Merge branch 'main' into layerwise-upcasting-hook

Author: a-r-r-o-w

src/diffusers/models/transformers/sana_transformer.py

@@ -82,6 +82,20 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         return hidden_states
 
 
+class SanaModulatedNorm(nn.Module):
+    def __init__(self, dim: int, elementwise_affine: bool = False, eps: float = 1e-6):
+        super().__init__()
+        self.norm = nn.LayerNorm(dim, elementwise_affine=elementwise_affine, eps=eps)
+
+    def forward(
+        self, hidden_states: torch.Tensor, temb: torch.Tensor, scale_shift_table: torch.Tensor
+    ) -> torch.Tensor:
+        hidden_states = self.norm(hidden_states)
+        shift, scale = (scale_shift_table[None] + temb[:, None].to(scale_shift_table.device)).chunk(2, dim=1)
+        hidden_states = hidden_states * (1 + scale) + shift
+        return hidden_states
+
+
 class SanaTransformerBlock(nn.Module):
     r"""
     Transformer block introduced in [Sana](https://huggingface.co/papers/2410.10629).
@@ -221,7 +235,7 @@ class SanaTransformer2DModel(ModelMixin, ConfigMixin, PeftAdapterMixin):
     """
 
     _supports_gradient_checkpointing = True
-    _no_split_modules = ["SanaTransformerBlock", "PatchEmbed"]
+    _no_split_modules = ["SanaTransformerBlock", "PatchEmbed", "SanaModulatedNorm"]
     _precision_sensitive_module_patterns = ["patch_embed", "norm"]
 
     @register_to_config
@@ -289,8 +303,7 @@ def __init__(
 
         # 4. Output blocks
         self.scale_shift_table = nn.Parameter(torch.randn(2, inner_dim) / inner_dim**0.5)
-
-        self.norm_out = nn.LayerNorm(inner_dim, elementwise_affine=False, eps=1e-6)
+        self.norm_out = SanaModulatedNorm(inner_dim, elementwise_affine=False, eps=1e-6)
         self.proj_out = nn.Linear(inner_dim, patch_size * patch_size * out_channels)
 
         self.gradient_checkpointing = False
@@ -463,13 +476,8 @@ def custom_forward(*inputs):
                 )
 
         # 3. Normalization
-        shift, scale = (
-            self.scale_shift_table[None] + embedded_timestep[:, None].to(self.scale_shift_table.device)
-        ).chunk(2, dim=1)
-        hidden_states = self.norm_out(hidden_states)
+        hidden_states = self.norm_out(hidden_states, embedded_timestep, self.scale_shift_table)
 
-        # 4. Modulation
-        hidden_states = hidden_states * (1 + scale) + shift
         hidden_states = self.proj_out(hidden_states)
 
         # 5. Unpatchify

tests/models/test_modeling_common.py

@@ -31,7 +31,7 @@
 import requests_mock
 import torch
 import torch.nn as nn
-from accelerate.utils.modeling import _get_proper_dtype, dtype_byte_size
+from accelerate.utils.modeling import _get_proper_dtype, compute_module_sizes, dtype_byte_size
 from huggingface_hub import ModelCard, delete_repo, snapshot_download
 from huggingface_hub.utils import is_jinja_available
 from parameterized import parameterized
@@ -1094,7 +1094,7 @@ def test_cpu_offload(self):
         torch.manual_seed(0)
         base_output = model(**inputs_dict)
 
-        model_size = compute_module_persistent_sizes(model)[""]
+        model_size = compute_module_sizes(model)[""]
         # We test several splits of sizes to make sure it works.
         max_gpu_sizes = [int(p * model_size) for p in self.model_split_percents[1:]]
         with tempfile.TemporaryDirectory() as tmp_dir:
@@ -1124,7 +1124,7 @@ def test_disk_offload_without_safetensors(self):
         torch.manual_seed(0)
         base_output = model(**inputs_dict)
 
-        model_size = compute_module_persistent_sizes(model)[""]
+        model_size = compute_module_sizes(model)[""]
         with tempfile.TemporaryDirectory() as tmp_dir:
             model.cpu().save_pretrained(tmp_dir, safe_serialization=False)
 
@@ -1158,7 +1158,7 @@ def test_disk_offload_with_safetensors(self):
         torch.manual_seed(0)
         base_output = model(**inputs_dict)
 
-        model_size = compute_module_persistent_sizes(model)[""]
+        model_size = compute_module_sizes(model)[""]
         with tempfile.TemporaryDirectory() as tmp_dir:
             model.cpu().save_pretrained(tmp_dir)
 
@@ -1186,7 +1186,7 @@ def test_model_parallelism(self):
         torch.manual_seed(0)
         base_output = model(**inputs_dict)
 
-        model_size = compute_module_persistent_sizes(model)[""]
+        model_size = compute_module_sizes(model)[""]
         # We test several splits of sizes to make sure it works.
         max_gpu_sizes = [int(p * model_size) for p in self.model_split_percents[1:]]
         with tempfile.TemporaryDirectory() as tmp_dir:
@@ -1197,6 +1197,7 @@ def test_model_parallelism(self):
                 new_model = self.model_class.from_pretrained(tmp_dir, device_map="auto", max_memory=max_memory)
                 # Making sure part of the model will actually end up offloaded
                 self.assertSetEqual(set(new_model.hf_device_map.values()), {0, 1})
+                print(f" new_model.hf_device_map:{new_model.hf_device_map}")
 
                 self.check_device_map_is_respected(new_model, new_model.hf_device_map)
 

tests/models/transformers/test_models_transformer_sana.py

@@ -14,7 +14,6 @@
 
 import unittest
 
-import pytest
 import torch
 
 from diffusers import SanaTransformer2DModel
@@ -33,6 +32,7 @@ class SanaTransformerTests(ModelTesterMixin, unittest.TestCase):
     model_class = SanaTransformer2DModel
     main_input_name = "hidden_states"
     uses_custom_attn_processor = True
+    model_split_percents = [0.7, 0.7, 0.9]
 
     @property
     def dummy_input(self):
@@ -81,27 +81,3 @@ def prepare_init_args_and_inputs_for_common(self):
     def test_gradient_checkpointing_is_applied(self):
         expected_set = {"SanaTransformer2DModel"}
         super().test_gradient_checkpointing_is_applied(expected_set=expected_set)
-
-    @pytest.mark.xfail(
-        condition=torch.device(torch_device).type == "cuda",
-        reason="Test currently fails.",
-        strict=True,
-    )
-    def test_cpu_offload(self):
-        return super().test_cpu_offload()
-
-    @pytest.mark.xfail(
-        condition=torch.device(torch_device).type == "cuda",
-        reason="Test currently fails.",
-        strict=True,
-    )
-    def test_disk_offload_with_safetensors(self):
-        return super().test_disk_offload_with_safetensors()
-
-    @pytest.mark.xfail(
-        condition=torch.device(torch_device).type == "cuda",
-        reason="Test currently fails.",
-        strict=True,
-    )
-    def test_disk_offload_without_safetensors(self):
-        return super().test_disk_offload_without_safetensors()