phi-3.5 rpp results

logs/l40-4gpu-8.txt

@@ -175,3 +175,12 @@ You seem to be using the pipelines sequentially on GPU. In order to maximize eff
 The `seen_tokens` attribute is deprecated and will be removed in v4.41. Use the `cache_position` model input instead.
 2024-08-26 14:30:49,269 [WARNING] [logging.py:328] You are not running the flash-attention implementation, expect numerical differences.
 You seem to be using the pipelines sequentially on GPU. In order to maximize efficiency please use a dataset
+[nltk_data] Downloading package punkt to
+[nltk_data]     /common/home/users/d/dh.huang.2023/nltk_data...
+[nltk_data]   Package punkt is already up-to-date!
+2024-08-26 19:44:35,757 [WARNING] [modeling_phi3.py:62] `flash-attention` package not found, consider installing for better performance: No module named 'flash_attn'.
+2024-08-26 19:44:35,757 [WARNING] [modeling_phi3.py:66] Current `flash-attention` does not support `window_size`. Either upgrade or use `attn_implementation='eager'`.
+
+The `seen_tokens` attribute is deprecated and will be removed in v4.41. Use the `cache_position` model input instead.
+2024-08-26 19:45:11,738 [WARNING] [logging.py:328] You are not running the flash-attention implementation, expect numerical differences.
+You seem to be using the pipelines sequentially on GPU. In order to maximize efficiency please use a dataset