[V1][Quantization] Add CUDA graph compatible v1 GGUF support (#18646)

Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Signed-off-by: Isotr0py <2037008807@qq.com>

tests/kernels/quantization/test_gguf.py

@@ -8,7 +8,6 @@ from gguf import GGMLQuantizationType, GGUFReader, ReaderTensor, dequantize
 from huggingface_hub import snapshot_download
 
 import vllm._custom_ops as ops
-from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.fused_moe import fused_experts
 from vllm.model_executor.layers.quantization.gguf import _fused_moe_gguf
 from vllm.platforms import current_platform
@@ -176,12 +175,11 @@ def test_moe(num_tokens: int, hidden_size: int, dtype: torch.dtype,
 
     w2_dequant = torch.tensor(dequantize(w2.data, quant_type),
                               device="cuda").to(dtype)
-    act = SiluAndMul()
 
     output = _fused_moe_gguf(x, torch.tensor(w13.data, device="cuda"),
                              torch.tensor(w2.data,
                                           device="cuda"), topk_weights,
-                             topk_ids, quant_type, quant_type, act)
+                             topk_ids, quant_type, quant_type, "silu")
 
     ref_output = fused_experts(x, w13_dequant, w2_dequant, topk_weights,
                                topk_ids).reshape(output.shape)

tests/models/quantization/test_gguf.py

@@ -78,8 +78,12 @@ DOLPHIN_CONFIG = GGUFTestConfig(
 )
 
 MODELS = [
-    LLAMA_CONFIG, QWEN2_CONFIG, PHI3_CONFIG, GPT2_CONFIG, STABLELM_CONFIG,
-    DOLPHIN_CONFIG
+    LLAMA_CONFIG,
+    QWEN2_CONFIG,
+    PHI3_CONFIG,
+    GPT2_CONFIG,
+    # STABLELM_CONFIG,  # enable this when v1 support head_size=80
+    DOLPHIN_CONFIG,
     # STARCODER_CONFIG, # broken
 ]
 

vllm/engine/arg_utils.py

@@ -1291,14 +1291,6 @@ class EngineArgs:
                                recommend_to_remove=False)
             return False
 
-        # Some quantization is not compatible with torch.compile.
-        V1_UNSUPPORTED_QUANT = ["gguf"]
-        if model_config.quantization in V1_UNSUPPORTED_QUANT:
-            _raise_or_fallback(
-                feature_name=f"--quantization {model_config.quantization}",
-                recommend_to_remove=False)
-            return False
-
         # No Embedding Models so far.
         if model_config.task not in ["generate"]:
             _raise_or_fallback(feature_name=f"--task {model_config.task}",

vllm/model_executor/layers/linear.py

@@ -587,8 +587,6 @@ class MergedColumnParallelLinear(ColumnParallelLinear):
                 param.shard_id.append(loaded_shard_id)
                 param.shard_id_map[loaded_shard_id] = len(param.data_container)
                 param.data_container.append(loaded_weight)
-                if len(param.data_container) == 2:
-                    self.qweight = param.materialize_nested()
                 return
 
         param_data = param.data
@@ -982,8 +980,6 @@ class QKVParallelLinear(ColumnParallelLinear):
                 param.shard_id.append(loaded_shard_id)
                 param.shard_id_map[loaded_shard_id] = len(param.data_container)
                 param.data_container.append(loaded_weight)
-                if len(param.data_container) == 3:
-                    self.qweight = param.materialize_nested()
                 return
 
         param_data = param.data

vllm/model_executor/layers/quantization/gguf.py

@@ -9,7 +9,6 @@ from torch.nn.parameter import Parameter, UninitializedParameter
 
 from vllm import _custom_ops as ops
 from vllm.logger import init_logger
-from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.fused_moe.layer import (FusedMoE,
                                                         FusedMoEMethodBase)
 from vllm.model_executor.layers.linear import LinearBase, LinearMethodBase
@@ -19,6 +18,7 @@ from vllm.model_executor.layers.quantization.base_config import (
 from vllm.model_executor.layers.vocab_parallel_embedding import (
     VocabParallelEmbedding)
 from vllm.model_executor.utils import set_weight_attrs
+from vllm.utils import direct_register_custom_op
 
 logger = init_logger(__name__)
 
@@ -96,8 +96,8 @@ MMVQ_QUANT_TYPES = STANDARD_QUANT_TYPES | KQUANT_TYPES | IMATRIX_QUANT_TYPES
 MMQ_QUANT_TYPES = STANDARD_QUANT_TYPES | KQUANT_TYPES
 
 
-def _fuse_mul_mat(x: torch.Tensor, qweight: torch.Tensor,
-                  qweight_type: int) -> torch.Tensor:
+def _fused_mul_mat_gguf(x: torch.Tensor, qweight: torch.Tensor,
+                        qweight_type: int) -> torch.Tensor:
     # HACK: when doing chunked prefill we don't generate output tokens
     # so input to logits generator is empty which causes invalid parameter
     if x.shape[0] == 0:
@@ -130,6 +130,30 @@ def _fuse_mul_mat(x: torch.Tensor, qweight: torch.Tensor,
     return y
 
 
+def _fused_mul_mat_gguf_fake(
+    x: torch.Tensor,
+    qweight: torch.Tensor,
+    qweight_type: int,
+) -> torch.Tensor:
+    return torch.empty(x.shape[0],
+                       qweight.shape[0],
+                       dtype=x.dtype,
+                       device=x.device)
+
+
+try:
+    direct_register_custom_op(
+        op_name="_fused_mul_mat_gguf",
+        op_func=_fused_mul_mat_gguf,
+        mutates_args=[],
+        fake_impl=_fused_mul_mat_gguf_fake,
+    )
+    fused_mul_mat_gguf = torch.ops.vllm._fused_mul_mat_gguf
+
+except AttributeError as error:
+    raise error
+
+
 def _fused_moe_gguf(
     x: torch.Tensor,
     w1: torch.Tensor,
@@ -138,8 +162,21 @@ def _fused_moe_gguf(
     topk_ids: torch.Tensor,
     qweight_type: int,
     qweight_type2: int,
-    act,
+    activation: str,
 ) -> torch.Tensor:
+
+    def act(x: torch.Tensor):
+        d = x.shape[-1] // 2
+        output_shape = (x.shape[:-1] + (d, ))
+        out = torch.empty(output_shape, dtype=x.dtype, device=x.device)
+        if activation == "silu":
+            torch.ops._C.silu_and_mul(out, x)
+        elif activation == "gelu":
+            torch.ops._C.gelu_and_mul(out, x)
+        else:
+            raise ValueError(f"Unsupported activation: {activation}")
+        return out
+
     # lazy import to avoid triggering triton import in CPU backend
     from vllm.model_executor.layers.fused_moe.fused_moe import (
         moe_align_block_size)
@@ -189,12 +226,12 @@ def _fused_moe_gguf(
             for ww, ii in zip(w, idx):
                 expert_up = w1[ii]
 
-                out = _fuse_mul_mat(inp, expert_up, qweight_type)
+                out = fused_mul_mat_gguf(inp, expert_up, qweight_type)
                 out = act(out)
 
                 expert_down = w2[ii]
-                current_state = _fuse_mul_mat(out, expert_down,
-                                              qweight_type2).mul_(ww)
+                current_state = fused_mul_mat_gguf(out, expert_down,
+                                                   qweight_type2).mul_(ww)
                 if current_hidden_state is None:
                     current_hidden_state = current_state
                 else:
@@ -203,6 +240,78 @@ def _fused_moe_gguf(
     return out_hidden_states
 
 
+def _fused_moe_gguf_fake(
+    x: torch.Tensor,
+    w1: torch.Tensor,
+    w2: torch.Tensor,
+    topk_weights: torch.Tensor,
+    topk_ids: torch.Tensor,
+    qweight_type: int,
+    qweight_type2: int,
+    activation: str,
+) -> torch.Tensor:
+    return torch.empty_like(x)
+
+
+try:
+    direct_register_custom_op(
+        op_name="_fused_moe_gguf",
+        op_func=_fused_moe_gguf,
+        mutates_args=[],
+        fake_impl=_fused_moe_gguf_fake,
+    )
+    fused_moe_gguf = torch.ops.vllm._fused_moe_gguf
+
+except AttributeError as error:
+    raise error
+
+
+def _apply_gguf_embedding(
+    x: torch.Tensor,
+    qweight: torch.Tensor,
+    qweight_type: int,
+    hidden_size: int,
+    dtype: Optional[torch.dtype] = None,
+) -> torch.Tensor:
+    if qweight_type in UNQUANTIZED_TYPES:
+        return torch.embedding(qweight, x)
+    elif qweight_type in DEQUANT_TYPES:
+        block_size, type_size = gguf.GGML_QUANT_SIZES[qweight_type]
+        x_flat = x.flatten()
+        assert (hidden_size == qweight.shape[1] // type_size * block_size)
+        quant = torch.index_select(qweight, dim=0, index=x_flat)
+        dequant = ops.ggml_dequantize(quant, qweight_type, hidden_size,
+                                      x_flat.shape[0], dtype)
+        return dequant.view(*x.shape, hidden_size)
+    else:
+        qweight_type = WeightType(qweight_type)
+        raise NotImplementedError(
+            f"Unsupported GGUF quantization type: {qweight_type}")
+
+
+def _apply_gguf_embedding_fake(
+    x: torch.Tensor,
+    qweight: torch.Tensor,
+    qweight_type: int,
+    hidden_size: int,
+    dtype: Optional[torch.dtype] = None,
+) -> torch.Tensor:
+    return torch.empty(x.shape[0], hidden_size, dtype=dtype, device=x.device)
+
+
+try:
+    direct_register_custom_op(
+        op_name="_apply_gguf_embedding",
+        op_func=_apply_gguf_embedding,
+        mutates_args=[],
+        fake_impl=_apply_gguf_embedding_fake,
+    )
+    apply_gguf_embedding = torch.ops.vllm._apply_gguf_embedding
+
+except AttributeError as error:
+    raise error
+
+
 class GGUFLinearMethod(LinearMethodBase):
     """Linear method for GGUF.
 
@@ -249,26 +358,76 @@ class GGUFLinearMethod(LinearMethodBase):
         set_weight_attrs(qweight_type, extra_weight_attrs)
         layer.register_parameter("qweight_type", qweight_type)
 
+    def process_weights_after_loading(self, layer: torch.nn.Module):
+        qweight_type = layer.qweight_type.weight_type
+        if not (qweight_type in UNQUANTIZED_TYPES
+                or qweight_type in DEQUANT_TYPES):
+            qweight_type = WeightType(qweight_type)
+            raise ValueError(
+                f"Unsupported GGUF quantization type {qweight_type} in "
+                f"layer {layer}.")
+        # For MergedColumnParallelLinear and QKVParallelLinear, we need to
+        # materialize the padded weight parameter for CUDA Graph compatibility.
+        self._create_padded_weight_param(layer)
+
+    def _create_padded_weight_param(self, layer: torch.nn.Module):
+        """Create padded weight parameter for GGUF MergedLinear layer."""
+        qweight = layer.qweight
+        shard_id_map = qweight.shard_id_map
+        shard_id = qweight.shard_id
+        if len(data_container := qweight.data_container) > 1:
+            dtype = {data.dtype for data in data_container}
+            assert len(dtype) == 1, ValueError(
+                f"Data container has mixed dtypes: {dtype}")
+            dtype = next(iter(dtype))
+            # concat dim0 and pad dim1
+            padded_side = max(x.size(1) for x in data_container)
+            concat_side = sum(x.size(0) for x in data_container)
+            # Pad the quantized weights to dense tensor, and create a map
+            # with the location of each shard in the padded tensor.
+            padded_data = torch.zeros((concat_side, padded_side),
+                                      dtype=dtype,
+                                      device=qweight.device)
+            # (dim0_start, dim0_end, dim1_size)
+            shard_offset_map = dict[str, tuple[int, int, int]]()
+            for idx in shard_id:
+                id_in_container = shard_id_map[idx]
+                start = sum(
+                    x.size(0) for x in data_container[:id_in_container])
+                end = start + data_container[id_in_container].size(0)
+                size = data_container[id_in_container].size(1)
+                padded_data[start:end, :size] = data_container[id_in_container]
+                shard_offset_map[idx] = (start, end, size)
+            qweight.data_container.clear()
+            padded_param = Parameter(padded_data, requires_grad=False)
+            set_weight_attrs(padded_param, vars(qweight))
+            set_weight_attrs(padded_param,
+                             {"shard_offset_map": shard_offset_map})
+            layer.register_parameter("qweight", padded_param)
+
     def apply(self,
               layer: torch.nn.Module,
               x: torch.Tensor,
               bias: Optional[torch.Tensor] = None) -> torch.Tensor:
-        shard_id = getattr(layer.qweight, "shard_id", None)
+        shard_id = layer.qweight.shard_id
 
         if shard_id:
             # dequantize shard weights respectively
             shard_id = ["q", "k", "v"] if "q" in shard_id else shard_id
-            qweight = layer.qweight.unbind(0)
+            qweight = layer.qweight
             result = []
             for idx in shard_id:
-                q_idx = layer.qweight.shard_id_map[idx]
+                start, end, offset = layer.qweight.shard_offset_map[idx]
                 qweight_type = layer.qweight_type.shard_weight_type[idx]
-                result.append(_fuse_mul_mat(x, qweight[q_idx], qweight_type))
+                result.append(
+                    fused_mul_mat_gguf(
+                        x, qweight[start:end, :offset].contiguous(),
+                        qweight_type))
             out = torch.cat(result, axis=1)
         else:
             qweight = layer.qweight
             qweight_type = layer.qweight_type.weight_type
-            out = _fuse_mul_mat(x, qweight, qweight_type)
+            out = fused_mul_mat_gguf(x, qweight, qweight_type)
         if bias is not None:
             out.add_(bias)
         return out
@@ -338,7 +497,6 @@ class GGUFMoEMethod(FusedMoEMethodBase):
 
         set_weight_attrs(w2_qweight_type, extra_weight_attrs)
         layer.register_parameter("w2_qweight_type", w2_qweight_type)
-        self.act = SiluAndMul()
 
     def apply(
         self,
@@ -375,10 +533,10 @@ class GGUFMoEMethod(FusedMoEMethodBase):
             custom_routing_function=custom_routing_function,
             scoring_func=scoring_func,
             e_score_correction_bias=e_score_correction_bias)
-        return _fused_moe_gguf(x, layer.w13_qweight, layer.w2_qweight,
-                               topk_weights, topk_ids,
-                               layer.w13_qweight_type.weight_type,
-                               layer.w2_qweight_type.weight_type, self.act)
+        return fused_moe_gguf(x, layer.w13_qweight, layer.w2_qweight,
+                              topk_weights, topk_ids,
+                              layer.w13_qweight_type.weight_type,
+                              layer.w2_qweight_type.weight_type, activation)
 
 
 class GGUFEmbeddingMethod(GGUFLinearMethod):
@@ -392,34 +550,15 @@ class GGUFEmbeddingMethod(GGUFLinearMethod):
                   x: torch.Tensor) -> torch.Tensor:
         qweight = layer.qweight
         qweight_type = layer.qweight_type.weight_type
+        hidden_size = qweight.tensor_shape[1]
 
-        block_size, type_size = gguf.GGML_QUANT_SIZES[qweight_type]
-        hidden_size = qweight.shape[1] // type_size * block_size
-        if qweight_type < 2:
-            return torch.embedding(qweight, x)
-        x_flat = x.flatten()
-        quant = torch.index_select(qweight, dim=0, index=x_flat)
-        dequant = ops.ggml_dequantize(quant, qweight_type, hidden_size,
-                                      x_flat.shape[0], self.params_dtype)
-        return dequant.view(*x.shape, hidden_size)
+        return apply_gguf_embedding(x,
+                                    qweight,
+                                    qweight_type,
+                                    hidden_size,
+                                    dtype=self.params_dtype)
 
 
 class GGUFUninitializedParameter(UninitializedParameter):
     cls_to_become = Parameter
     data_container: list[torch.Tensor]
-
-    def materialize_nested(self) -> Parameter:
-        dtype = {data.dtype for data in self.data_container}
-        assert len(dtype) == 1, ValueError(
-            f"Data container has mixed dtypes: {dtype}")
-        dtype = next(iter(dtype))
-        nested_data = torch.nested.nested_tensor(self.data_container,
-                                                 device=self.device,
-                                                 dtype=dtype)
-        self.data_container.clear()
-        param = torch.Tensor._make_subclass(self.cls_to_become,
-                                            nested_data,
-                                            require_grad=False)
-        for k, v in self.__dict__.items():
-            setattr(param, k, v)
-        return param