114514

vllm/config.py

@@ -155,7 +155,7 @@ class ModelConfig:
         self.tokenizer_mode = tokenizer_mode
 
     def _verify_quantization(self) -> None:
-        supported_quantization = ["awq", "gptq", "squeezellm", "marlin"]
+        supported_quantization = ["awq", "gptq", "squeezellm", "marlin","myq"]
         rocm_not_supported_quantization = ["awq", "marlin"]
         if self.quantization is not None:
             self.quantization = self.quantization.lower()

vllm/engine/arg_utils.py

@@ -208,7 +208,7 @@ class EngineArgs:
         parser.add_argument('--quantization',
                             '-q',
                             type=str,
-                            choices=['awq', 'gptq', 'squeezellm', None],
+                            choices=['awq', 'gptq', 'squeezellm','myq', None],
                             default=EngineArgs.quantization,
                             help='Method used to quantize the weights. If '
                             'None, we first check the `quantization_config` '

vllm/model_executor/layers/quantization/__init__.py

@@ -3,11 +3,13 @@ from typing import Type
 from vllm.model_executor.layers.quantization.base_config import QuantizationConfig
 from vllm.model_executor.layers.quantization.awq import AWQConfig
 from vllm.model_executor.layers.quantization.gptq import GPTQConfig
+from vllm.model_executor.layers.quantization.myq import MYQConfig
 from vllm.model_executor.layers.quantization.squeezellm import SqueezeLLMConfig
 from vllm.model_executor.layers.quantization.marlin import MarlinConfig
 
 _QUANTIZATION_CONFIG_REGISTRY = {
     "awq": AWQConfig,
+    "myq": MYQConfig,
     "gptq": GPTQConfig,
     "squeezellm": SqueezeLLMConfig,
     "marlin": MarlinConfig,

vllm/model_executor/layers/quantization/myq.py

+import enum
+from enum import Enum
+from typing import Any, Dict, List, Optional
+from fractions import Fraction
+
+import torch
+from torch.nn.parameter import Parameter
+
+from vllm._C import ops
+from vllm.model_executor.layers.linear import (LinearMethodBase,
+                                               set_weight_attrs)
+from vllm.model_executor.layers.quantization.base_config import (
+    QuantizationConfig)
+
+
+class MYQConfig(QuantizationConfig):
+    """Config class for MYQ.
+
+    Reference: https://arxiv.org/abs/2210.17323
+    """
+
+    def __init__(
+        self,
+        weight_bits: int,
+        group_size: int,
+        desc_act: bool,
+    ) -> None:
+        self.weight_bits = weight_bits
+        self.group_size = group_size
+        self.desc_act = desc_act
+        self.pack_factor = Fraction(32, self.weight_bits)
+        if self.weight_bits not in [2, 3, 4, 8]:
+            raise ValueError(
+                "Currently, only 2/3/4/8-bit weight quantization is supported for "
+                f"MYQ, but got {self.weight_bits} bits.")
+
+    def __repr__(self) -> str:
+        return (f"MYQConfig(weight_bits={self.weight_bits}, "
+                f"group_size={self.group_size}, "
+                f"desc_act={self.desc_act})")
+
+    @classmethod
+    def get_name(cls) -> str:
+        return "myq"
+
+    @classmethod
+    def get_supported_act_dtypes(cls) -> List[torch.dtype]:
+        return [torch.half]
+
+    @classmethod
+    # Need to figure it out
+    def get_min_capability(cls) -> int:
+        return 60
+
+    @classmethod
+    def get_config_filenames(cls) -> List[str]:
+        return ["quantize_config.json"]
+
+    @classmethod
+    def from_config(cls, config: Dict[str, Any]) -> "MYQConfig":
+        weight_bits = cls.get_from_keys(config, ["bits"])
+        group_size = cls.get_from_keys(config, ["group_size"])
+        desc_act = cls.get_from_keys(config, ["desc_act"])
+        return cls(weight_bits, group_size, desc_act)
+
+    def get_linear_method(self) -> "MYQLinearMethod":
+        return MYQLinearMethod(self)
+
+    def get_scaled_act_names(self) -> List[str]:
+        return []
+
+
+class ExllamaState(Enum):
+
+    UNUSED = enum.auto()
+    UNINITIALIZED = enum.auto()
+    READY = enum.auto()
+
+
+class MYQLinearMethod(LinearMethodBase):
+    """Linear method for MYQ.
+
+    Args:
+        quant_config: The MYQ quantization config.
+    """
+
+    def __init__(self, quant_config: MYQConfig):
+        self.quant_config = quant_config
+
+    def create_weights(
+        self,
+        input_size_per_partition: int,
+        output_size_per_partition: int,
+        input_size: int,
+        output_size: int,
+        params_dtype: torch.dtype,
+    ) -> Dict[str, Any]:
+        del output_size  # Unused.
+        if input_size_per_partition % self.quant_config.group_size != 0:
+            raise ValueError(
+                "The input size is not aligned with the quantized "
+                "weight shape. This can be caused by too large "
+                "tensor parallel size.")
+        if output_size_per_partition % self.quant_config.pack_factor.numerator != 0:
+            raise ValueError(
+                "The output size is not aligned with the quantized "
+                "weight shape. This can be caused by too large "
+                "tensor parallel size.")
+
+        if self.quant_config.group_size != -1:
+            group_size = self.quant_config.group_size
+        else:
+            group_size = input_size
+        exllama_state = ExllamaState.UNINITIALIZED
+        scale_and_zero_size = input_size // group_size
+        scale_and_zero_input_dim = None
+        if input_size != input_size_per_partition and self.quant_config.group_size != -1:
+            # For act-order models, we cannot use Exllama for row parallel layer
+            if self.quant_config.desc_act:
+                exllama_state = ExllamaState.UNUSED
+            else:
+                # we need to partition qzeros and scales for exllama kernel
+                scale_and_zero_size = input_size_per_partition // group_size
+                scale_and_zero_input_dim = 0
+
+        qweight = Parameter(
+            torch.empty(
+                input_size_per_partition // self.quant_config.pack_factor,
+                output_size_per_partition,
+                dtype=torch.int32,
+            ),
+            requires_grad=False,
+        )
+        set_weight_attrs(
+            qweight, {
+                "input_dim": 0,
+                "output_dim": 1,
+                "packed_dim": 0,
+                "pack_factor": self.quant_config.pack_factor,
+            })
+        g_idx = Parameter(
+            torch.tensor(
+                [
+                    i // self.quant_config.group_size
+                    for i in range(input_size_per_partition)
+                ],
+                dtype=torch.int32,
+            ),
+            requires_grad=False,
+        )
+        # Ignore warning from fused linear layers such as QKVParallelLinear.
+        set_weight_attrs(g_idx, {"input_dim": 0, "ignore_warning": True})
+        qzeros = Parameter(
+            torch.empty(
+                scale_and_zero_size,
+                output_size_per_partition // self.quant_config.pack_factor,
+                dtype=torch.int32,
+            ),
+            requires_grad=False,
+        )
+        set_weight_attrs(
+            qzeros, {
+                "input_dim": scale_and_zero_input_dim,
+                "output_dim": 1,
+                "packed_dim": 1,
+                "pack_factor": self.quant_config.pack_factor,
+            })
+        scales = Parameter(
+            torch.empty(
+                scale_and_zero_size,
+                output_size_per_partition,
+                dtype=params_dtype,
+            ),
+            requires_grad=False,
+        )
+        set_weight_attrs(scales, {
+            "input_dim": scale_and_zero_input_dim,
+            "output_dim": 1,
+        })
+        return {
+            "qweight": qweight,
+            "g_idx": g_idx,
+            "qzeros": qzeros,
+            "scales": scales,
+            "exllama_state": exllama_state,
+        }
+
+    def apply_weights(self,
+                      weights: Dict[str, Any],
+                      x: torch.Tensor,
+                      bias: Optional[torch.Tensor] = None) -> torch.Tensor:
+        qweight = weights["qweight"]
+        out_shape = x.shape[:-1] + (qweight.shape[-1], )
+        reshaped_x = x.reshape(-1, x.shape[-1])
+        # exllama needs to shuffle the weight after the weight is loaded
+        # here we do the shuffle on first forward pass
+        if weights["exllama_state"] == ExllamaState.UNINITIALIZED:
+            if self.quant_config.desc_act:
+                weights["g_idx"] = torch.argsort(weights["g_idx"]).to(
+                    torch.int)
+            else:
+                weights["g_idx"] = torch.empty((1, 1), device="meta")
+            weights["exllama_state"] = ExllamaState.READY
+            ops.gptq_shuffle(weights["qweight"], weights["g_idx"],
+                             self.quant_config.weight_bits)
+        output = ops.gptq_gemm(reshaped_x, weights["qweight"],
+                               weights["qzeros"], weights["scales"],
+                               weights["g_idx"],
+                               weights["exllama_state"] == ExllamaState.READY,
+                               self.quant_config.weight_bits)
+        if bias is not None:
+            output = output + bias
+        return output.reshape(out_shape)