Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
Update src/display/about.py
Browse files- src/display/about.py +3 -3
src/display/about.py
CHANGED
@@ -45,12 +45,12 @@ I use LM-Evaluation-Harness-Turkish, a version of the LM Evaluation Harness adap
|
|
45 |
|
46 |
## How to Reproduce Results:
|
47 |
|
48 |
-
|
49 |
-
|
50 |
```python
|
51 |
lm_eval --model vllm --model_args pretrained=Trendyol/Trendyol-LLM-7b-chat-v1.0 --tasks truthfulqa_mc2_tr,truthfulqa_mc1_tr,mmlu_tr,winogrande_tr,gsm8k_tr,arc_challenge_tr,hellaswag_tr --output /workspace/Trendyol-LLM-7b-chat-v1.0
|
52 |
```
|
53 |
-
|
54 |
|
55 |
## Notes:
|
56 |
|
|
|
45 |
|
46 |
## How to Reproduce Results:
|
47 |
|
48 |
+
1) Set Up the repo: Clone the "lm-evaluation-harness_turkish" from https://github.com/malhajar17/lm-evaluation-harness_turkish and follow the installation instructions.
|
49 |
+
2) Run Evaluations: To get the results as on the leaderboard (Some tests might show small variations), use the following command, adjusting for your model. For example, with the Trendyol model:
|
50 |
```python
|
51 |
lm_eval --model vllm --model_args pretrained=Trendyol/Trendyol-LLM-7b-chat-v1.0 --tasks truthfulqa_mc2_tr,truthfulqa_mc1_tr,mmlu_tr,winogrande_tr,gsm8k_tr,arc_challenge_tr,hellaswag_tr --output /workspace/Trendyol-LLM-7b-chat-v1.0
|
52 |
```
|
53 |
+
3) Report Results: I take the average of truthfulqa_mc1_tr and truthfulqa_mc2_tr scores and report it as truthfulqa. The results file generated is then uploaded to the OpenLLM Turkish Leaderboard.
|
54 |
|
55 |
## Notes:
|
56 |
|