[Misc] update fp8 to use vLLMParameter (#7437)

tests/weight_loading/models.txt

@@ -15,3 +15,4 @@ compressed-tensors, nm-testing/Phi-3-mini-128k-instruct-FP8, main
 compressed-tensors, neuralmagic/Phi-3-medium-128k-instruct-quantized.w4a16, main
 awq, casperhansen/mixtral-instruct-awq, main
 awq_marlin, casperhansen/mixtral-instruct-awq, main
+fp8, neuralmagic/Meta-Llama-3-8B-Instruct-FP8-KV, main

vllm/model_executor/layers/linear.py

@@ -22,7 +22,7 @@ logger = init_logger(__name__)
 
 WEIGHT_LOADER_V2_SUPPORTED = [
     "CompressedTensorsLinearMethod", "AWQMarlinLinearMethod",
-    "AWQLinearMethod", "GPTQMarlinLinearMethod"
+    "AWQLinearMethod", "GPTQMarlinLinearMethod", "Fp8LinearMethod"
 ]
 
 
@@ -349,6 +349,11 @@ class ColumnParallelLinear(LinearBase):
         param_data.copy_(loaded_weight)
 
     def weight_loader_v2(self, param: Parameter, loaded_weight: torch.Tensor):
+        # Special case for loading scales off disk, which often do not
+        # have a shape (such as in the case of AutoFP8).
+        if len(loaded_weight.shape) == 0:
+            assert loaded_weight.numel() == 1
+            loaded_weight = loaded_weight.reshape(1)
         param.load_column_parallel_weight(loaded_weight=loaded_weight)
 
     def forward(self, input_):
@@ -1021,6 +1026,13 @@ class RowParallelLinear(LinearBase):
 
     def weight_loader_v2(self, param: BasevLLMParameter,
                          loaded_weight: torch.Tensor):
+
+        # Special case for loading scales off disk, which often do not
+        # have a shape (such as in the case of AutoFP8).
+        if len(loaded_weight.shape) == 0:
+            assert loaded_weight.numel() == 1
+            loaded_weight = loaded_weight.reshape(1)
+
         param.load_row_parallel_weight(loaded_weight=loaded_weight)
 
     def forward(self, input_):

vllm/model_executor/layers/quantization/fp8.py

@@ -19,9 +19,10 @@ from vllm.model_executor.layers.quantization.utils.quant_utils import (
     is_layer_skipped)
 from vllm.model_executor.layers.quantization.utils.w8a8_utils import (
     all_close_1d, apply_fp8_linear, convert_to_channelwise,
-    create_per_tensor_scale_param, cutlass_fp8_supported,
+    cutlass_fp8_supported, normalize_e4m3fn_to_e4m3fnuz, per_tensor_dequantize,
-    normalize_e4m3fn_to_e4m3fnuz, per_tensor_dequantize,
     requantize_with_max_scale)
+from vllm.model_executor.parameter import (ModelWeightParameter,
+                                           PerTensorScaleParameter)
 from vllm.model_executor.utils import set_weight_attrs
 from vllm.platforms import current_platform
 from vllm.utils import is_hip, print_warning_once
@@ -137,6 +138,7 @@ class Fp8LinearMethod(LinearMethodBase):
     ):
         del input_size, output_size
         output_size_per_partition = sum(output_partition_sizes)
+        weight_loader = extra_weight_attrs.get("weight_loader")
 
         layer.logical_widths = output_partition_sizes
 
@@ -148,34 +150,41 @@ class Fp8LinearMethod(LinearMethodBase):
         weight_dtype = (torch.float8_e4m3fn
                         if self.quant_config.is_checkpoint_fp8_serialized else
                         params_dtype)
-        weight = Parameter(torch.empty(output_size_per_partition,
+
+        weight = ModelWeightParameter(data=torch.empty(
+            output_size_per_partition,
             input_size_per_partition,
             dtype=weight_dtype),
-                           requires_grad=False)
+                                      input_dim=1,
+                                      output_dim=0,
+                                      weight_loader=weight_loader)
         layer.register_parameter("weight", weight)
-        set_weight_attrs(weight, {
-            **extra_weight_attrs,
-            "input_dim": 1,
-            "output_dim": 0,
-        })
 
         # If checkpoint is serialized fp8, load them.
         # Otherwise, wait until process_weights_after_loading.
         if self.quant_config.is_checkpoint_fp8_serialized:
             # WEIGHT SCALE
-            scale = create_per_tensor_scale_param(output_partition_sizes,
+            scale = PerTensorScaleParameter(data=torch.empty(
-                                                  **extra_weight_attrs)
+                len(output_partition_sizes), dtype=torch.float32),
+                                            weight_loader=weight_loader)
+
+            scale[:] = torch.finfo(torch.float32).min
             layer.register_parameter("weight_scale", scale)
 
             # INPUT ACTIVATION SCALE
             if self.quant_config.activation_scheme == "static":
-                scale = create_per_tensor_scale_param(output_partition_sizes,
+                scale = PerTensorScaleParameter(data=torch.empty(
-                                                      **extra_weight_attrs)
+                    len(output_partition_sizes), dtype=torch.float32),
+                                                weight_loader=weight_loader)
+
+                scale[:] = torch.finfo(torch.float32).min
                 layer.register_parameter("input_scale", scale)
             else:
                 layer.register_parameter("input_scale", None)
 
     def process_weights_after_loading(self, layer: Module) -> None:
+        layer.weight = torch.nn.Parameter(layer.weight.data,
+                                          requires_grad=False)
         # If checkpoint not serialized fp8, quantize the weights.
         if not self.quant_config.is_checkpoint_fp8_serialized:
             qweight, weight_scale = ops.scaled_fp8_quant(layer.weight,
@@ -197,6 +206,11 @@ class Fp8LinearMethod(LinearMethodBase):
         # If checkpoint is fp8, handle that there are N scales for N
         # shards in a fused module
         else:
+            layer.weight_scale = torch.nn.Parameter(layer.weight_scale.data,
+                                                    requires_grad=False)
+            if self.quant_config.activation_scheme == "static":
+                layer.input_scale = torch.nn.Parameter(layer.input_scale.data,
+                                                       requires_grad=False)
             # If using marlin (w8a16), kernel uses channelwise weights,
             # so extend the weight scales to be channelwise.
             if self.use_marlin:

vllm/model_executor/parameter.py

@@ -208,10 +208,17 @@ class PerTensorScaleParameter(BasevLLMParameter):
         if isinstance(shard_id, int):
             return shard_id
 
+        # if not int, assume shard_id for qkv
+        # map to int and return
         assert isinstance(shard_id, str)
         assert shard_id in self.qkv_idxs
         return self.qkv_idxs[shard_id]
 
+    # For row parallel layers, no sharding needed
+    # load weight into parameter as is
+    def load_row_parallel_weight(self, *args, **kwargs):
+        super().load_row_parallel_weight(*args, **kwargs)
+
     def load_merged_column_weight(self, *args, **kwargs):
         self._load_into_shard_id(*args, **kwargs)
 
@@ -219,7 +226,7 @@ class PerTensorScaleParameter(BasevLLMParameter):
         self._load_into_shard_id(*args, **kwargs)
 
     def load_column_parallel_weight(self, *args, **kwargs):
-        self._load_into_shard_id(*args, **kwargs)
+        super().load_row_parallel_weight(*args, **kwargs)
 
     def _load_into_shard_id(self, loaded_weight: torch.Tensor,
                             shard_id: Union[str, int], **kwargs):