@@ -145,15 +145,21 @@ def test_qwen3_0_6B(self):
145145 self .__class__ .__name__ ,
146146 "qwen3_0.6B" ,
147147 "examples/torchtitan/configs/MI300X/qwen3_0.6B-pretrain.yaml" ,
148- extra_args = ["--n_layers" , "4" , "--training.steps" , "30 " ],
148+ extra_args = ["--model. n_layers" , "4" , "--training.steps" , "3 " ],
149149 )
150150
151- # def test_qwen3_1_7B(self):
152- # run_script(self.__class__.__name__, "qwen3_1.7B",
153- # "examples/torchtitan/configs/MI300X/qwen3_1.7B-pretrain.yaml",
154- # extra_args=["--n_layers", "4", "--training.steps", "30"])
151+ def test_qwen3_1_7B (self ):
152+ run_script (
153+ self .__class__ .__name__ ,
154+ "qwen3_1.7B" ,
155+ "examples/torchtitan/configs/MI300X/qwen3_1.7B-pretrain.yaml" ,
156+ extra_args = ["--model.n_layers" , "4" , "--training.steps" , "3" ],
157+ )
155158
156- # def test_qwen3_32B(self):
157- # run_script(self.__class__.__name__, "qwen3_32B",
158- # "examples/torchtitan/configs/MI300X/qwen3_32B-pretrain.yaml",
159- # extra_args=["--n_layers", "4", "--training.steps", "30"])
159+ def test_qwen3_32B (self ):
160+ run_script (
161+ self .__class__ .__name__ ,
162+ "qwen3_32B" ,
163+ "examples/torchtitan/configs/MI300X/qwen3_32B-pretrain.yaml" ,
164+ extra_args = ["--model.n_layers" , "4" , "--training.steps" , "3" ],
165+ )
0 commit comments