|
Download README.md from logitlab/README: direct link, hf CLI and curl.
- Browser
- Download file 700 Bytes
-
https://huggingface.co/spaces/logitlab/README/resolve/main/README.md
- Command line
-
hf download hf://spaces/logitlab/README/README.md
-
curl -L -o README.md https://huggingface.co/spaces/logitlab/README/resolve/main/README.md
700 Bytes
| title: README | |
| emoji: ⚡ | |
| colorFrom: gray | |
| colorTo: purple | |
| sdk: static | |
| pinned: false | |
| # Logitlab | |
| Small models that answer from their logits. | |
| We build **snap**, an open-source engine that turns a language model into a | |
| typed decision function: one forward pass, a softmax over the option letters, | |
| no generated text. Choice, yes/no, score and numeric questions come back as | |
| calibrated probabilities, locally, through a Jev-compatible | |
| `POST /v1/systemone` API. | |
| - **Engine:** [github.com/emnlmn/snap](https://github.com/emnlmn/snap) (MIT) | |
| - **Docs:** [emnlmn.github.io/snap](https://emnlmn.github.io/snap/) | |
| - **Models:** fine-tuned for the letter readout, released as GGUF. `snap1` is coming soon. |