Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,18 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
datasets:
|
| 4 |
+
- 40umov/dostoevsky
|
| 5 |
+
language:
|
| 6 |
+
- ru
|
| 7 |
+
base_model:
|
| 8 |
+
- Vikhrmodels/Vikhr-Nemo-12B-Instruct-R-21-09-24
|
| 9 |
+
library_name: unsloth
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
Vikhr-Nemo fine-tuned with contrastive Russian literature.
|
| 13 |
+
|
| 14 |
+
* Base model: https://huggingface.co/Vikhrmodels/Vikhr-Nemo-12B-Instruct-R-21-09-24
|
| 15 |
+
* Dataset: https://huggingface.co/datasets/40umov/dostoevsky
|
| 16 |
+
* Method: [ORPO](https://arxiv.org/abs/2403.07691)
|
| 17 |
+
* Training config: https://github.com/IlyaGusev/saiga/blob/main/configs/models/doestoevsky_nemo_12b_orpo_m1.json
|
| 18 |
+
* WandB: https://wandb.ai/ilyagusev/rulm_self_instruct/runs/4v4pcgej?nw=nwuserilyagusev
|