Commit
•
3bf09cb
1
Parent(s):
cc7773c
Adding Evaluation Results (#23)
Browse files- Adding Evaluation Results (c9f05a848372dbbb1938e8bb80f9ec149ce2080d)
Co-authored-by: Open LLM Leaderboard PR Bot <leaderboard-pr-bot@users.noreply.huggingface.co>
README.md
CHANGED
@@ -43,4 +43,17 @@ torchrun --nproc_per_node=4 --master_port=3045 train.py \
|
|
43 |
--fsdp "shard_grad_op auto_wrap" \
|
44 |
--fsdp_transformer_layer_cls_to_wrap 'LLaMADecoderLayer' \
|
45 |
--tf32 True --report_to="wandb"
|
46 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
--fsdp "shard_grad_op auto_wrap" \
|
44 |
--fsdp_transformer_layer_cls_to_wrap 'LLaMADecoderLayer' \
|
45 |
--tf32 True --report_to="wandb"
|
46 |
+
```
|
47 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
48 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_chavinlo__alpaca-native)
|
49 |
+
|
50 |
+
| Metric | Value |
|
51 |
+
|-----------------------|---------------------------|
|
52 |
+
| Avg. | 41.96 |
|
53 |
+
| ARC (25-shot) | 52.3 |
|
54 |
+
| HellaSwag (10-shot) | 77.09 |
|
55 |
+
| MMLU (5-shot) | 41.6 |
|
56 |
+
| TruthfulQA (0-shot) | 37.58 |
|
57 |
+
| Winogrande (5-shot) | 69.46 |
|
58 |
+
| GSM8K (5-shot) | 1.44 |
|
59 |
+
| DROP (3-shot) | 14.23 |
|