Spaces:
Running
Running
Create organization card
Browse files
README.md
CHANGED
|
@@ -1,10 +1,30 @@
|
|
| 1 |
---
|
| 2 |
-
title:
|
| 3 |
-
emoji: 馃憗
|
| 4 |
-
colorFrom: red
|
| 5 |
-
colorTo: indigo
|
| 6 |
sdk: static
|
| 7 |
pinned: false
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Tuned Tensor
|
|
|
|
|
|
|
|
|
|
| 3 |
sdk: static
|
| 4 |
pinned: false
|
| 5 |
---
|
| 6 |
|
| 7 |
+
# Tuned Tensor
|
| 8 |
+
|
| 9 |
+
**Behaviour control infrastructure for open-weight language models.**
|
| 10 |
+
|
| 11 |
+
Tuned Tensor helps teams define how a model should behave, fine-tune open-weight models against that behaviour spec, evaluate the result, and feed the findings back into the next run.
|
| 12 |
+
|
| 13 |
+
## Workflow
|
| 14 |
+
|
| 15 |
+
1. **Specify behaviour** with system prompts, guidelines, constraints, examples, and a base model.
|
| 16 |
+
2. **Fine-tune** by compiling training data and tuning supported open-weight models.
|
| 17 |
+
3. **Evaluate** with per-example scores, failure inspection, and regression checks.
|
| 18 |
+
4. **Iterate** by refining the spec and starting the next run.
|
| 19 |
+
|
| 20 |
+
## Get started
|
| 21 |
+
|
| 22 |
+
```bash
|
| 23 |
+
npm i -g @tuned-tensor/cli
|
| 24 |
+
tt init
|
| 25 |
+
tt runs start <spec-id>
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
Every feature is available through the open-source `tt` CLI or the REST API.
|
| 29 |
+
|
| 30 |
+
[Website](https://tunedtensor.com/) 路 [Docs](https://tunedtensor.com/docs) 路 [CLI](https://github.com/tunedtensor/tuned-tensor-cli) 路 [Agent skill](https://tunedtensor.com/skill.md)
|