[pull] master from comfyanonymous:master

comfy/model_management.py

@@ -1098,13 +1098,14 @@ def cast_to_device(tensor, device, dtype, copy=False):
             MAX_PINNED_MEMORY = get_total_memory(torch.device("cpu")) * 0.95
         logging.info("Enabled pinned memory {}".format(MAX_PINNED_MEMORY // (1024 * 1024)))
 
+PINNING_ALLOWED_TYPES = set(["Parameter", "QuantizedTensor"])
 
 def pin_memory(tensor):
     global TOTAL_PINNED_MEMORY
     if MAX_PINNED_MEMORY <= 0:
         return False
 
-    if type(tensor) is not torch.nn.parameter.Parameter:
+    if type(tensor).__name__ not in PINNING_ALLOWED_TYPES:
         return False
 
     if not is_device_cpu(tensor.device):
@@ -1124,6 +1125,9 @@ def pin_memory(tensor):
         return False
 
     ptr = tensor.data_ptr()
+    if ptr == 0:
+        return False
+
     if torch.cuda.cudart().cudaHostRegister(ptr, size, 1) == 0:
         PINNED_MEMORY[ptr] = size
         TOTAL_PINNED_MEMORY += size

comfy/model_patcher.py

@@ -231,7 +231,6 @@ def __init__(self, model, load_device, offload_device, size=0, weight_inplace_up
         self.object_patches_backup = {}
         self.weight_wrapper_patches = {}
         self.model_options = {"transformer_options":{}}
-        self.model_size()
         self.load_device = load_device
         self.offload_device = offload_device
         self.weight_inplace_update = weight_inplace_update
@@ -286,7 +285,7 @@ def lowvram_patch_counter(self):
         return self.model.lowvram_patch_counter
 
     def clone(self):
-        n = self.__class__(self.model, self.load_device, self.offload_device, self.size, weight_inplace_update=self.weight_inplace_update)
+        n = self.__class__(self.model, self.load_device, self.offload_device, self.model_size(), weight_inplace_update=self.weight_inplace_update)
         n.patches = {}
         for k in self.patches:
             n.patches[k] = self.patches[k][:]

comfy/ops.py

@@ -540,115 +540,118 @@ def forward(self, *args, **kwargs):
 # ==============================================================================
 from .quant_ops import QuantizedTensor, QUANT_ALGOS
 
-class MixedPrecisionOps(disable_weight_init):
-    _layer_quant_config = {}
-    _compute_dtype = torch.bfloat16
-
-    class Linear(torch.nn.Module, CastWeightBiasOp):
-        def __init__(
-            self,
-            in_features: int,
-            out_features: int,
-            bias: bool = True,
-            device=None,
-            dtype=None,
-        ) -> None:
-            super().__init__()
-
-            self.factory_kwargs = {"device": device, "dtype": MixedPrecisionOps._compute_dtype}
-            # self.factory_kwargs = {"device": device, "dtype": dtype}
-
-            self.in_features = in_features
-            self.out_features = out_features
-            if bias:
-                self.bias = torch.nn.Parameter(torch.empty(out_features, **self.factory_kwargs))
-            else:
-                self.register_parameter("bias", None)
 
-            self.tensor_class = None
+def mixed_precision_ops(layer_quant_config={}, compute_dtype=torch.bfloat16, full_precision_mm=False):
+    class MixedPrecisionOps(manual_cast):
+        _layer_quant_config = layer_quant_config
+        _compute_dtype = compute_dtype
+        _full_precision_mm = full_precision_mm
+
+        class Linear(torch.nn.Module, CastWeightBiasOp):
+            def __init__(
+                self,
+                in_features: int,
+                out_features: int,
+                bias: bool = True,
+                device=None,
+                dtype=None,
+            ) -> None:
+                super().__init__()
+
+                self.factory_kwargs = {"device": device, "dtype": MixedPrecisionOps._compute_dtype}
+                # self.factory_kwargs = {"device": device, "dtype": dtype}
+
+                self.in_features = in_features
+                self.out_features = out_features
+                if bias:
+                    self.bias = torch.nn.Parameter(torch.empty(out_features, **self.factory_kwargs))
+                else:
+                    self.register_parameter("bias", None)
 
-        def reset_parameters(self):
-            return None
+                self.tensor_class = None
+                self._full_precision_mm = MixedPrecisionOps._full_precision_mm
 
-        def _load_from_state_dict(self, state_dict, prefix, local_metadata,
-                                  strict, missing_keys, unexpected_keys, error_msgs):
+            def reset_parameters(self):
+                return None
 
-            device = self.factory_kwargs["device"]
-            layer_name = prefix.rstrip('.')
-            weight_key = f"{prefix}weight"
-            weight = state_dict.pop(weight_key, None)
-            if weight is None:
-                raise ValueError(f"Missing weight for layer {layer_name}")
+            def _load_from_state_dict(self, state_dict, prefix, local_metadata,
+                                    strict, missing_keys, unexpected_keys, error_msgs):
 
-            manually_loaded_keys = [weight_key]
+                device = self.factory_kwargs["device"]
+                layer_name = prefix.rstrip('.')
+                weight_key = f"{prefix}weight"
+                weight = state_dict.pop(weight_key, None)
+                if weight is None:
+                    raise ValueError(f"Missing weight for layer {layer_name}")
 
-            if layer_name not in MixedPrecisionOps._layer_quant_config:
-                self.weight = torch.nn.Parameter(weight.to(device=device, dtype=MixedPrecisionOps._compute_dtype), requires_grad=False)
-            else:
-                quant_format = MixedPrecisionOps._layer_quant_config[layer_name].get("format", None)
-                if quant_format is None:
-                    raise ValueError(f"Unknown quantization format for layer {layer_name}")
-
-                qconfig = QUANT_ALGOS[quant_format]
-                self.layout_type = qconfig["comfy_tensor_layout"]
-
-                weight_scale_key = f"{prefix}weight_scale"
-                layout_params = {
-                    'scale': state_dict.pop(weight_scale_key, None),
-                    'orig_dtype': MixedPrecisionOps._compute_dtype,
-                    'block_size': qconfig.get("group_size", None),
-                }
-                if layout_params['scale'] is not None:
-                    manually_loaded_keys.append(weight_scale_key)
-
-                self.weight = torch.nn.Parameter(
-                    QuantizedTensor(weight.to(device=device), self.layout_type, layout_params),
-                    requires_grad=False
-                )
-
-                for param_name in qconfig["parameters"]:
-                    param_key = f"{prefix}{param_name}"
-                    _v = state_dict.pop(param_key, None)
-                    if _v is None:
-                        continue
-                    setattr(self, param_name, torch.nn.Parameter(_v.to(device=device), requires_grad=False))
-                    manually_loaded_keys.append(param_key)
-
-            super()._load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)
-
-            for key in manually_loaded_keys:
-                if key in missing_keys:
-                    missing_keys.remove(key)
-
-        def _forward(self, input, weight, bias):
-            return torch.nn.functional.linear(input, weight, bias)
+                manually_loaded_keys = [weight_key]
 
-        def forward_comfy_cast_weights(self, input):
-            weight, bias, offload_stream = cast_bias_weight(self, input, offloadable=True)
-            x = self._forward(input, weight, bias)
-            uncast_bias_weight(self, weight, bias, offload_stream)
-            return x
+                if layer_name not in MixedPrecisionOps._layer_quant_config:
+                    self.weight = torch.nn.Parameter(weight.to(device=device, dtype=MixedPrecisionOps._compute_dtype), requires_grad=False)
+                else:
+                    quant_format = MixedPrecisionOps._layer_quant_config[layer_name].get("format", None)
+                    if quant_format is None:
+                        raise ValueError(f"Unknown quantization format for layer {layer_name}")
+
+                    qconfig = QUANT_ALGOS[quant_format]
+                    self.layout_type = qconfig["comfy_tensor_layout"]
+
+                    weight_scale_key = f"{prefix}weight_scale"
+                    layout_params = {
+                        'scale': state_dict.pop(weight_scale_key, None),
+                        'orig_dtype': MixedPrecisionOps._compute_dtype,
+                        'block_size': qconfig.get("group_size", None),
+                    }
+                    if layout_params['scale'] is not None:
+                        manually_loaded_keys.append(weight_scale_key)
+
+                    self.weight = torch.nn.Parameter(
+                        QuantizedTensor(weight.to(device=device), self.layout_type, layout_params),
+                        requires_grad=False
+                    )
+
+                    for param_name in qconfig["parameters"]:
+                        param_key = f"{prefix}{param_name}"
+                        _v = state_dict.pop(param_key, None)
+                        if _v is None:
+                            continue
+                        setattr(self, param_name, torch.nn.Parameter(_v.to(device=device), requires_grad=False))
+                        manually_loaded_keys.append(param_key)
+
+                super()._load_from_state_dict(state_dict, prefix, local_metadata, strict, missing_keys, unexpected_keys, error_msgs)
+
+                for key in manually_loaded_keys:
+                    if key in missing_keys:
+                        missing_keys.remove(key)
+
+            def _forward(self, input, weight, bias):
+                return torch.nn.functional.linear(input, weight, bias)
 
-        def forward(self, input, *args, **kwargs):
-            run_every_op()
+            def forward_comfy_cast_weights(self, input):
+                weight, bias, offload_stream = cast_bias_weight(self, input, offloadable=True)
+                x = self._forward(input, weight, bias)
+                uncast_bias_weight(self, weight, bias, offload_stream)
+                return x
 
-            if self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
-                return self.forward_comfy_cast_weights(input, *args, **kwargs)
-            if (getattr(self, 'layout_type', None) is not None and
-                getattr(self, 'input_scale', None) is not None and
-                not isinstance(input, QuantizedTensor)):
-                input = QuantizedTensor.from_float(input, self.layout_type, scale=self.input_scale, dtype=self.weight.dtype)
-            return self._forward(input, self.weight, self.bias)
+            def forward(self, input, *args, **kwargs):
+                run_every_op()
 
+                if self._full_precision_mm or self.comfy_cast_weights or len(self.weight_function) > 0 or len(self.bias_function) > 0:
+                    return self.forward_comfy_cast_weights(input, *args, **kwargs)
+                if (getattr(self, 'layout_type', None) is not None and
+                    getattr(self, 'input_scale', None) is not None and
+                    not isinstance(input, QuantizedTensor)):
+                    input = QuantizedTensor.from_float(input, self.layout_type, scale=self.input_scale, dtype=self.weight.dtype)
+                return self._forward(input, self.weight, self.bias)
+    return MixedPrecisionOps
 
 def pick_operations(weight_dtype, compute_dtype, load_device=None, disable_fast_fp8=False, fp8_optimizations=False, scaled_fp8=None, model_config=None):
+    fp8_compute = comfy.model_management.supports_fp8_compute(load_device) # TODO: if we support more ops this needs to be more granular
+
     if model_config and hasattr(model_config, 'layer_quant_config') and model_config.layer_quant_config:
-        MixedPrecisionOps._layer_quant_config = model_config.layer_quant_config
-        MixedPrecisionOps._compute_dtype = compute_dtype
         logging.info(f"Using mixed precision operations: {len(model_config.layer_quant_config)} quantized layers")
-        return MixedPrecisionOps
+        return mixed_precision_ops(model_config.layer_quant_config, compute_dtype, full_precision_mm=not fp8_compute)
 
-    fp8_compute = comfy.model_management.supports_fp8_compute(load_device)
     if scaled_fp8 is not None:
         return scaled_fp8_ops(fp8_matrix_mult=fp8_compute and fp8_optimizations, scale_input=fp8_optimizations, override_dtype=scaled_fp8)
 

comfy/quant_ops.py

@@ -228,6 +228,14 @@ def dequant_arg(arg):
         new_kwargs = dequant_arg(kwargs)
         return func(*new_args, **new_kwargs)
 
+    def data_ptr(self):
+        return self._qdata.data_ptr()
+
+    def is_pinned(self):
+        return self._qdata.is_pinned()
+
+    def is_contiguous(self):
+        return self._qdata.is_contiguous()
 
 # ==============================================================================
 # Generic Utilities (Layout-Agnostic Operations)
@@ -338,6 +346,18 @@ def generic_copy_(func, args, kwargs):
     return func(*args, **kwargs)
 
 
+@register_generic_util(torch.ops.aten.to.dtype)
+def generic_to_dtype(func, args, kwargs):
+    """Handle .to(dtype) calls - dtype conversion only."""
+    src = args[0]
+    if isinstance(src, QuantizedTensor):
+        # For dtype-only conversion, just change the orig_dtype, no real cast is needed
+        target_dtype = args[1] if len(args) > 1 else kwargs.get('dtype')
+        src._layout_params["orig_dtype"] = target_dtype
+        return src
+    return func(*args, **kwargs)
+
+
 @register_generic_util(torch.ops.aten._has_compatible_shallow_copy_type.default)
 def generic_has_compatible_shallow_copy_type(func, args, kwargs):
     return True
@@ -385,8 +405,8 @@ def quantize(cls, tensor, scale=None, dtype=torch.float8_e4m3fn):
 
         tensor_scaled = tensor * (1.0 / scale).to(tensor.dtype)
         # TODO: uncomment this if it's actually needed because the clamp has a small performance penality'
-        # lp_amax = torch.finfo(dtype).max
-        # torch.clamp(tensor_scaled, min=-lp_amax, max=lp_amax, out=tensor_scaled)
+        lp_amax = torch.finfo(dtype).max
+        torch.clamp(tensor_scaled, min=-lp_amax, max=lp_amax, out=tensor_scaled)
         qdata = tensor_scaled.to(dtype, memory_format=torch.contiguous_format)
 
         layout_params = {

comfy/sd.py

@@ -917,7 +917,12 @@ class CLIPType(Enum):
 def load_clip(ckpt_paths, embedding_directory=None, clip_type=CLIPType.STABLE_DIFFUSION, model_options={}):
     clip_data = []
     for p in ckpt_paths:
-        clip_data.append(comfy.utils.load_torch_file(p, safe_load=True))
+        sd, metadata = comfy.utils.load_torch_file(p, safe_load=True, return_metadata=True)
+        if metadata is not None:
+            quant_metadata = metadata.get("_quantization_metadata", None)
+            if quant_metadata is not None:
+                sd["_quantization_metadata"] = quant_metadata
+        clip_data.append(sd)
     return load_text_encoder_state_dicts(clip_data, embedding_directory=embedding_directory, clip_type=clip_type, model_options=model_options)
 
 
@@ -1142,6 +1147,8 @@ class EmptyClass:
 
     parameters = 0
     for c in clip_data:
+        if "_quantization_metadata" in c:
+            c.pop("_quantization_metadata")
         parameters += comfy.utils.calculate_parameters(c)
         tokenizer_data, model_options = comfy.text_encoders.long_clipl.model_options_long_clip(c, tokenizer_data, model_options)
 

comfy/sd1_clip.py

@@ -109,13 +109,23 @@ def __init__(self, device="cpu", max_length=77,
 
         operations = model_options.get("custom_operations", None)
         scaled_fp8 = None
+        quantization_metadata = model_options.get("quantization_metadata", None)
 
         if operations is None:
-            scaled_fp8 = model_options.get("scaled_fp8", None)
-            if scaled_fp8 is not None:
-                operations = comfy.ops.scaled_fp8_ops(fp8_matrix_mult=False, override_dtype=scaled_fp8)
+            layer_quant_config = None
+            if quantization_metadata is not None:
+                layer_quant_config = json.loads(quantization_metadata).get("layers", None)
+
+            if layer_quant_config is not None:
+                operations = comfy.ops.mixed_precision_ops(layer_quant_config, dtype, full_precision_mm=True)
+                logging.info(f"Using MixedPrecisionOps for text encoder: {len(layer_quant_config)} quantized layers")
             else:
-                operations = comfy.ops.manual_cast
+                # Fallback to scaled_fp8_ops for backward compatibility
+                scaled_fp8 = model_options.get("scaled_fp8", None)
+                if scaled_fp8 is not None:
+                    operations = comfy.ops.scaled_fp8_ops(fp8_matrix_mult=False, override_dtype=scaled_fp8)
+                else:
+                    operations = comfy.ops.manual_cast
 
         self.operations = operations
         self.transformer = model_class(config, dtype, device, self.operations)

comfy/text_encoders/hunyuan_video.py

@@ -18,6 +18,9 @@ def llama_detect(state_dict, prefix=""):
     if scaled_fp8_key in state_dict:
         out["llama_scaled_fp8"] = state_dict[scaled_fp8_key].dtype
 
+    if "_quantization_metadata" in state_dict:
+        out["llama_quantization_metadata"] = state_dict["_quantization_metadata"]
+
     return out