Text Classification
PEFT
Safetensors
decision-making
system-one
calibrated-probabilities
lora
qwen3
indic
hinglish
hindi
non-generative
Instructions to use cmul8-hf/nirnaya with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use cmul8-hf/nirnaya with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
Remove Repository contents section from v1.0 card
Browse files
README.md
CHANGED
|
@@ -152,17 +152,6 @@ Until the inference package lands, the checkpoint can be loaded via the training
|
|
| 152 |
|
| 153 |
---
|
| 154 |
|
| 155 |
-
## Repository contents
|
| 156 |
-
|
| 157 |
-
- **LoRA adapter** (safetensors, ~130 MB fp32) — merges onto the Qwen3-4B backbone at load time.
|
| 158 |
-
- **Decision heads** — PointerHead and NoulHead weights (`heads.pt`).
|
| 159 |
-
- **Config JSON** — model configuration (embedding dims, format version, temperatures placeholder).
|
| 160 |
-
- **`meta.json`** — training metadata (best_loss, step count, timestamp).
|
| 161 |
-
|
| 162 |
-
**Not included:** the Qwen3-4B backbone weights (fetched at load time from `Qwen/Qwen3-4B` under Apache-2.0). You will need permission to download the base model.
|
| 163 |
-
|
| 164 |
-
---
|
| 165 |
-
|
| 166 |
## License
|
| 167 |
|
| 168 |
- **Nirṇaya v1.0 weights (this repository):** [Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)](https://creativecommons.org/licenses/by-nc/4.0/).
|
|
|
|
| 152 |
|
| 153 |
---
|
| 154 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 155 |
## License
|
| 156 |
|
| 157 |
- **Nirṇaya v1.0 weights (this repository):** [Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)](https://creativecommons.org/licenses/by-nc/4.0/).
|