Text Generation
Transformers
Safetensors
English
babylm
babylm-2026
strict-small
linear-attention
state-tracking
delta-rule
custom_code
Instructions to use SecludedCorner/bind2_0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use SecludedCorner/bind2_0 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="SecludedCorner/bind2_0", trust_remote_code=True)# Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("SecludedCorner/bind2_0", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use SecludedCorner/bind2_0 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "SecludedCorner/bind2_0" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "SecludedCorner/bind2_0", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/SecludedCorner/bind2_0
- SGLang
How to use SecludedCorner/bind2_0 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 "SecludedCorner/bind2_0" \ --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": "SecludedCorner/bind2_0", "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 "SecludedCorner/bind2_0" \ --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": "SecludedCorner/bind2_0", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use SecludedCorner/bind2_0 with Docker Model Runner:
docker model run hf.co/SecludedCorner/bind2_0
| # bind2_0 — delta-rule backbone with a forced bottleneck | |
| Gated-delta-rule backbone; the parameter-matching reference for the bind2_1 arms. | |
| > Research artifact. Not a competition entry. | |
| ## What this is | |
| A **frozen snapshot** of the code for one model generation. It is not a living repository: it is never updated after release. Files shared with another generation are duplicated here on purpose, so that the code closure is complete and nothing has to be fetched from a sibling package. | |
| **Self-contained in code, not in data.** Every module the code imports ships here. Large or non-redistributable inputs -- corpora, official evaluation sets, token streams -- do not, and are listed under "Inputs you must supply" in `BUILD.md` with how to obtain or regenerate each. Do not read the completeness of `src/` as a claim that the package runs with no further downloads. | |
| ## Contents | |
| | | count | | |
| |---|---| | |
| | source files (`src/`) | 2 | | |
| | shipped data (`data/`) | 1 | | |
| | externally-obtained inputs | 1 | | |
| `MANIFEST.json` records, for every artifact, its original path in the working tree, its SHA-256, its size, and the git commit it came from. That is the traceability record: it is what lets you prove which state of the code produced a published result. | |
| ## External inputs (not redistributed here) | |
| - `alaya-strict-small/tokens_u16.bin` — regenerable from the bind1 package's tokenizer chain | |
| ## Building and running | |
| See `BUILD.md`. It states plainly what works, what does not, and why. | |
| ## Availability of this package | |
| If a model card, paper or dataset page points at this package, that pointer must resolve for whoever reads it. A public card that links private code makes a reproducibility promise it cannot keep — and a licence line on unreachable code claims a release that has not happened. Before publishing any pointer to this package, confirm the package is actually reachable by the audience that will read the pointer. | |
| ## Licence | |
| MIT (see `LICENSE`). Code only — model weights and evaluation artifacts are distributed separately and carry their own terms. | |