Instructions to use infosave/Ternary-Bonsai-2-27B-cmf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- cortiq
How to use infosave/Ternary-Bonsai-2-27B-cmf with cortiq:
# one Rust binary, no additional dependencies cargo install cortiq-cli # or a prebuilt binary from github.com/infosave2007/cmf/releases hf download infosave/Ternary-Bonsai-2-27B-cmf --include "*.cmf" --local-dir . ls *.cmf # some repos ship more than one quantization
cortiq run FILE.cmf --prompt "What is the capital of France?"
cortiq serve FILE.cmf --port 8080 # OpenAI-compatible server
- Notebooks
- Google Colab
- Kaggle
| { | |
| "model_filename": "Ternary-Bonsai-2-27B-Q2TP-Affine.cmf", | |
| "model_sha256": "fb0f1a9cdb0434bc9ce50eb6473bc2ca4a71d6994ad61314bbd7e2f11c4e29e8", | |
| "corpus_sha256": "181716ce3f0a71b94f7d5d4428232c241b88f3e26d68f4ed50a191506ddb78b5", | |
| "token_ids_sha256": "6dd2b70cb1d4a18a2955737b9dc3711bc20cd951c48a086ce144dcfbff8eb47b", | |
| "policy": {"input_lengths": [513, 513], "starts": [0, 512], "targets": 1024, "bos": false}, | |
| "runtime": {"backend": "Vulkan", "gpu": "NVIDIA GeForce RTX 4090 24GB", "q2tp_gpu_opt_in": true, "gpu_probe": false}, | |
| "chunks": [ | |
| {"start": 0, "input_len": 513, "nll": 1085.211889936, "count": 512}, | |
| {"start": 512, "input_len": 513, "nll": 771.316238749, "count": 512} | |
| ], | |
| "nll": 1856.528128685, | |
| "mean_nll": 1.813015751, | |
| "ppl": 6.128902833, | |
| "reference_mean_nll": 1.81274952826, | |
| "delta_nats_per_token": 0.000266223 | |
| } | |