apple mps: training support for SDXL (ControlNet, LoRA, Dreambooth, T2I)

examples/controlnet/train_controlnet_sdxl.py

@@ -125,7 +125,11 @@ def log_validation(vae, unet, controlnet, args, accelerator, weight_dtype, step,
         )
 
     image_logs = []
-    inference_ctx = contextlib.nullcontext() if is_final_validation else torch.autocast("cuda")
+    inference_ctx = (
+        contextlib.nullcontext()
+        if (is_final_validation or torch.backends.mps.is_available())
+        else torch.autocast("cuda")
+    )
 
     for validation_prompt, validation_image in zip(validation_prompts, validation_images):
         validation_image = Image.open(validation_image).convert("RGB")
@@ -792,6 +796,12 @@ def main(args):
 
     logging_dir = Path(args.output_dir, args.logging_dir)
 
+    if torch.backends.mps.is_available() and args.mixed_precision == "bf16":
+        # due to pytorch#99272, MPS does not yet support bfloat16.
+        raise ValueError(
+            "Mixed precision training with bfloat16 is not supported on MPS. Please use fp16 (recommended) or fp32 instead."
+        )
+
     accelerator_project_config = ProjectConfiguration(project_dir=args.output_dir, logging_dir=logging_dir)
 
     accelerator = Accelerator(

examples/dreambooth/train_dreambooth_lora_sdxl.py

@@ -14,7 +14,6 @@
 # See the License for the specific language governing permissions and
 
 import argparse
-import contextlib
 import gc
 import itertools
 import json
@@ -208,11 +207,18 @@ def log_validation(
     generator = torch.Generator(device=accelerator.device).manual_seed(args.seed) if args.seed else None
     # Currently the context determination is a bit hand-wavy. We can improve it in the future if there's a better
     # way to condition it. Reference: https://github.com/huggingface/diffusers/pull/7126#issuecomment-1968523051
-    inference_ctx = (
-        contextlib.nullcontext() if "playground" in args.pretrained_model_name_or_path else torch.cuda.amp.autocast()
-    )
+    enable_autocast = True
+    if torch.backends.mps.is_available() or (
+        accelerator.mixed_precision == "fp16" or accelerator.mixed_precision == "bf16"
+    ):
+        enable_autocast = False
+    if "playground" in args.pretrained_model_name_or_path:
+        enable_autocast = False
 
-    with inference_ctx:
+    with torch.autocast(
+        accelerator.device.type,
+        enabled=enable_autocast,
+    ):
         images = [pipeline(**pipeline_args, generator=generator).images[0] for _ in range(args.num_validation_images)]
 
     for tracker in accelerator.trackers:
@@ -230,7 +236,8 @@ def log_validation(
             )
 
     del pipeline
-    torch.cuda.empty_cache()
+    if torch.cuda.is_available():
+        torch.cuda.empty_cache()
 
     return images
 
@@ -967,6 +974,12 @@ def main(args):
     if args.do_edm_style_training and args.snr_gamma is not None:
         raise ValueError("Min-SNR formulation is not supported when conducting EDM-style training.")
 
+    if torch.backends.mps.is_available() and args.mixed_precision == "bf16":
+        # due to pytorch#99272, MPS does not yet support bfloat16.
+        raise ValueError(
+            "Mixed precision training with bfloat16 is not supported on MPS. Please use fp16 (recommended) or fp32 instead."
+        )
+
     logging_dir = Path(args.output_dir, args.logging_dir)
 
     accelerator_project_config = ProjectConfiguration(project_dir=args.output_dir, logging_dir=logging_dir)
@@ -1009,7 +1022,8 @@ def main(args):
         cur_class_images = len(list(class_images_dir.iterdir()))
 
         if cur_class_images < args.num_class_images:
-            torch_dtype = torch.float16 if accelerator.device.type == "cuda" else torch.float32
+            has_supported_fp16_accelerator = torch.cuda.is_available() or torch.backends.mps.is_available()
+            torch_dtype = torch.float16 if has_supported_fp16_accelerator else torch.float32
             if args.prior_generation_precision == "fp32":
                 torch_dtype = torch.float32
             elif args.prior_generation_precision == "fp16":
@@ -1134,6 +1148,12 @@ def main(args):
     elif accelerator.mixed_precision == "bf16":
         weight_dtype = torch.bfloat16
 
+    if torch.backends.mps.is_available() and weight_dtype == torch.bfloat16:
+        # due to pytorch#99272, MPS does not yet support bfloat16.
+        raise ValueError(
+            "Mixed precision training with bfloat16 is not supported on MPS. Please use fp16 (recommended) or fp32 instead."
+        )
+
     # Move unet, vae and text_encoder to device and cast to weight_dtype
     unet.to(accelerator.device, dtype=weight_dtype)
 
@@ -1278,7 +1298,7 @@ def load_model_hook(models, input_dir):
 
     # Enable TF32 for faster training on Ampere GPUs,
     # cf https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices
-    if args.allow_tf32:
+    if args.allow_tf32 and torch.cuda.is_available():
         torch.backends.cuda.matmul.allow_tf32 = True
 
     if args.scale_lr:
@@ -1455,7 +1475,8 @@ def compute_text_embeddings(prompt, text_encoders, tokenizers):
     if not args.train_text_encoder and not train_dataset.custom_instance_prompts:
         del tokenizers, text_encoders
         gc.collect()
-        torch.cuda.empty_cache()
+        if torch.cuda.is_available():
+            torch.cuda.empty_cache()
 
     # If custom instance prompts are NOT provided (i.e. the instance prompt is used for all images),
     # pack the statically computed variables appropriately here. This is so that we don't

examples/instruct_pix2pix/train_instruct_pix2pix_sdxl.py

@@ -71,12 +71,7 @@
 
 
 def log_validation(
-    pipeline,
-    args,
-    accelerator,
-    generator,
-    global_step,
-    is_final_validation=False,
+    pipeline, args, accelerator, generator, global_step, is_final_validation=False, enable_autocast=True
 ):
     logger.info(
         f"Running validation... \n Generating {args.num_validation_images} images with prompt:"
@@ -96,7 +91,7 @@ def log_validation(
         else Image.open(image_url_or_path).convert("RGB")
     )(args.val_image_url_or_path)
 
-    with torch.autocast(str(accelerator.device).replace(":0", ""), enabled=accelerator.mixed_precision == "fp16"):
+    with torch.autocast(accelerator.device.type, enabled=enable_autocast):
         edited_images = []
         # Run inference
         for val_img_idx in range(args.num_validation_images):
@@ -497,6 +492,13 @@ def main():
             ),
         )
     logging_dir = os.path.join(args.output_dir, args.logging_dir)
+
+    if torch.backends.mps.is_available() and args.mixed_precision == "bf16":
+        # due to pytorch#99272, MPS does not yet support bfloat16.
+        raise ValueError(
+            "Mixed precision training with bfloat16 is not supported on MPS. Please use fp16 (recommended) or fp32 instead."
+        )
+
     accelerator_project_config = ProjectConfiguration(project_dir=args.output_dir, logging_dir=logging_dir)
     accelerator = Accelerator(
         gradient_accumulation_steps=args.gradient_accumulation_steps,
@@ -981,6 +983,13 @@ def collate_fn(examples):
     if accelerator.is_main_process:
         accelerator.init_trackers("instruct-pix2pix-xl", config=vars(args))
 
+    # Some configurations require autocast to be disabled.
+    enable_autocast = True
+    if torch.backends.mps.is_available() or (
+        accelerator.mixed_precision == "fp16" or accelerator.mixed_precision == "bf16"
+    ):
+        enable_autocast = False
+
     # Train!
     total_batch_size = args.train_batch_size * accelerator.num_processes * args.gradient_accumulation_steps
 
@@ -1193,6 +1202,7 @@ def collate_fn(examples):
                         generator,
                         global_step,
                         is_final_validation=False,
+                        enable_autocast=enable_autocast,
                     )
 
                     if args.use_ema:
@@ -1242,6 +1252,7 @@ def collate_fn(examples):
                 generator,
                 global_step,
                 is_final_validation=True,
+                enable_autocast=enable_autocast,
             )
 
     accelerator.end_training()

examples/text_to_image/train_text_to_image_lora_sdxl.py

@@ -501,6 +501,12 @@ def main(args):
 
     logging_dir = Path(args.output_dir, args.logging_dir)
 
+    if torch.backends.mps.is_available() and args.mixed_precision == "bf16":
+        # due to pytorch#99272, MPS does not yet support bfloat16.
+        raise ValueError(
+            "Mixed precision training with bfloat16 is not supported on MPS. Please use fp16 (recommended) or fp32 instead."
+        )
+
     accelerator_project_config = ProjectConfiguration(project_dir=args.output_dir, logging_dir=logging_dir)
     kwargs = DistributedDataParallelKwargs(find_unused_parameters=True)
     accelerator = Accelerator(
@@ -973,6 +979,13 @@ def collate_fn(examples):
     if accelerator.is_main_process:
         accelerator.init_trackers("text2image-fine-tune", config=vars(args))
 
+    # Some configurations require autocast to be disabled.
+    enable_autocast = True
+    if torch.backends.mps.is_available() or (
+        accelerator.mixed_precision == "fp16" or accelerator.mixed_precision == "bf16"
+    ):
+        enable_autocast = False
+
     # Train!
     total_batch_size = args.train_batch_size * accelerator.num_processes * args.gradient_accumulation_steps
 
@@ -1199,7 +1212,10 @@ def compute_time_ids(original_size, crops_coords_top_left):
                 generator = torch.Generator(device=accelerator.device).manual_seed(args.seed) if args.seed else None
                 pipeline_args = {"prompt": args.validation_prompt}
 
-                with torch.cuda.amp.autocast():
+                with torch.autocast(
+                    accelerator.device.type,
+                    enabled=enable_autocast,
+                ):
                     images = [
                         pipeline(**pipeline_args, generator=generator).images[0]
                         for _ in range(args.num_validation_images)

examples/text_to_image/train_text_to_image_sdxl.py

@@ -590,6 +590,12 @@ def main(args):
 
     accelerator_project_config = ProjectConfiguration(project_dir=args.output_dir, logging_dir=logging_dir)
 
+    if torch.backends.mps.is_available() and args.mixed_precision == "bf16":
+        # due to pytorch#99272, MPS does not yet support bfloat16.
+        raise ValueError(
+            "Mixed precision training with bfloat16 is not supported on MPS. Please use fp16 (recommended) or fp32 instead."
+        )
+
     accelerator = Accelerator(
         gradient_accumulation_steps=args.gradient_accumulation_steps,
         mixed_precision=args.mixed_precision,
@@ -980,6 +986,13 @@ def unwrap_model(model):
         model = model._orig_mod if is_compiled_module(model) else model
         return model
 
+    # Some configurations require autocast to be disabled.
+    enable_autocast = True
+    if torch.backends.mps.is_available() or (
+        accelerator.mixed_precision == "fp16" or accelerator.mixed_precision == "bf16"
+    ):
+        enable_autocast = False
+
     # Train!
     total_batch_size = args.train_batch_size * accelerator.num_processes * args.gradient_accumulation_steps
 
@@ -1213,7 +1226,10 @@ def compute_time_ids(original_size, crops_coords_top_left):
                 generator = torch.Generator(device=accelerator.device).manual_seed(args.seed) if args.seed else None
                 pipeline_args = {"prompt": args.validation_prompt}
 
-                with torch.cuda.amp.autocast():
+                with torch.autocast(
+                    accelerator.device.type,
+                    enabled=enable_autocast,
+                ):
                     images = [
                         pipeline(**pipeline_args, generator=generator, num_inference_steps=25).images[0]
                         for _ in range(args.num_validation_images)
@@ -1268,7 +1284,7 @@ def compute_time_ids(original_size, crops_coords_top_left):
         if args.validation_prompt and args.num_validation_images > 0:
             pipeline = pipeline.to(accelerator.device)
             generator = torch.Generator(device=accelerator.device).manual_seed(args.seed) if args.seed else None
-            with torch.cuda.amp.autocast():
+            with torch.autocast(accelerator.device.type, enabled=enable_autocast):
                 images = [
                     pipeline(args.validation_prompt, num_inference_steps=25, generator=generator).images[0]
                     for _ in range(args.num_validation_images)