[bitsandbbytes] follow-ups (huggingface#9730)

* bnb follow ups.

* add a warning when dtypes mismatch.

* fx-copies

* clear cache.

* check_if_quantized_param

* add a check on shape.

* updates

* docs

* improve readability.

* resources.

* fix

docs/source/en/quantization/bitsandbytes.md

@@ -59,19 +59,7 @@ model_8bit = FluxTransformer2DModel.from_pretrained(
 model_8bit.transformer_blocks.layers[-1].norm2.weight.dtype
 ```
 
-Once a model is quantized, you can push the model to the Hub with the [`~ModelMixin.push_to_hub`] method. The quantization `config.json` file is pushed first, followed by the quantized model weights.
-
-```py
-from diffusers import FluxTransformer2DModel, BitsAndBytesConfig
-
-quantization_config = BitsAndBytesConfig(load_in_8bit=True)
-
-model_8bit = FluxTransformer2DModel.from_pretrained(
-    "black-forest-labs/FLUX.1-dev", 
-    subfolder="transformer",
-    quantization_config=quantization_config
-)
-```
+Once a model is quantized, you can push the model to the Hub with the [`~ModelMixin.push_to_hub`] method. The quantization `config.json` file is pushed first, followed by the quantized model weights. You can also save the serialized 4-bit models locally with [`~ModelMixin.save_pretrained`].
 
 </hfoption>
 <hfoption id="4-bit">
@@ -131,7 +119,7 @@ from diffusers import FluxTransformer2DModel, BitsAndBytesConfig
 quantization_config = BitsAndBytesConfig(load_in_4bit=True)
 
 model_4bit = FluxTransformer2DModel.from_pretrained(
-    "sayakpaul/flux.1-dev-nf4-pkg", subfolder="transformer"
+    "hf-internal-testing/flux.1-dev-nf4-pkg", subfolder="transformer"
 )
 ```
 
@@ -264,4 +252,9 @@ double_quant_model = SD3Transformer2DModel.from_pretrained(
     quantization_config=double_quant_config,
 )
 model.dequantize()
-```
+```
+
+## Resources
+
+* [End-to-end notebook showing Flux.1 Dev inference in a free-tier Colab](https://gist.github.com/sayakpaul/c76bd845b48759e11687ac550b99d8b4)
+* [Training](https://gist.github.com/sayakpaul/05afd428bc089b47af7c016e42004527)

src/diffusers/models/model_loading_utils.py

@@ -211,21 +211,28 @@ def load_model_dict_into_meta(
                     set_module_kwargs["dtype"] = dtype
 
         # bnb params are flattened.
-        if not is_quant_method_bnb and empty_state_dict[param_name].shape != param.shape:
-            model_name_or_path_str = f"{model_name_or_path} " if model_name_or_path is not None else ""
-            raise ValueError(
-                f"Cannot load {model_name_or_path_str}because {param_name} expected shape {empty_state_dict[param_name]}, but got {param.shape}. If you want to instead overwrite randomly initialized weights, please make sure to pass both `low_cpu_mem_usage=False` and `ignore_mismatched_sizes=True`. For more information, see also: https://github.com/huggingface/diffusers/issues/1619#issuecomment-1345604389 as an example."
-            )
+        if empty_state_dict[param_name].shape != param.shape:
+            if (
+                is_quant_method_bnb
+                and hf_quantizer.pre_quantized
+                and hf_quantizer.check_if_quantized_param(model, param, param_name, state_dict, param_device=device)
+            ):
+                hf_quantizer.check_quantized_param_shape(param_name, empty_state_dict[param_name].shape, param.shape)
+            elif not is_quant_method_bnb:
+                model_name_or_path_str = f"{model_name_or_path} " if model_name_or_path is not None else ""
+                raise ValueError(
+                    f"Cannot load {model_name_or_path_str} because {param_name} expected shape {empty_state_dict[param_name]}, but got {param.shape}. If you want to instead overwrite randomly initialized weights, please make sure to pass both `low_cpu_mem_usage=False` and `ignore_mismatched_sizes=True`. For more information, see also: https://github.com/huggingface/diffusers/issues/1619#issuecomment-1345604389 as an example."
+                )
 
-        if not is_quantized or (
-            not hf_quantizer.check_quantized_param(model, param, param_name, state_dict, param_device=device)
+        if is_quantized and (
+            hf_quantizer.check_if_quantized_param(model, param, param_name, state_dict, param_device=device)
         ):
+            hf_quantizer.create_quantized_param(model, param, param_name, device, state_dict, unexpected_keys)
+        else:
             if accepts_dtype:
                 set_module_tensor_to_device(model, param_name, device, value=param, **set_module_kwargs)
             else:
                 set_module_tensor_to_device(model, param_name, device, value=param)
-        else:
-            hf_quantizer.create_quantized_param(model, param, param_name, device, state_dict, unexpected_keys)
 
     return unexpected_keys
 

src/diffusers/quantizers/__init__.py

@@ -12,5 +12,5 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from .auto import DiffusersAutoQuantizationConfig, DiffusersAutoQuantizer
+from .auto import DiffusersAutoQuantizer
 from .base import DiffusersQuantizer

src/diffusers/quantizers/auto.py

@@ -33,10 +33,10 @@
 }
 
 
-class DiffusersAutoQuantizationConfig:
+class DiffusersAutoQuantizer:
     """
-    The auto diffusers quantization config class that takes care of automatically dispatching to the correct
-    quantization config given a quantization config stored in a dictionary.
+     The auto diffusers quantizer class that takes care of automatically instantiating to the correct
+    `DiffusersQuantizer` given the `QuantizationConfig`.
     """
 
     @classmethod
@@ -60,31 +60,11 @@ def from_dict(cls, quantization_config_dict: Dict):
         target_cls = AUTO_QUANTIZATION_CONFIG_MAPPING[quant_method]
         return target_cls.from_dict(quantization_config_dict)
 
-    @classmethod
-    def from_pretrained(cls, pretrained_model_name_or_path, **kwargs):
-        model_config = cls.load_config(pretrained_model_name_or_path, **kwargs)
-        if getattr(model_config, "quantization_config", None) is None:
-            raise ValueError(
-                f"Did not found a `quantization_config` in {pretrained_model_name_or_path}. Make sure that the model is correctly quantized."
-            )
-        quantization_config_dict = model_config.quantization_config
-        quantization_config = cls.from_dict(quantization_config_dict)
-        # Update with potential kwargs that are passed through from_pretrained.
-        quantization_config.update(kwargs)
-        return quantization_config
-
-
-class DiffusersAutoQuantizer:
-    """
-     The auto diffusers quantizer class that takes care of automatically instantiating to the correct
-    `DiffusersQuantizer` given the `QuantizationConfig`.
-    """
-
     @classmethod
     def from_config(cls, quantization_config: Union[QuantizationConfigMixin, Dict], **kwargs):
         # Convert it to a QuantizationConfig if the q_config is a dict
         if isinstance(quantization_config, dict):
-            quantization_config = DiffusersAutoQuantizationConfig.from_dict(quantization_config)
+            quantization_config = cls.from_dict(quantization_config)
 
         quant_method = quantization_config.quant_method
 
@@ -107,7 +87,16 @@ def from_config(cls, quantization_config: Union[QuantizationConfigMixin, Dict],
 
     @classmethod
     def from_pretrained(cls, pretrained_model_name_or_path, **kwargs):
-        quantization_config = DiffusersAutoQuantizationConfig.from_pretrained(pretrained_model_name_or_path, **kwargs)
+        model_config = cls.load_config(pretrained_model_name_or_path, **kwargs)
+        if getattr(model_config, "quantization_config", None) is None:
+            raise ValueError(
+                f"Did not found a `quantization_config` in {pretrained_model_name_or_path}. Make sure that the model is correctly quantized."
+            )
+        quantization_config_dict = model_config.quantization_config
+        quantization_config = cls.from_dict(quantization_config_dict)
+        # Update with potential kwargs that are passed through from_pretrained.
+        quantization_config.update(kwargs)
+
         return cls.from_config(quantization_config)
 
     @classmethod
@@ -129,7 +118,7 @@ def merge_quantization_configs(
             warning_msg = ""
 
         if isinstance(quantization_config, dict):
-            quantization_config = DiffusersAutoQuantizationConfig.from_dict(quantization_config)
+            quantization_config = cls.from_dict(quantization_config)
 
         if warning_msg != "":
             warnings.warn(warning_msg)

src/diffusers/quantizers/base.py

@@ -134,7 +134,7 @@ def adjust_max_memory(self, max_memory: Dict[str, Union[int, str]]) -> Dict[str,
         """adjust max_memory argument for infer_auto_device_map() if extra memory is needed for quantization"""
         return max_memory
 
-    def check_quantized_param(
+    def check_if_quantized_param(
         self,
         model: "ModelMixin",
         param_value: "torch.Tensor",
@@ -152,10 +152,13 @@ def create_quantized_param(self, *args, **kwargs) -> "torch.nn.Parameter":
         """
         takes needed components from state_dict and creates quantized param.
         """
-        if not hasattr(self, "check_quantized_param"):
-            raise AttributeError(
-                f"`.create_quantized_param()` method is not supported by quantizer class {self.__class__.__name__}."
-            )
+        return
+
+    def check_quantized_param_shape(self, *args, **kwargs):
+        """
+        checks if the quantized param has expected shape.
+        """
+        return True
 
     def validate_environment(self, *args, **kwargs):
         """

src/diffusers/quantizers/bitsandbytes/bnb_quantizer.py

@@ -106,7 +106,7 @@ def adjust_target_dtype(self, target_dtype: "torch.dtype") -> "torch.dtype":
         else:
             raise ValueError(f"Wrong `target_dtype` ({target_dtype}) provided.")
 
-    def check_quantized_param(
+    def check_if_quantized_param(
         self,
         model: "ModelMixin",
         param_value: "torch.Tensor",
@@ -204,6 +204,16 @@ def create_quantized_param(
 
         module._parameters[tensor_name] = new_value
 
+    def check_quantized_param_shape(self, param_name, current_param_shape, loaded_param_shape):
+        n = current_param_shape.numel()
+        inferred_shape = (n,) if "bias" in param_name else ((n + 1) // 2, 1)
+        if loaded_param_shape != inferred_shape:
+            raise ValueError(
+                f"Expected the flattened shape of the current param ({param_name}) to be {loaded_param_shape} but is {inferred_shape}."
+            )
+        else:
+            return True
+
     def adjust_max_memory(self, max_memory: Dict[str, Union[int, str]]) -> Dict[str, Union[int, str]]:
         # need more space for buffers that are created during quantization
         max_memory = {key: val * 0.90 for key, val in max_memory.items()}
@@ -330,7 +340,6 @@ def __init__(self, quantization_config, **kwargs):
         if self.quantization_config.llm_int8_skip_modules is not None:
             self.modules_to_not_convert = self.quantization_config.llm_int8_skip_modules
 
-    # Copied from diffusers.quantizers.bitsandbytes.bnb_quantizer.BnB4BitDiffusersQuantizer.validate_environment with 4-bit->8-bit
     def validate_environment(self, *args, **kwargs):
         if not torch.cuda.is_available():
             raise RuntimeError("No GPU found. A GPU is needed for quantization.")
@@ -404,7 +413,7 @@ def adjust_target_dtype(self, target_dtype: "torch.dtype") -> "torch.dtype":
             logger.info("target_dtype {target_dtype} is replaced by `torch.int8` for 8-bit BnB quantization")
         return torch.int8
 
-    def check_quantized_param(
+    def check_if_quantized_param(
         self,
         model: "ModelMixin",
         param_value: "torch.Tensor",

tests/quantization/bnb/test_4bit.py

@@ -13,10 +13,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 import gc
+import os
 import tempfile
 import unittest
 
 import numpy as np
+import safetensors.torch
 
 from diffusers import BitsAndBytesConfig, DiffusionPipeline, FluxTransformer2DModel, SD3Transformer2DModel
 from diffusers.utils import logging
@@ -118,6 +120,9 @@ def get_dummy_inputs(self):
 
 class BnB4BitBasicTests(Base4bitTests):
     def setUp(self):
+        gc.collect()
+        torch.cuda.empty_cache()
+
         # Models
         self.model_fp16 = SD3Transformer2DModel.from_pretrained(
             self.model_name, subfolder="transformer", torch_dtype=torch.float16
@@ -232,7 +237,7 @@ def test_linear_are_4bit(self):
 
     def test_config_from_pretrained(self):
         transformer_4bit = FluxTransformer2DModel.from_pretrained(
-            "sayakpaul/flux.1-dev-nf4-pkg", subfolder="transformer"
+            "hf-internal-testing/flux.1-dev-nf4-pkg", subfolder="transformer"
         )
         linear = get_some_linear_layer(transformer_4bit)
         self.assertTrue(linear.weight.__class__ == bnb.nn.Params4bit)
@@ -312,9 +317,42 @@ def test_bnb_4bit_wrong_config(self):
         with self.assertRaises(ValueError):
             _ = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_quant_storage="add")
 
+    def test_bnb_4bit_errors_loading_incorrect_state_dict(self):
+        r"""
+        Test if loading with an incorrect state dict raises an error.
+        """
+        with tempfile.TemporaryDirectory() as tmpdirname:
+            nf4_config = BitsAndBytesConfig(load_in_4bit=True)
+            model_4bit = SD3Transformer2DModel.from_pretrained(
+                self.model_name, subfolder="transformer", quantization_config=nf4_config
+            )
+            model_4bit.save_pretrained(tmpdirname)
+            del model_4bit
+
+            with self.assertRaises(ValueError) as err_context:
+                state_dict = safetensors.torch.load_file(
+                    os.path.join(tmpdirname, "diffusion_pytorch_model.safetensors")
+                )
+
+                # corrupt the state dict
+                key_to_target = "context_embedder.weight"  # can be other keys too.
+                compatible_param = state_dict[key_to_target]
+                corrupted_param = torch.randn(compatible_param.shape[0] - 1, 1)
+                state_dict[key_to_target] = bnb.nn.Params4bit(corrupted_param, requires_grad=False)
+                safetensors.torch.save_file(
+                    state_dict, os.path.join(tmpdirname, "diffusion_pytorch_model.safetensors")
+                )
+
+                _ = SD3Transformer2DModel.from_pretrained(tmpdirname)
+
+            assert key_to_target in str(err_context.exception)
+
 
 class BnB4BitTrainingTests(Base4bitTests):
     def setUp(self):
+        gc.collect()
+        torch.cuda.empty_cache()
+
         nf4_config = BitsAndBytesConfig(
             load_in_4bit=True,
             bnb_4bit_quant_type="nf4",
@@ -360,6 +398,9 @@ def test_training(self):
 @require_transformers_version_greater("4.44.0")
 class SlowBnb4BitTests(Base4bitTests):
     def setUp(self) -> None:
+        gc.collect()
+        torch.cuda.empty_cache()
+
         nf4_config = BitsAndBytesConfig(
             load_in_4bit=True,
             bnb_4bit_quant_type="nf4",
@@ -447,8 +488,10 @@ def test_moving_to_cpu_throws_warning(self):
 @require_transformers_version_greater("4.44.0")
 class SlowBnb4BitFluxTests(Base4bitTests):
     def setUp(self) -> None:
-        # TODO: Copy sayakpaul/flux.1-dev-nf4-pkg to testing repo.
-        model_id = "sayakpaul/flux.1-dev-nf4-pkg"
+        gc.collect()
+        torch.cuda.empty_cache()
+
+        model_id = "hf-internal-testing/flux.1-dev-nf4-pkg"
         t5_4bit = T5EncoderModel.from_pretrained(model_id, subfolder="text_encoder_2")
         transformer_4bit = FluxTransformer2DModel.from_pretrained(model_id, subfolder="transformer")
         self.pipeline_4bit = DiffusionPipeline.from_pretrained(

tests/quantization/bnb/test_mixed_int8.py

@@ -117,6 +117,9 @@ def get_dummy_inputs(self):
 
 class BnB8bitBasicTests(Base8bitTests):
     def setUp(self):
+        gc.collect()
+        torch.cuda.empty_cache()
+
         # Models
         self.model_fp16 = SD3Transformer2DModel.from_pretrained(
             self.model_name, subfolder="transformer", torch_dtype=torch.float16
@@ -238,7 +241,7 @@ def test_llm_skip(self):
 
     def test_config_from_pretrained(self):
         transformer_8bit = FluxTransformer2DModel.from_pretrained(
-            "sayakpaul/flux.1-dev-int8-pkg", subfolder="transformer"
+            "hf-internal-testing/flux.1-dev-int8-pkg", subfolder="transformer"
         )
         linear = get_some_linear_layer(transformer_8bit)
         self.assertTrue(linear.weight.__class__ == bnb.nn.Int8Params)
@@ -296,6 +299,9 @@ def test_device_and_dtype_assignment(self):
 
 class BnB8bitTrainingTests(Base8bitTests):
     def setUp(self):
+        gc.collect()
+        torch.cuda.empty_cache()
+
         mixed_int8_config = BitsAndBytesConfig(load_in_8bit=True)
         self.model_8bit = SD3Transformer2DModel.from_pretrained(
             self.model_name, subfolder="transformer", quantization_config=mixed_int8_config
@@ -337,6 +343,9 @@ def test_training(self):
 @require_transformers_version_greater("4.44.0")
 class SlowBnb8bitTests(Base8bitTests):
     def setUp(self) -> None:
+        gc.collect()
+        torch.cuda.empty_cache()
+
         mixed_int8_config = BitsAndBytesConfig(load_in_8bit=True)
         model_8bit = SD3Transformer2DModel.from_pretrained(
             self.model_name, subfolder="transformer", quantization_config=mixed_int8_config
@@ -427,8 +436,10 @@ def test_generate_quality_dequantize(self):
 @require_transformers_version_greater("4.44.0")
 class SlowBnb8bitFluxTests(Base8bitTests):
     def setUp(self) -> None:
-        # TODO: Copy sayakpaul/flux.1-dev-int8-pkg to testing repo.
-        model_id = "sayakpaul/flux.1-dev-int8-pkg"
+        gc.collect()
+        torch.cuda.empty_cache()
+
+        model_id = "hf-internal-testing/flux.1-dev-int8-pkg"
         t5_8bit = T5EncoderModel.from_pretrained(model_id, subfolder="text_encoder_2")
         transformer_8bit = FluxTransformer2DModel.from_pretrained(model_id, subfolder="transformer")
         self.pipeline_8bit = DiffusionPipeline.from_pretrained(
@@ -466,6 +477,9 @@ def test_quality(self):
 @slow
 class BaseBnb8bitSerializationTests(Base8bitTests):
     def setUp(self):
+        gc.collect()
+        torch.cuda.empty_cache()
+
         quantization_config = BitsAndBytesConfig(
             load_in_8bit=True,
         )