Instructions to use ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47 with Transformers:

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

pipe = pipeline("text-generation", model="ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47", trust_remote_code=True)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47

SGLang

How to use ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47 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 "ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47" \
    --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": "ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47",
		"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 "ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47" \
        --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": "ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47 with Docker Model Runner:
```
docker model run hf.co/ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47
```

seedvar-looped-moe-1e18-d704-seed47

Seed-variance run for Sparse Layers are Critical to Scaling Looped Language Models (arXiv:2605.09165), trained to measure run-to-run noise in the 1e18-FLOP benchmark numbers.


architecture	`looped-moe`
d_model	704
effective layers	16
compute budget	1e18 FLOPs
training steps	60,782
parameters (stored)	216,427,904
peak LR	0.01
batch size	16
data-order seed	47
init seed	42
muP width_ratio	5.5 (d_base=128)

Important: what varies across these four seeds

Only the training data order. The initialization seed is fixed at 42 for all four runs, as is the validation-batch order. The spread across seeds 42-45 therefore measures data-order variance, which is a lower bound on full run-to-run variance — a study that also varied initialization would be expected to show equal or greater spread. Do not read these error bars as total training noise.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
m = AutoModelForCausalLM.from_pretrained(
    "ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47", trust_remote_code=True)
tok = AutoTokenizer.from_pretrained("gpt2")

Evaluated with OLMES 5-shot core_9mcqa::olmes. When evaluating, pass max_length=1024 — the RoPE buffer is sized to the 1024-token training context and longer sequences overflow it.

Downloads last month: 17

Safetensors

Model size

0.2B params

Tensor type

F32

Paper for ml-ryanlee/seedvar-looped-moe-1e18-d704-seed47

Sparse Layers are Critical to Scaling Looped Language Models

Paper • 2605.09165 • Published May 9