Multilingual-Transfer
Collection
Pretraining models to find what allows multilingual transfer • 24 items • Updated • 2
llama3-7b-en-ru
LLaMA-3 7B pretrained on English-Russian bilingual data, 134k steps.
Validation Results
Final checkpoint: step 133,600
| Validation set | Cross-entropy loss (nats) | Perplexity |
|---|---|---|
| English (en) | 2.2223 | 9.23 |
RU (ru) |
0.7653 | 2.15 |
Note: Russian validation loss is on a Russian validation set (different scale from Arabic models).
Evaluation Results
EEE-format evaluation results are stored under eval_results/eee/ in this repository.
Tasks: Global MMLU (EN/AR/RU), PIQA, ECLeKTic, Fictive Entity (2-rate mix).
Citation
Part of the The-CoLab multilingual-transfer collection.
- Downloads last month
- 775
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support