Commit
•
93ecc1e
1
Parent(s):
a5feffa
Update README.md
Browse files
README.md
CHANGED
@@ -17,16 +17,15 @@ This model is an fine-tuned version of a Llama3 8b Large Language Model (LLM) fo
|
|
17 |
- **Training Dataset**: Cleaned Turkish raw data with 5 billion tokens, custom Turkish instruction sets
|
18 |
- **Training Method**: Initially with DORA, followed by fine-tuning with LORA
|
19 |
|
20 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
|
22 |
-
Metric Value
|
23 |
-
Avg.
|
24 |
-
AI2 Reasoning Challenge_tr
|
25 |
-
HellaSwag_tr
|
26 |
-
MMLU_tr
|
27 |
-
TruthfulQA_tr
|
28 |
-
Winogrande _tr
|
29 |
-
GSM8k_tr
|
30 |
|
31 |
## Usage Examples
|
32 |
|
|
|
17 |
- **Training Dataset**: Cleaned Turkish raw data with 5 billion tokens, custom Turkish instruction sets
|
18 |
- **Training Method**: Initially with DORA, followed by fine-tuning with LORA
|
19 |
|
20 |
+
## Benchmark Results
|
21 |
+
|
22 |
+
- **Winogrande_tr**: 56.16
|
23 |
+
- **TruthfulQA_tr_v0.2**: 47.46
|
24 |
+
- **Mmlu_tr_v0.2**: 46.46
|
25 |
+
- **HellaSwag_tr_v0.2**: 48.87
|
26 |
+
- **GSM8k_tr_v0.2**: 25.43
|
27 |
+
- **Arc_tr_v0.2**: 41.97
|
28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
|
30 |
## Usage Examples
|
31 |
|