow/MoA-150M with Docker Model Runner:
```
docker model run hf.co/reaperdoesntknow/MoA-150M
```

MoA-150M / config.json

reaperdoesntknow

Upload MoAMetricLM

63361ab verified 8 months ago

raw

history blame contribute delete

1.28 kB

	{
	"alpha_init": 1.5,
	"architectures": [
	"MoAMetricLM"
	],
	"attn_drop": 0.1,
	"attn_heads": 32,
	"bos_token_id": null,
	"conv_kernel": 5,
	"conv_mult": 2,
	"dim": 512,
	"discrepancy_modulation": true,
	"drop_path": 0.0,
	"dtype": "float32",
	"enable_feature_gates": true,
	"enable_router_gates": true,
	"energy_amplification": 3.1415,
	"eos_token_id": 151643,
	"ff_mult": 3,
	"ffn_hidden": 1536,
	"head_feature_heads": 32,
	"layer_scale_init_value": 0.0001,
	"learn_alpha": true,
	"learn_radius": true,
	"lr_rank": 32,
	"maha_init": 1.0,
	"max_position_embeddings": 2048,
	"max_seq_len_cached": 8192,
	"metric": "maha_diag",
	"mixer_hidden": 1536,
	"model_type": "moa_metric",
	"mqa_q_heads": 64,
	"n_branches": 3,
	"n_token_router_heads": 4,
	"num_hidden_layers": 8,
	"num_layers": 8,
	"origin_init_scale": 0.0,
	"pad_token_id": 151643,
	"proj_drop": 0.1,
	"r_basis": 32,
	"radius_init": 5.0,
	"router_bias_heads": 4,
	"router_dropout": 0.1,
	"router_hidden": 1536,
	"router_init_temperature": 1.5,
	"router_temperature": 2.0,
	"router_topk": 3,
	"shared_kv_ratio": 0.55,
	"theta_base": 100000.0,
	"ti_reg_samples": 0,
	"ti_reg_weight": 0.0,
	"transformers_version": "4.56.1",
	"use_balls": true,
	"vocab_size": 151665
	}