Instructions to use briscoooe/tiny-cube-deep20 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use briscoooe/tiny-cube-deep20 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="briscoooe/tiny-cube-deep20")# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("briscoooe/tiny-cube-deep20") model = AutoModelForCausalLM.from_pretrained("briscoooe/tiny-cube-deep20", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use briscoooe/tiny-cube-deep20 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "briscoooe/tiny-cube-deep20" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "briscoooe/tiny-cube-deep20", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/briscoooe/tiny-cube-deep20
- SGLang
How to use briscoooe/tiny-cube-deep20 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 "briscoooe/tiny-cube-deep20" \ --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": "briscoooe/tiny-cube-deep20", "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 "briscoooe/tiny-cube-deep20" \ --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": "briscoooe/tiny-cube-deep20", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use briscoooe/tiny-cube-deep20 with Docker Model Runner:
docker model run hf.co/briscoooe/tiny-cube-deep20
| """ | |
| Checks a generated dataset by replaying it: applies each recorded solution to its | |
| recorded state and confirms the cube ends solved. | |
| This is the one check worth running on every dataset before training on it. A | |
| generator bug that corrupts states or mislabels them is silent otherwise -- the | |
| model simply learns the wrong function and the failure only surfaces much later, | |
| as a bad benchmark number that looks like a modelling problem. | |
| """ | |
| import argparse, json, sys | |
| from gen_data import SOLVED, apply_sequence | |
| def main(): | |
| p = argparse.ArgumentParser(description=__doc__) | |
| p.add_argument("path") | |
| p.add_argument("--limit", type=int, default=0, help="Check only the first N rows (0 = all).") | |
| args = p.parse_args() | |
| checked = failed = 0 | |
| lengths = [] | |
| with open(args.path) as fh: | |
| for line in fh: | |
| if args.limit and checked >= args.limit: | |
| break | |
| row = json.loads(line) | |
| if len(row["state"]) != 54: | |
| failed += 1 | |
| continue | |
| moves = row["solution"].split() | |
| lengths.append(len(moves)) | |
| if apply_sequence(row["state"], moves) != SOLVED: | |
| failed += 1 | |
| checked += 1 | |
| mean = sum(lengths) / len(lengths) if lengths else 0 | |
| print(f"checked {checked}, solved {checked - failed}, FAILED {failed}") | |
| print(f"solution length: mean {mean:.2f} min {min(lengths, default=0)} max {max(lengths, default=0)}") | |
| sys.exit(1 if failed else 0) | |
| if __name__ == "__main__": | |
| main() | |