[FA3 Feature] Support multi modal Llama-3.2-11B-Vision-Instruct (#5103)

benchmark/mmmu/bench_sglang.py

@@ -86,8 +86,8 @@ def eval_mmmu(args):
 
 if __name__ == "__main__":
     parser = argparse.ArgumentParser()
-    args = add_common_sglang_args_and_parse(parser)
     EvalArgs.add_cli_args(parser)
+    args = add_common_sglang_args_and_parse(parser)
     args = parser.parse_args()
 
     eval_mmmu(args)

python/sglang/srt/layers/attention/flashattention_backend.py

@@ -42,6 +42,16 @@ class FlashAttentionMetadata:
     # Page table, the index of KV Cache Tables/Blocks
     page_table: torch.Tensor = None
 
+    # Encoder metadata
+    # Cumulative sequence lengths for encoder key
+    encoder_cu_seqlens_k: torch.Tensor = None
+    # Maximum sequence length for encoder key
+    encoder_max_seq_len_k: int = 0
+    # Sequence lengths for the forward batch
+    encoder_lens_int32: torch.Tensor = None
+    # Page table for the encoder
+    encoder_page_table: torch.Tensor = None
+
     @dataclass
     class LocalAttentionMetadata:
         local_query_start_loc: torch.Tensor = None  # cu_seqlens_q for local attention
@@ -435,6 +445,30 @@ class FlashAttentionBackend(AttentionBackend):
                 )
                 metadata.local_attn_metadata = local_metadata
 
+        # Encoder metadata for cross attention
+        if forward_batch.encoder_lens is not None:
+            assert (
+                forward_batch.encoder_lens.numel() == 1
+            ), "Only encoder size 1 is supported for now"
+
+            metadata.encoder_lens_int32 = forward_batch.encoder_lens.to(torch.int32)
+            metadata.encoder_cu_seqlens_k = torch.nn.functional.pad(
+                torch.cumsum(metadata.encoder_lens_int32, dim=0, dtype=torch.int32),
+                (1, 0),
+            )
+            metadata.encoder_max_seq_len_k = metadata.encoder_lens_int32.max().item()
+            metadata.encoder_page_table = forward_batch.req_to_token_pool.req_to_token[
+                forward_batch.req_pool_indices, : metadata.encoder_max_seq_len_k
+            ]
+
+            # Currently only support forward_batch.encoder_lens.numel() == 1
+            metadata.page_table = forward_batch.req_to_token_pool.req_to_token[
+                forward_batch.req_pool_indices,
+                metadata.encoder_max_seq_len_k : (
+                    metadata.encoder_max_seq_len_k + metadata.max_seq_len_k
+                ),
+            ]
+
         # Convert the page table to a strided format which is needed by FA3 API
         if self.page_size > 1:
             self.strided_indices = torch.arange(
@@ -486,6 +520,7 @@ class FlashAttentionBackend(AttentionBackend):
             if layer.sliding_window_size is not None
             else (-1, -1)
         )
+        causal = not layer.is_cross_attention
 
         # Check if we should use local attention
         use_local_attn = (
@@ -521,6 +556,12 @@ class FlashAttentionBackend(AttentionBackend):
             value_cache = value_cache.view(
                 -1, self.page_size, layer.tp_v_head_num, layer.head_dim
             )
+            if layer.is_cross_attention:
+                page_table = metadata.encoder_page_table
+                cache_seqlens = metadata.encoder_lens_int32
+                cu_seqlens_k = metadata.encoder_cu_seqlens_k
+                window_size = (-1, -1)
+
             o = flash_attn_with_kvcache(
                 q=q.contiguous().view(-1, layer.tp_q_head_num, layer.head_dim),
                 k_cache=key_cache,
@@ -531,7 +572,7 @@ class FlashAttentionBackend(AttentionBackend):
                 cu_seqlens_k_new=cu_seqlens_k if not use_local_attn else None,
                 max_seqlen_q=max_seqlen_q,
                 softmax_scale=layer.scaling,
-                causal=True,
+                causal=causal,
                 window_size=window_size,
                 softcap=layer.logit_cap,
                 k_descale=layer.k_scale,
@@ -614,6 +655,7 @@ class FlashAttentionBackend(AttentionBackend):
             if layer.sliding_window_size is not None
             else (-1, -1)
         )
+        causal = not layer.is_cross_attention
 
         if not self.use_mla:
             # Do multi-head attention
@@ -627,17 +669,27 @@ class FlashAttentionBackend(AttentionBackend):
             )
 
             q_reshaped = q.contiguous().view(-1, layer.tp_q_head_num, layer.head_dim)
+            if layer.is_cross_attention:
+                page_table = metadata.encoder_page_table
+                cache_seqlens = metadata.encoder_lens_int32
+                cu_seqlens_k = metadata.encoder_cu_seqlens_k
+                window_size = (-1, -1)
+            else:
+                page_table = metadata.page_table
+                cache_seqlens = metadata.cache_seqlens_int32
+                cu_seqlens_k = metadata.cu_seqlens_k
+
             o = flash_attn_with_kvcache(
                 q=q_reshaped,
                 k_cache=key_cache,
                 v_cache=value_cache,
-                page_table=metadata.page_table,
+                page_table=page_table,
-                cache_seqlens=metadata.cache_seqlens_int32,
+                cache_seqlens=cache_seqlens,
                 cu_seqlens_q=metadata.cu_seqlens_q,
-                cu_seqlens_k_new=metadata.cu_seqlens_k,
+                cu_seqlens_k_new=cu_seqlens_k,
                 max_seqlen_q=1,
                 softmax_scale=layer.scaling,
-                causal=True,
+                causal=causal,
                 window_size=window_size,
                 softcap=layer.logit_cap,
                 k_descale=layer.k_scale,
@@ -733,6 +785,21 @@ class FlashAttentionBackend(AttentionBackend):
             ),
         }
 
+        self.encoder_metadata = {
+            "encoder_page_table": torch.zeros(
+                max_bs,
+                self.max_context_len,
+                dtype=torch.int32,
+                device=self.device,
+            ),
+            "encoder_lens_int32": torch.zeros(
+                max_bs, dtype=torch.int32, device=self.device
+            ),
+            "encoder_cu_seqlens_k": torch.zeros(
+                max_bs + 1, dtype=torch.int32, device=self.device
+            ),
+        }
+
     def init_forward_metadata_capture_cuda_graph(
         self,
         bs: int,
@@ -818,6 +885,19 @@ class FlashAttentionBackend(AttentionBackend):
 
             self.target_verify_metadata[bs] = metadata
 
+        if encoder_lens is not None:
+            encoder_bs = encoder_lens.numel()
+            metadata.encoder_lens_int32 = self.encoder_metadata["encoder_lens_int32"][
+                :encoder_bs
+            ]
+            metadata.encoder_cu_seqlens_k = self.encoder_metadata[
+                "encoder_cu_seqlens_k"
+            ][: (encoder_bs + 1)]
+
+            metadata.encoder_page_table = self.encoder_metadata["encoder_page_table"][
+                req_pool_indices, :
+            ]
+
         self.forward_metadata = metadata
 
     def init_forward_metadata_replay_cuda_graph(
@@ -903,6 +983,30 @@ class FlashAttentionBackend(AttentionBackend):
             page_table = self.req_to_token[req_pool_indices, : metadata.max_seq_len_k]
             metadata.page_table[:, : metadata.max_seq_len_k].copy_(page_table)
 
+        if encoder_lens is not None:
+            # Only support encoder size 1 for now
+            metadata.encoder_max_seq_len_k = encoder_lens[0]
+            metadata.encoder_lens_int32.copy_(encoder_lens[:1])
+            metadata.encoder_cu_seqlens_k.copy_(
+                torch.nn.functional.pad(
+                    torch.cumsum(metadata.encoder_lens_int32, dim=0, dtype=torch.int32),
+                    (1, 0),
+                )
+            )
+
+            metadata.encoder_page_table[:, : metadata.encoder_max_seq_len_k].copy_(
+                self.req_to_token[req_pool_indices, : metadata.encoder_max_seq_len_k]
+            )
+
+            # Update the regular page table
+            page_table = self.req_to_token[
+                req_pool_indices,
+                metadata.encoder_max_seq_len_k : (
+                    metadata.encoder_max_seq_len_k + metadata.max_seq_len_k
+                ),
+            ]
+            metadata.page_table[:, : metadata.max_seq_len_k].copy_(page_table)
+
         self.forward_metadata = metadata
 
     def get_cuda_graph_seq_len_fill_value(self):
@@ -956,7 +1060,7 @@ class FlashAttentionMultiStepBackend:
                 forward_batch.batch_size * self.topk,
                 forward_batch.req_pool_indices,
                 forward_batch.seq_lens,
-                encoder_lens=None,
+                encoder_lens=forward_batch.encoder_lens,
                 forward_mode=ForwardMode.DECODE,
                 spec_info=forward_batch.spec_info,
             )
@@ -973,7 +1077,7 @@ class FlashAttentionMultiStepBackend:
                 forward_batch.req_pool_indices,
                 forward_batch.seq_lens,
                 forward_batch.seq_lens_sum,
-                encoder_lens=None,
+                encoder_lens=forward_batch.encoder_lens,
                 forward_mode=ForwardMode.DECODE,
                 spec_info=forward_batch.spec_info,
                 seq_lens_cpu=forward_batch.seq_lens_cpu,

python/sglang/srt/model_executor/model_runner.py

@@ -886,7 +886,7 @@ class ModelRunner:
                 "Please use `--attention-backend flashinfer`."
             )
             logger.warning(
-                "FlashAttention v3 Backend is in Beta. Multimodal, FP8, and Speculative Decoding are not supported."
+                "FlashAttention v3 Backend is in Beta. FP8 is not supported."
             )
             from sglang.srt.layers.attention.flashattention_backend import (
                 FlashAttentionBackend,