[OMNIML-2932] [feat] nvfp4 awq support

tensorrt_llm/_torch/modules/attention.py

@@ -358,8 +358,9 @@ def create_output(self, q: torch.Tensor):
         out_dtype = q.dtype
 
         if self.attn_backend == "TRTLLM":
-            if self.has_quant_scale and (self.attn.has_fp8_kv_cache
-                                         or self.attn.has_fp4_kv_cache):
+            # Don't use FP8 output if o_proj has pre_quant_scale - keep BF16 for better precision
+            if self.has_quant_scale and self.o_proj.pre_quant_scale is None and (
+                    self.attn.has_fp8_kv_cache or self.attn.has_fp4_kv_cache):
                 out_dtype = torch.float8_e4m3fn
         output = q.new_empty([num_tokens, hidden_size], dtype=out_dtype)
         return output
@@ -390,8 +391,14 @@ def _attn_impl(
 
         out_scale = None
         out_scale_sf = None
-        if self.has_quant_scale:
+        # Don't set out_scale if o_proj has pre_quant_scale - this prevents FP8/FP4 output
+        # and keeps attention output in BF16 for better precision when applying pre_quant_scale
+        if self.has_quant_scale and self.o_proj.pre_quant_scale is None:
             out_scale = self.o_proj.inv_input_scale
+        if hasattr(
+                self.o_proj,
+                'pre_quant_scale') and self.o_proj.pre_quant_scale is not None:
+            enable_attn_nvfp4_output = False
         if self.o_proj.has_nvfp4 and self.support_nvfp4_output and enable_attn_nvfp4_output:
             out_scale_sf = self.o_proj.input_scale
 

tensorrt_llm/_torch/modules/fused_moe/fused_moe_cutlass.py

@@ -290,6 +290,14 @@ def forward_chunk(
             elif self.has_int8_woq_per_channel:
                 use_int8_woq_per_channel = True
             elif self.has_nvfp4:
+                # Apply pre_quant_scale if it exists (for NVFP4_AWQ)
+                if hasattr(
+                        self,
+                        'fc31_act_scale') and self.fc31_act_scale is not None:
+                    assert not isinstance(
+                        x, Fp4QuantizedTensor
+                    ), "Fp4QuantizedTensor is not expected for AWQ quantization."
+                    x = x * self.fc31_act_scale
                 if run_post_quant_allgather or self.enable_alltoall:
                     if isinstance(x, Fp4QuantizedTensor):
                         assert not x.is_sf_swizzled, "Fp4QuantizedTensor should not be swizzled before communication"

tensorrt_llm/_torch/modules/fused_moe/fused_moe_trtllm_gen.py

@@ -187,7 +187,6 @@ def forward_impl(
     ) -> torch.Tensor:
 
         assert x.dtype == torch.bfloat16
-
         # DeepSeekV3 style routing
         if isinstance(self.routing_method, DeepSeekV3MoeRoutingMethod):
             top_k = self.routing_method.routing_impl.top_k
@@ -232,6 +231,14 @@ def forward_impl(
                 else:
                     x_row = x.shape[0]
                     x_col = x.shape[1]
+
+                    # Apply pre_quant_scale if it exists (for NVFP4_AWQ)
+                    # fc31_act_scale shape: (1, hidden_size)
+                    # x shape: (num_tokens, hidden_size)
+                    if hasattr(self, 'fc31_act_scale'
+                               ) and self.fc31_act_scale is not None:
+                        x = x * self.fc31_act_scale
+
                     x, x_sf = torch.ops.trtllm.fp4_quantize(
                         x, self.fc31_input_scale, self.scaling_vector_size,
                         False, False)
@@ -297,6 +304,9 @@ def forward_impl(
             scale_factor_use_ue8m0 = False
             is_scale_factor_swizzled = False  # use linear layout here
 
+            if hasattr(self,
+                       'fc31_act_scale') and self.fc31_act_scale is not None:
+                x = x * self.fc31_act_scale
             if not run_post_quant_allgather:
                 hidden_states_fp4, hidden_states_scale_linear_fp4 = (
                     torch.ops.trtllm.fp4_quantize(

tensorrt_llm/_torch/modules/fused_moe/quantization.py

@@ -1576,6 +1576,10 @@ def create_weights(self,
                                  requires_grad=False)
         module.register_parameter("fc2_alpha", fc2_alpha)
 
+        # Optional per-channel act scale for NVFP4_AWQ (pre_quant_scale support)
+        # This will be initialized in load_quant_scales if pre_quant_scale exists
+        module.register_parameter("fc31_act_scale", None)
+
         super().create_weights(module, weight_dtype, w3_w1_weight_shape,
                                w2_weight_shape)
 
@@ -1678,12 +1682,32 @@ def load_all_fp4_weight_scales_and_alphas(
                                              dst_fc2_alpha[expert_idx])
 
     def load_quant_scales(self, module: torch.nn.Module, weights: Dict):
+        # Check if pre_quant_scale exists in the checkpoint (for NVFP4_AWQ)
+        has_pre_quant_scale = False
+        if module.weight_loading_mode == MoEWeightLoadingMode.VANILLA:
+            # Check if any expert has pre_quant_scale
+            has_pre_quant_scale = f"0.w1.pre_quant_scale" in weights
+
         # Step1: Load input scales.
         tmp_fc31_input_scale = torch.empty(module.num_experts,
                                            dtype=torch.float32)
         tmp_fc2_input_scale = torch.empty(module.num_experts,
                                           dtype=torch.float32)
 
+        # If pre_quant_scale exists, we need a per-channel act scale for fc31
+        # All experts share the same input, so pre_quant_scale should be identical across experts
+        if has_pre_quant_scale:
+            from ..linear import TensorParallelMode, load_weight_shard
+
+            # Create fc31_act_scale parameter (for gate_up_proj / w3_w1)
+            # Shape: (1, hidden_size) - single vector for all experts (they share the same input)
+            fc31_act_scale = nn.Parameter(torch.empty(1,
+                                                      module.hidden_size,
+                                                      dtype=module.dtype,
+                                                      device='cuda'),
+                                          requires_grad=False)
+            module.register_parameter("fc31_act_scale", fc31_act_scale)
+
         for expert_id in range(module.num_experts):
             if module.weight_loading_mode == MoEWeightLoadingMode.VANILLA:
                 w1_input_scale = weights[f"{expert_id}.w1.input_scale"]
@@ -1710,6 +1734,68 @@ def load_quant_scales(self, module: torch.nn.Module, weights: Dict):
         module.fc2_input_scale.data.copy_(
             tmp_fc2_input_scale.max().reciprocal())
 
+        # Load pre_quant_scale if it exists (for NVFP4_AWQ)
+        if has_pre_quant_scale:
+            from ..linear import TensorParallelMode, load_weight_shard
+
+            # Load fc31 (w3/w1) pre_quant_scales
+            # All experts should have identical pre_quant_scale since they share the same input
+            all_w3_pre_quant_scales = []
+            all_w1_pre_quant_scales = []
+            for expert_id in module.initial_local_expert_ids:
+                w3_pre_quant_scale = load_weight_shard(
+                    weights[f"{expert_id}.w3.pre_quant_scale"],
+                    module.tp_size,
+                    module.tp_rank,
+                    TensorParallelMode.ROW,
+                    device='cuda')
+                w1_pre_quant_scale = load_weight_shard(
+                    weights[f"{expert_id}.w1.pre_quant_scale"],
+                    module.tp_size,
+                    module.tp_rank,
+                    TensorParallelMode.ROW,
+                    device='cuda')
+                all_w3_pre_quant_scales.append(w3_pre_quant_scale)
+                all_w1_pre_quant_scales.append(w1_pre_quant_scale)
+
+            # Verify that all experts have identical pre_quant_scale
+            # (they should be the same since all experts share the same input)
+            w3_reference = all_w3_pre_quant_scales[0]
+            w1_reference = all_w1_pre_quant_scales[0]
+
+            for i, (w3_scale, w1_scale) in enumerate(
+                    zip(all_w3_pre_quant_scales[1:],
+                        all_w1_pre_quant_scales[1:]), 1):
+                if not torch.allclose(
+                        w3_scale, w3_reference, rtol=1e-5, atol=1e-8):
+                    max_diff = (w3_scale - w3_reference).abs().max()
+                    logger.warning(
+                        f"MoE pre_quant_scale: expert {module.initial_local_expert_ids[i]} w3.pre_quant_scale "
+                        f"differs from expert {module.initial_local_expert_ids[0]}! Max diff: {max_diff:.6e}. "
+                        f"All experts should have identical pre_quant_scale since they share the same input. "
+                        f"Using the first expert's value.")
+                    break
+                if not torch.allclose(
+                        w1_scale, w1_reference, rtol=1e-5, atol=1e-8):
+                    max_diff = (w1_scale - w1_reference).abs().max()
+                    logger.warning(
+                        f"MoE pre_quant_scale: expert {module.initial_local_expert_ids[i]} w1.pre_quant_scale "
+                        f"differs from expert {module.initial_local_expert_ids[0]}! Max diff: {max_diff:.6e}. "
+                        f"All experts should have identical pre_quant_scale since they share the same input. "
+                        f"Using the first expert's value.")
+                    break
+
+            # Take the maximum pre_quant_scale between w3 and w1 from the first expert
+            # (all experts should have the same values)
+            # Shape: (hidden_size,)
+            # Keep on CUDA device (w3_reference and w1_reference are already on CUDA)
+            fc31_pre_quant_scale = torch.max(w3_reference, w1_reference).to(
+                dtype=module.dtype, device='cuda')
+
+            # Store as a single vector since all experts share the same pre_quant_scale
+            # This will be broadcasted to all tokens in the forward pass
+            module.fc31_act_scale.data.copy_(fc31_pre_quant_scale.unsqueeze(0))
+
         # Step2: Load weight block scales and alphas.
         self.load_all_fp4_weight_scales_and_alphas(
             module, weights, module.initial_local_expert_ids,

tensorrt_llm/_torch/modules/linear.py

@@ -774,6 +774,10 @@ def create_weights(self, module: Linear, in_features: int,
         module.inv_kv_scales = Parameter(torch.ones(3, dtype=torch.float32),
                                          requires_grad=False)
 
+        # NOTE: Not in all linear we have this tensor - pre_quant_scale is computed as an average and merged with the
+        # LayerNorm for QKV and Gate/Up projection layers when possible. we can see the tensor only for o_proj and down_proj
+        module.pre_quant_scale = None
+
         if bias:
             module.bias = Parameter(torch.empty((out_features), dtype=dtype),
                                     requires_grad=False)
@@ -783,10 +787,28 @@ def create_weights(self, module: Linear, in_features: int,
     def apply(self, module: Linear, input: torch.Tensor,
               bias: Optional[torch.Tensor]):
         if isinstance(input, Fp4QuantizedTensor):
+            # Input is already quantized - this should not happen if pre_quant_scale exists
+            # because we disable FP4 output for attention output when pre_quant_scale is present
+            if module.pre_quant_scale is not None:
+                raise RuntimeError(
+                    "Received FP4 quantized input but pre_quant_scale exists. "
+                    "This indicates FP4 output was not properly disabled for the previous layer."
+                )
             act_fp4, act_sf = input.fp4_tensor, input.scaling_factor
         elif isinstance(input, tuple):
+            # Input is a tuple of (fp4_tensor, scaling_factor)
+            if module.pre_quant_scale is not None:
+                raise RuntimeError(
+                    "Received FP4 quantized tuple input but pre_quant_scale exists. "
+                    "This indicates FP4 output was not properly disabled for the previous layer."
+                )
             act_fp4, act_sf = input
         else:
+            # Input is a regular tensor () - apply pre_quant_scale if it exists (for NVFP4_AWQ)
+            if module.pre_quant_scale is not None:
+                assert input.dtype == module.pre_quant_scale.dtype, "Input dtype and pre_quant_scale dtype must match"
+                input = input * module.pre_quant_scale
+
             act_fp4, act_sf = torch.ops.trtllm.fp4_quantize(
                 input, module.input_scale, module.scaling_vector_size, False)
 
@@ -874,6 +896,24 @@ def load_weights_vanilla(self, module: Linear, weights: List[Dict]) -> None:
         copy_weight(module.alpha, alpha)
         module.scalar_alpha = alpha.item()
 
+        # Load pre_quant_scale if it exists (for NVFP4_AWQ)
+        if "pre_quant_scale" in weights[0]:
+            device = module.weight.device
+            pre_quant_scale = load_weight_shard(
+                weights[0]["pre_quant_scale"],
+                module.tp_size,
+                module.tp_rank,
+                # pre_quant_scale applies to activation as opposed to weight, so flip tp_mode the other way around
+                TensorParallelMode.flip(module.tp_mode),
+                device,
+            )
+
+            module.pre_quant_scale = Parameter(
+                torch.ones((module.in_features, ), dtype=pre_quant_scale.dtype),
+                requires_grad=False).to(device=device)
+
+            copy_weight(module.pre_quant_scale, pre_quant_scale)
+
     def load_weights_fused_qkv_linear(self, module: Linear,
                                       weights: List[Dict]) -> None:
         q_weight, k_weight, v_weight = load_weights_fused_qkv_helper(
@@ -930,6 +970,25 @@ def load_weights_fused_gate_up_linear(self, module: Linear,
         copy_weight(module.alpha, alpha)
         module.scalar_alpha = alpha.item()
 
+        # Load pre_quant_scale if it exists (for NVFP4_AWQ)
+        # NOTE: pre_quant_scale is the same for gate and up since modelopt checks which layer shared the same input
+        if "pre_quant_scale" in weights[0]:
+            device = module.weight.device
+            pre_quant_scale = load_weight_shard(
+                weights[0]["pre_quant_scale"],
+                module.tp_size,
+                module.tp_rank,
+                # pre_quant_scale applies to activation as opposed to weight, so flip tp_mode the other way around
+                TensorParallelMode.flip(module.tp_mode),
+                device,
+            )
+
+            module.pre_quant_scale = Parameter(
+                torch.ones((module.in_features, ), dtype=pre_quant_scale.dtype),
+                requires_grad=False).to(device=device)
+
+            copy_weight(module.pre_quant_scale, pre_quant_scale)
+
 
 class W4A8NVFP4FP8LinearMethod(LinearMethodBase):
 

tensorrt_llm/quantization/mode.py

@@ -44,6 +44,7 @@ class QuantAlgo(StrEnum, metaclass=BaseEnumMeta):
     W4A8_MXFP4_FP8 = auto()
     W4A8_MXFP4_MXFP8 = auto()
     W4A16_MXFP4 = auto()
+    NVFP4_AWQ = auto()
     NO_QUANT = auto()
 
 
@@ -410,6 +411,9 @@ def from_quant_algo(
             quant_mode = QuantMode.from_description(use_fp8_block_scales=True)
         elif quant_algo == QuantAlgo.NVFP4:
             quant_mode = QuantMode.from_description(use_nvfp4=True)
+        elif quant_algo == QuantAlgo.NVFP4_AWQ:
+            # NVFP4_AWQ uses the same QuantMode as NVFP4, distinction is at QuantAlgo level
+            quant_mode = QuantMode.from_description(use_nvfp4=True)
         elif quant_algo == QuantAlgo.W4A8_NVFP4_FP8:
             quant_mode = QuantMode.from_description(use_w4a8_nvfp4_fp8=True)
         elif quant_algo == QuantAlgo.W4A8_MXFP4_FP8:

tensorrt_llm/quantization/quantize.py

@@ -72,6 +72,14 @@ def quantize_layers(
                 else:
                     quant_mode = quant_config.quant_mode
                 init_params["quant_mode"] = quant_mode
+
+                # Auto-detect pre_quant_scale based on quant_algo
+                # For AWQ-based quantization methods that use pre_quant_scale
+                if quant_config.quant_algo in [
+                        QuantAlgo.W4A16_AWQ, QuantAlgo.NVFP4_AWQ,
+                        QuantAlgo.W4A8_AWQ
+                ]:
+                    init_params["pre_quant_scale"] = True
             if "bias" in init_params and not isinstance(module,
                                                         MixtureOfExperts):
                 init_params["bias"] = init_params["bias"] is not None