Instructions to use StepLaw/StepLaw-N_268M-D_4.0B-LR9.766e-04-BS524288 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-LR9.766e-04-BS524288 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-LR9.766e-04-BS524288")

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("StepLaw/StepLaw-N_268M-D_4.0B-LR9.766e-04-BS524288", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use StepLaw/StepLaw-N_268M-D_4.0B-LR9.766e-04-BS524288 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-LR9.766e-04-BS524288"
# 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-LR9.766e-04-BS524288",
		"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-LR9.766e-04-BS524288

SGLang

How to use StepLaw/StepLaw-N_268M-D_4.0B-LR9.766e-04-BS524288 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-LR9.766e-04-BS524288" \
    --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-LR9.766e-04-BS524288",
		"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-LR9.766e-04-BS524288" \
        --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-LR9.766e-04-BS524288",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use StepLaw/StepLaw-N_268M-D_4.0B-LR9.766e-04-BS524288 with Docker Model Runner:
```
docker model run hf.co/StepLaw/StepLaw-N_268M-D_4.0B-LR9.766e-04-BS524288
```

brain-zhang commited on Apr 15, 2025

Commit

c27fbc7

verified ·

1 Parent(s): 0b6af71

Upload README.md with huggingface_hub

Browse files

Files changed (1) hide show

README.md +4 -7

README.md CHANGED Viewed

@@ -23,17 +23,17 @@ This model is part of the [StepLaw-N_268M-D_4.0B](https://huggingface.co/collect
 - **Feed-forward network size (FFN)**: 9552
 - **Attention heads**: 16
 - **Layers**: 8
-- **Parameter count**: 268MM
 ### Training Parameters
 - **Learning rate (lr)**: 9.766e-04
-- **Batch size (bs)**: 256
 - **Training iterations**: 9536
 - **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 9.766e-04 and batch size 256 for 9536 iterations, using a total of 5.0B training tokens.
 ## Usage Example
@@ -48,7 +48,4 @@ model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True)
 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))
-```## Part of StepLaw Project
-StepLaw is an initiative to provide thousands of models for optimal hyperparameter research.
-Visit [StepLaw Project](https://step-law.github.io/) for more information.

 - **Feed-forward network size (FFN)**: 9552
 - **Attention heads**: 16
 - **Layers**: 8
+- **Parameter count**: 268M
 ### Training Parameters
 - **Learning rate (lr)**: 9.766e-04
+- **Batch size (bs)**: 524288
 - **Training iterations**: 9536
 - **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 9.766e-04 and batch size 524288 for 9536 iterations, using a total of 5.0B training tokens.
 ## Usage Example
 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))
+```