various tests fixes for flakey tests (#2110)

* add mhenrichsen/alpaca_2k_test with revision dataset download fixture for flaky tests

* log slowest tests

* pin pynvml==11.5.3

* fix load local hub path

* optimize for speed w smaller models and val_set_size

* replace pynvml

* make the resume from checkpoint e2e faster

* make tests smaller

tests/e2e/patched/test_fa_xentropy.py

@@ -51,11 +51,11 @@ class TestFAXentropyLlama:
                 "flash_attn_cross_entropy": True,
                 "load_in_8bit": True,
                 "adapter": "lora",
-                "lora_r": 32,
-                "lora_alpha": 64,
+                "lora_r": 8,
+                "lora_alpha": 16,
                 "lora_dropout": 0.05,
                 "lora_target_linear": True,
-                "val_set_size": 0.2,
+                "val_set_size": 0.05,
                 "special_tokens": {
                     "pad_token": "<|endoftext|>",
                 },