optimum/exporters/openvino/model_patcher.py [5047:5087]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        def attn_forward(
            self,
            hidden_states: torch.Tensor,
            attention_mask: Optional[torch.Tensor] = None,
            output_attentions: Optional[bool] = False,
        ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
            if output_attentions:
                return super().forward(
                    hidden_states=hidden_states,
                    attention_mask=attention_mask,
                    output_attentions=output_attentions,
                )

            batch_size, q_len, _ = hidden_states.size()

            query_states = self.q_proj(hidden_states)
            key_states = self.k_proj(hidden_states)
            value_states = self.v_proj(hidden_states)

            query_states = query_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
            key_states = key_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
            value_states = value_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)

            # SDPA with memory-efficient backend is currently (torch==2.1.2) bugged with non-contiguous inputs with custom attn_mask,
            # Reference: https://github.com/pytorch/pytorch/issues/112577.
            if query_states.device.type == "cuda" and attention_mask is not None:
                query_states = query_states.contiguous()
                key_states = key_states.contiguous()
                value_states = value_states.contiguous()

            # We dispatch to SDPA's Flash Attention or Efficient kernels via this `is_causal` if statement instead of an inline conditional assignment
            # in SDPA to support both torch.compile's dynamic shapes and full graph options. An inline conditional prevents dynamic shapes from compiling.
            is_causal = False

            attn_output = torch.nn.functional.scaled_dot_product_attention(
                query_states,
                key_states,
                value_states,
                attn_mask=attention_mask,
                dropout_p=self.dropout if self.training else 0.0,
                is_causal=is_causal,
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



optimum/exporters/openvino/model_patcher.py [6145:6185]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
        def attn_forward(
            self,
            hidden_states: torch.Tensor,
            attention_mask: Optional[torch.Tensor] = None,
            output_attentions: Optional[bool] = False,
        ) -> Tuple[torch.Tensor, Optional[torch.Tensor], Optional[Tuple[torch.Tensor]]]:
            if output_attentions:
                return super().forward(
                    hidden_states=hidden_states,
                    attention_mask=attention_mask,
                    output_attentions=output_attentions,
                )

            batch_size, q_len, _ = hidden_states.size()

            query_states = self.q_proj(hidden_states)
            key_states = self.k_proj(hidden_states)
            value_states = self.v_proj(hidden_states)

            query_states = query_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
            key_states = key_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)
            value_states = value_states.view(batch_size, q_len, self.num_heads, self.head_dim).transpose(1, 2)

            # SDPA with memory-efficient backend is currently (torch==2.1.2) bugged with non-contiguous inputs with custom attn_mask,
            # Reference: https://github.com/pytorch/pytorch/issues/112577.
            if query_states.device.type == "cuda" and attention_mask is not None:
                query_states = query_states.contiguous()
                key_states = key_states.contiguous()
                value_states = value_states.contiguous()

            # We dispatch to SDPA's Flash Attention or Efficient kernels via this `is_causal` if statement instead of an inline conditional assignment
            # in SDPA to support both torch.compile's dynamic shapes and full graph options. An inline conditional prevents dynamic shapes from compiling.
            is_causal = False

            attn_output = torch.nn.functional.scaled_dot_product_attention(
                query_states,
                key_states,
                value_states,
                attn_mask=attention_mask,
                dropout_p=self.dropout if self.training else 0.0,
                is_causal=is_causal,
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -