Use bf16 only if supported

bench.py

@@ -15,7 +15,7 @@
 real_data = True
 seed = 1337
 device = 'cuda' # examples: 'cpu', 'cuda', 'cuda:0', 'cuda:1', etc.
-dtype = 'bfloat16' # 'float32' or 'bfloat16' or 'float16'
+dtype = 'bfloat16' if torch.cuda.is_bf16_supported() else 'float16' # 'float32' or 'bfloat16' or 'float16'
 compile = True # use PyTorch 2.0 to compile the model to be faster
 profile = False # use pytorch profiler, or just simple benchmarking?
 exec(open('configurator.py').read()) # overrides from command line or config file

sample.py

@@ -18,7 +18,7 @@
 top_k = 200 # retain only the top_k most likely tokens, clamp others to have 0 probability
 seed = 1337
 device = 'cuda' # examples: 'cpu', 'cuda', 'cuda:0', 'cuda:1', etc.
-dtype = 'bfloat16' # 'float32' or 'bfloat16' or 'float16'
+dtype = 'bfloat16' if torch.cuda.is_bf16_supported() else 'float16' # 'float32' or 'bfloat16' or 'float16'
 compile = False # use PyTorch 2.0 to compile the model to be faster
 exec(open('configurator.py').read()) # overrides from command line or config file
 # -----------------------------------------------------------------------------

train.py

@@ -70,7 +70,7 @@
 backend = 'nccl' # 'nccl', 'gloo', etc.
 # system
 device = 'cuda' # examples: 'cpu', 'cuda', 'cuda:0', 'cuda:1' etc., or try 'mps' on macbooks
-dtype = 'bfloat16' # 'float32', 'bfloat16', or 'float16', the latter will auto implement a GradScaler
+dtype = 'bfloat16' if torch.cuda.is_bf16_supported() else 'float16' # 'float32', 'bfloat16', or 'float16', the latter will auto implement a GradScaler
 compile = True # use PyTorch 2.0 to compile the model to be faster
 # -----------------------------------------------------------------------------
 config_keys = [k for k,v in globals().items() if not k.startswith('_') and isinstance(v, (int, float, bool, str))]