Instructions to use StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072 with Transformers:

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

pipe = pipeline("text-generation", model="StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072")

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072

SGLang

How to use StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072 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 "StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072" \
    --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": "StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072",
		"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 "StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072" \
        --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": "StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072 with Docker Model Runner:
```
docker model run hf.co/StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072
```

Wandb Model Name: step2v2_0618_h1024_ffnh9552_numh16_numl8_lr7.812e-03_bs64_ti38146_mlr1.00e-05

This model is part of the StepLaw-N_268M-D_4.0B collection.

Model Specifications

Architecture

Hidden size (H): 1024
Feed-forward network size (FFN): 9552
Attention heads: 16
Layers: 8
Parameter count: 268M

Training Parameters

Learning rate (lr): 7.812e-03
Batch size (bs): 131072
Training iterations: 38146
Training tokens (D): 5.0B

Model Description

StepLaw models are trained with various hyperparameter settings to enable research on scaling laws and hyperparameter optimization. This specific model was trained with learning rate 7.812e-03 and batch size 131072 for 38146 iterations, using a total of 5.0B training tokens.

Usage Example

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True, use_fast=False)
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True)

# Generate text
inputs = tokenizer("A long time ago in a galaxy far, far away", return_tensors="pt")
outputs = model.generate(**inputs, max_length=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Downloads last month: 3

Safetensors

Model size

0.4B params

Tensor type

BF16

Collection including StepLaw/StepLaw-N_268M-D_4.0B-LR7.812e-03-BS131072

StepLaw-N_268M-D_4.0B

Collection

Models with 268M parameters trained with 4.0B tokens. • 118 items • Updated Apr 10, 2025