Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
datasets:
|
| 3 |
- mlabonne/orpo-dpo-mix-40k
|
|
|
|
| 1 |
+
This model is a fine-tuned version of [meta-llama/Llama-3.2-1B, optimized with ORPO (Optimized Regularization for Prompt Optimization) Trainer. Fine-tuning was performed using a subset of the [meta-llama/Llama-3.2-1B dataset, with only 100 samples selected to enable rapid training with ORPO’s efficient approach.
|
| 2 |
+
|
| 3 |
+
**Fine-tuning Method:** ORPO
|
| 4 |
+
**Dataset:** mlabonne/orpo-dpo-mix-40k
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
**Evaluation**
|
| 8 |
+
|
| 9 |
+
The model was evaluated on the following benchmarks, with the following performance metrics:
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
| Tasks |Version|Filter|n-shot| Metric | |Value | |Stderr|
|
| 13 |
+
|---------|------:|------|-----:|--------|---|-----:|---|-----:|
|
| 14 |
+
|hellaswag| 1|none | 0|acc |↑ | 0.4772 |± | 0.0050 |
|
| 15 |
+
| | |none | 0|acc_norm|↑ |0.6366 |± | 0.0048 |
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
|
| 20 |
---
|
| 21 |
datasets:
|
| 22 |
- mlabonne/orpo-dpo-mix-40k
|