Enable loading model from hub that has already been converted #35

Job	Run time
build (3.10, macos-15, test_modeling_gemma2.py)	55m 27s
build (3.10, macos-15, test_modeling_gemma.py)	53m 57s
build (3.10, macos-15, test_modeling_llama.py)	9m 21s
build (3.10, macos-15, test_modeling_olmo.py)	6m 52s
build (3.10, macos-15, test_modeling.py)	2m 12s
build (3.10, macos-15, test_modeling_qwen2.py)	3m 52s
build (3.11, macos-15, test_modeling_gemma2.py)	1h 1m 22s
build (3.11, macos-15, test_modeling_gemma.py)	52m 54s
build (3.11, macos-15, test_modeling_llama.py)	10m 22s
build (3.11, macos-15, test_modeling_olmo.py)	7m 15s
build (3.11, macos-15, test_modeling.py)	1m 17s
build (3.11, macos-15, test_modeling_qwen2.py)	4m 30s
build (3.12, macos-15, test_modeling_gemma2.py)	59m 48s
build (3.12, macos-15, test_modeling_gemma.py)	52m 27s
build (3.12, macos-15, test_modeling_llama.py)	10m 47s
build (3.12, macos-15, test_modeling_olmo.py)	6m 40s
build (3.12, macos-15, test_modeling.py)	1m 1s
build (3.12, macos-15, test_modeling_qwen2.py)	4m 21s
	6h 44m 25s

Provide feedback