feat: add support for qwen2 vl model

server/text_generation_server/layers/rotary.py

@@ -89,6 +89,8 @@ class PositionRotaryEmbedding(nn.Module):
 
             if rope_type == "linear":
                 pass
+            elif rope_type == "default":
+                pass
             elif rope_type == "dynamic":
                 scaling_factor = rope_scaling["factor"]
                 return DynamicPositionRotaryEmbedding(
@@ -275,6 +277,32 @@ class PositionRotaryEmbedding(nn.Module):
         # Note: this unsqueeze is not necessary on RoCm + VLLM ROPE implementation, but we leave it as is to avoid yet an other controlflow.
         return cos.unsqueeze(1), sin.unsqueeze(1)
 
+    def get_cos_sin_hack(
+        self, position_ids: torch.Tensor, max_s: int, dtype: torch.dtype
+    ):
+        # TODO: avoid always computing, use the cache and update it if necessary
+        inv_freq_expanded = (
+            self.inv_freq[None, None, :, None]
+            .float()
+            .expand(3, position_ids.shape[1], -1, 1)
+        )
+
+        position_ids_expanded = position_ids[
+            :, :, None, :
+        ].float()  # shape (3, bs, 1, positions)
+        freqs = (inv_freq_expanded.float() @ position_ids_expanded.float()).transpose(
+            2, 3
+        )
+        emb = torch.cat((freqs, freqs), dim=-1)
+        cos = emb.cos().to(dtype)
+        sin = emb.sin().to(dtype)
+
+        # Update cached values
+        self._cos_cached = cos
+        self._sin_cached = sin
+
+        return cos, sin
+
 
 class SuRotaryEmbedding(PositionRotaryEmbedding):
     def __init__(

server/text_generation_server/layers/tensor_parallel.py

@@ -144,7 +144,7 @@ class TensorParallelColumnLinear(SuperLayer):
             num_key_value_heads=num_key_value_heads,
         )
         if bias:
-            raise NotImplementedError("packed_qkv only implemented for baichuan")
+            bias = weights.get_tensor(f"{prefix}.bias")
         else:
             bias = None
         linear = get_linear(weight, bias)

server/text_generation_server/models/__init__.py

@@ -146,6 +146,9 @@ try:
     from text_generation_server.models.custom_modeling.idefics2 import (
         Idefics2ForConditionalGeneration,
     )
+    from text_generation_server.models.custom_modeling.qwen2_vl import (
+        Qwen2VLForConditionalGeneration,
+    )
     from text_generation_server.layers.attention import SUPPORTS_WINDOWING
 except ImportError as e:
     log_master(logger.warning, f"Could not import Flash Attention enabled models: {e}")
@@ -275,6 +278,11 @@ class ModelType(enum.Enum):
         "name": "Qwen 2",
         "url": "https://huggingface.co/collections/Qwen/qwen2-6659360b33528ced941e557f",
     }
+    QWEN2_VL = {
+        "type": "qwen2_vl",
+        "name": "Qwen 2 VL",
+        "url": "https://huggingface.co/collections/Qwen/qwen2-vl-66cee7455501d7126940800d",
+    }
     OPT = {
         "type": "opt",
         "name": "Opt",
@@ -1193,6 +1201,18 @@ def get_model(
             )
         else:
             raise NotImplementedError(FLASH_ATT_ERROR_MESSAGE.format("Idefics"))
+    if model_type == QWEN2_VL:
+        return VlmCausalLM(
+            model_id=model_id,
+            model_class=Qwen2VLForConditionalGeneration,
+            revision=revision,
+            quantize=quantize,
+            speculator=speculator,
+            dtype=dtype,
+            kv_cache_dtype=kv_cache_dtype,
+            trust_remote_code=trust_remote_code,
+            lora_adapter_ids=lora_adapter_ids,
+        )
     if model_type == MLLAMA:
         if FLASH_ATTENTION:
             return MllamaCausalLM(

server/text_generation_server/models/custom_modeling/flash_qwen2_modeling.py

@@ -49,6 +49,13 @@ def _load_gqa(config, prefix: str, weights):
     )
 
 
+def rotate_half(x):
+    """Rotates half the hidden dims of the input."""
+    x1 = x[..., : x.shape[-1] // 2]
+    x2 = x[..., x.shape[-1] // 2 :]
+    return torch.cat((-x2, x1), dim=-1)
+
+
 class Qwen2Attention(torch.nn.Module):
     def __init__(
         self,
@@ -61,6 +68,7 @@ class Qwen2Attention(torch.nn.Module):
             config.sliding_window if config.sliding_window is not None else -1
         )
         self.num_heads = config.num_attention_heads
+        self.mrope_section = config.rope_scaling.get("mrope_section", None)
         self.hidden_size = config.hidden_size
         self.head_size = self.hidden_size // self.num_heads
 
@@ -122,7 +130,28 @@ class Qwen2Attention(torch.nn.Module):
         query = query.view(-1, self.num_heads, self.head_size)
         kv = kv.view(-1, 2, self.num_key_value_heads, self.head_size)
 
-        self.rotary_emb(query, torch.select(kv, dim=1, index=0), cos, sin)
+        # TODO: correctly handle the multimodal case
+        if False:
+            self.rotary_emb(query, torch.select(kv, dim=1, index=0), cos, sin)
+        else:
+            # multimodal rotary
+            unsqueeze_dim = 1
+            mrope_section = self.mrope_section * 2
+            cos = torch.cat(
+                [m[i % 3] for i, m in enumerate(cos.split(mrope_section, dim=-1))],
+                dim=-1,
+            ).unsqueeze(unsqueeze_dim)
+            sin = torch.cat(
+                [m[i % 3] for i, m in enumerate(sin.split(mrope_section, dim=-1))],
+                dim=-1,
+            ).unsqueeze(unsqueeze_dim)
+
+            _query = query.transpose(0, 1).unsqueeze(0)
+            _key = torch.select(kv, dim=1, index=0).transpose(0, 1).unsqueeze(0)
+            q_embed = (_query * cos) + (rotate_half(_query) * sin)
+            k_embed = (_key * cos) + (rotate_half(_key) * sin)
+            query = q_embed.squeeze(0).transpose(0, 1)
+            kv[:, 0] = k_embed.squeeze(0).transpose(0, 1)
 
         if prefill_cache_indices is not None:
             kv_to_cache = kv[prefill_cache_indices]
@@ -306,12 +335,20 @@ class Qwen2Model(torch.nn.Module):
         max_s: int,
         true_max_s: int,
         prefill_cache_indices: Optional[torch.Tensor],
+        inputs_embeds: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
     ) -> torch.Tensor:
-        hidden_states = self.embed_tokens(input_ids)
+
+        # if inputs_embeds are supplied from an external model (vision model) then avoid embedding input_ids
+        if inputs_embeds is not None:
+            hidden_states = inputs_embeds.squeeze(0)
+        else:
+            hidden_states = self.embed_tokens(input_ids)
 
         # Get rotary cos and sin for this forward
         # Avoid to index in each layer
-        cos, sin = self.layers[0].self_attn.rotary_emb.get_cos_sin(
+        # TODO: fix how N-D position_ids are handled
+        cos, sin = self.layers[0].self_attn.rotary_emb.get_cos_sin_hack(
             position_ids, true_max_s, hidden_states.dtype
         )
 

server/text_generation_server/models/custom_modeling/qwen2_vl.py

@@ -0,0 +1,544 @@
+# coding=utf-8
+# Copyright 2024 the HuggingFace Inc. team. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""PyTorch Qwen2 VL model."""
+
+from typing import Optional, Tuple, List
+
+import torch
+import torch.utils.checkpoint
+from torch import nn
+from text_generation_server.utils.import_utils import SYSTEM
+
+if SYSTEM == "ipex":
+    import intel_extension_for_pytorch as ipex
+else:
+    import flash_attn_2_cuda
+
+from transformers.activations import ACT2FN
+import torch.nn.functional as F
+
+from text_generation_server.layers.layernorm import (
+    FastLayerNorm,
+)
+from text_generation_server.layers import (
+    TensorParallelColumnLinear,
+    FastLinear,
+)
+from text_generation_server.layers.attention import (
+    Seqlen,
+)
+from text_generation_server.models.custom_modeling.flash_qwen2_modeling import (
+    Qwen2Model,
+)
+
+
+# Copied from transformers.models.llama.modeling_llama.rotate_half
+def rotate_half(x):
+    """Rotates half the hidden dims of the input."""
+    x1 = x[..., : x.shape[-1] // 2]
+    x2 = x[..., x.shape[-1] // 2 :]
+    return torch.cat((-x2, x1), dim=-1)
+
+
+def apply_rotary_pos_emb_vision(
+    tensor: torch.Tensor, freqs: torch.Tensor
+) -> torch.Tensor:
+    orig_dtype = tensor.dtype
+    tensor = tensor.float()
+    cos = freqs.cos()
+    sin = freqs.sin()
+    cos = cos.unsqueeze(1).repeat(1, 1, 2).unsqueeze(0).float()
+    sin = sin.unsqueeze(1).repeat(1, 1, 2).unsqueeze(0).float()
+    output = (tensor * cos) + (rotate_half(tensor) * sin)
+    output = output.to(orig_dtype)
+    return output
+
+
+class Qwen2VLSdpaAttention(nn.Module):
+    def __init__(self, *, prefix, config, weights):
+        super().__init__()
+        self.embed_dim = config.embed_dim
+        self.head_dim = config.hidden_size // config.num_heads
+        self.num_heads = config.num_heads // weights.process_group.size()
+
+        self.qkv = TensorParallelColumnLinear.load_qkv(
+            config,
+            prefix=f"{prefix}.qkv",
+            weights=weights,
+            bias=True,
+            num_heads=self.num_heads,
+            num_key_value_heads=self.num_heads,
+        )
+
+        self.proj = TensorParallelColumnLinear.load(
+            config,
+            prefix=f"{prefix}.proj",
+            weights=weights,
+            bias=True,
+        )
+
+    def forward(
+        self,
+        hidden_state: torch.Tensor,
+        cu_seqlens: torch.Tensor,
+        rotary_pos_emb: torch.Tensor,
+        attention_mask: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        # apply the qkv linear layer to the hidden state
+        qkv = self.qkv(hidden_state)
+        query, key, value = qkv.split(
+            [self.embed_dim, self.embed_dim, self.embed_dim], dim=1
+        )
+
+        # reshape the query, key, and value tensors
+        _shape = (
+            hidden_state.shape[0],
+            self.num_heads,
+            self.embed_dim // self.num_heads,
+        )
+        query = query.view(*_shape)
+        key = key.view(*_shape)
+        value = value.view(*_shape)
+
+        # apply rotary positional embeddings
+        query = apply_rotary_pos_emb_vision(query.unsqueeze(0), rotary_pos_emb).squeeze(
+            0
+        )
+        key = apply_rotary_pos_emb_vision(key.unsqueeze(0), rotary_pos_emb).squeeze(0)
+        # TODO: make use of existing RotatoryPositionEmbedding class
+
+        # create the attention mask
+        attention_mask = torch.zeros(
+            [1, hidden_state.shape[0], hidden_state.shape[0]],
+            device=hidden_state.device,
+            dtype=torch.bool,
+        )
+        # TODO: avoid creating the mask in the forward pass, instead define the largest possible mask and slice it
+
+        # apply the cu_seqlens to the attention mask
+        for i in range(1, len(cu_seqlens)):
+            attention_mask[
+                ...,
+                cu_seqlens[i - 1] : cu_seqlens[i],
+                cu_seqlens[i - 1] : cu_seqlens[i],
+            ] = True
+
+        # transpose for the attention mechanism (batch, seqlen, hidden_dim) -> (seqlen, batch, hidden_dim)
+        query = query.transpose(0, 1)
+        key = key.transpose(0, 1)
+        value = value.transpose(0, 1)
+
+        # apply attention
+        attn_output = F.scaled_dot_product_attention(
+            query, key, value, attention_mask, dropout_p=0.0
+        )
+        attn_output = attn_output.transpose(0, 1)
+        attn_output = attn_output.reshape(hidden_state.shape[0], -1)
+        # TODO: prefer flash attention
+
+        attn_output = self.proj(attn_output)
+        return attn_output
+
+
+class Qwen2VLVisionMLP(nn.Module):
+    def __init__(self, *, prefix, config, weights):
+        super().__init__()
+        self.activation_fn = ACT2FN[config.hidden_act]
+        self.fc1 = TensorParallelColumnLinear.load(
+            prefix=f"{prefix}.fc1", weights=weights, config=config, bias=True
+        )
+        self.fc2 = TensorParallelRowLinear.load(
+            prefix=f"{prefix}.fc2", weights=weights, config=config, bias=True
+        )
+
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+        hidden_states = self.fc1(hidden_states)
+        hidden_states = self.activation_fn(hidden_states)
+        hidden_states = self.fc2(hidden_states)
+        return hidden_states
+
+
+class Qwen2VLVisionBlock(nn.Module):
+    def __init__(self, prefix, config, weights):
+        super().__init__()
+        self.attn = Qwen2VLSdpaAttention(
+            prefix=f"{prefix}.attn",
+            config=config,
+            weights=weights,
+        )
+        self.norm1 = FastLayerNorm.load(
+            prefix=f"{prefix}.norm1",
+            weights=weights,
+            eps=1e-6,
+        )
+        self.norm2 = FastLayerNorm.load(
+            prefix=f"{prefix}.norm2",
+            weights=weights,
+            eps=1e-6,
+        )
+        self.mlp = Qwen2VLVisionMLP(
+            prefix=f"{prefix}.mlp",
+            config=config,
+            weights=weights,
+        )
+
+    def forward(self, hidden_states, cu_seqlens, rotary_pos_emb) -> torch.Tensor:
+        hidden_states_post_norm1, res = self.norm1(hidden_states)
+        hidden_states = hidden_states + self.attn(
+            hidden_states_post_norm1, cu_seqlens, rotary_pos_emb
+        )
+        hidden_states_post_norm2, res = self.norm2(hidden_states)
+        hidden_states = hidden_states + self.mlp(hidden_states_post_norm2)
+        return hidden_states
+
+
+class Qwen2VLPatchMerger(nn.Module):
+    def __init__(self, *, prefix, config, weights):
+        super().__init__()
+        context_dim = 2560
+        spatial_merge_size: int = 2
+        self.hidden_size = 5120  # context_dim * (spatial_merge_size**2)
+        self.patch_merger_ln_q = FastLayerNorm.load(
+            prefix=f"{prefix}.ln_q",
+            weights=weights,
+            eps=1e-6,
+        )
+        self.fc1 = TensorParallelColumnLinear.load(
+            prefix=f"{prefix}.mlp.0", weights=weights, config=config, bias=True
+        )
+        self.fc2 = TensorParallelRowLinear.load(
+            prefix=f"{prefix}.mlp.2", weights=weights, config=config, bias=True
+        )
+
+    def forward(self, hidden_states, grid_thw) -> torch.Tensor:
+        hidden_states, _ = self.patch_merger_ln_q(hidden_states)
+        hidden_states = hidden_states.view(-1, self.hidden_size)
+        hidden_states = self.fc1(hidden_states)
+        hidden_states = F.gelu(hidden_states)
+        hidden_states = self.fc2(hidden_states)
+        return hidden_states
+
+
+class Qwen2VisionModel(nn.Module):
+    def __init__(self, *, prefix, config, weights):
+        super().__init__()
+        self.spatial_merge_size = config.spatial_merge_size
+        kernel_size = [config.temporal_patch_size, config.patch_size, config.patch_size]
+        self.patch_embedding = nn.Conv3d(
+            in_channels=config.in_chans,
+            out_channels=config.embed_dim,
+            kernel_size=kernel_size,
+            stride=kernel_size,
+            bias=False,
+        )
+        self.patch_embedding.weight = nn.Parameter(
+            weights.get_tensor(f"{prefix}.patch_embed.proj.weight"), requires_grad=False
+        )
+        head_dim = config.embed_dim // config.num_heads
+        # TODO: replace with static positional embeddings once implemented
+        theta = 10000.0
+        dim = head_dim // 2
+        inv_freq = 1.0 / (theta ** (torch.arange(0, dim, 2, dtype=torch.float) / dim))
+        self.register_buffer("inv_freq", inv_freq, persistent=False)
+
+        self.blocks = nn.ModuleList(
+            [
+                Qwen2VLVisionBlock(
+                    prefix=f"{prefix}.blocks.{i}",
+                    config=config,
+                    weights=weights,
+                )
+                for i in range(config.depth)
+            ]
+        )
+        self.merger = Qwen2VLPatchMerger(
+            prefix=f"{prefix}.merger",
+            config=config,
+            weights=weights,
+        )
+
+        self.temporal_patch_size = config.temporal_patch_size
+        self.spatial_patch_size = config.spatial_patch_size
+        self.in_channels = config.in_channels
+        self.embed_dim = config.embed_dim
+
+    def apply_class_embedding(self, hidden_state: torch.Tensor) -> torch.Tensor:
+        batch_size, _, hidden_size = hidden_state.shape
+        class_embedding = self.class_embedding.expand(batch_size, 1, hidden_size)
+        hidden_state = torch.cat([class_embedding, hidden_state], dim=1)
+        return hidden_state
+
+    def forward(
+        self,
+        pixel_values: torch.Tensor,
+        aspect_ratio_ids: Optional[torch.Tensor] = None,
+        attention_mask: Optional[torch.Tensor] = None,
+        grid_thw: Optional[torch.LongTensor] = None,
+    ) -> torch.Tensor:
+        # reshape the input tensor for processing
+        shape = (
+            -1,
+            self.in_channels,
+            self.temporal_patch_size,
+            self.spatial_patch_size,
+            self.spatial_patch_size,
+        )
+        pixel_values = pixel_values.view(shape).to(self.patch_embedding.weight.dtype)
+        hidden_states = self.patch_embedding(pixel_values).view(-1, self.embed_dim)
+        # TODO: revisit to see if we can avoid some of these reshapes
+
+        # find the position ids for the input tensor based on the grid_thw
+        pos_ids = []
+        for t, h, w in grid_thw:
+            hpos_ids = torch.arange(h).unsqueeze(1).expand(-1, w)
+            hpos_ids = hpos_ids.reshape(
+                h // self.spatial_merge_size,
+                self.spatial_merge_size,
+                w // self.spatial_merge_size,
+                self.spatial_merge_size,
+            )
+            hpos_ids = hpos_ids.permute(0, 2, 1, 3)
+            hpos_ids = hpos_ids.flatten()
+
+            wpos_ids = torch.arange(w).unsqueeze(0).expand(h, -1)
+            wpos_ids = wpos_ids.reshape(
+                h // self.spatial_merge_size,
+                self.spatial_merge_size,
+                w // self.spatial_merge_size,
+                self.spatial_merge_size,
+            )
+            wpos_ids = wpos_ids.permute(0, 2, 1, 3)
+            wpos_ids = wpos_ids.flatten()
+            pos_ids.append(torch.stack([hpos_ids, wpos_ids], dim=-1).repeat(t, 1))
+
+        pos_ids = torch.cat(pos_ids, dim=0)
+        max_grid_size = grid_thw[:, 1:].max()
+
+        # apply the positional embeddings to the position ids
+        seq = torch.arange(
+            max_grid_size, device=self.inv_freq.device, dtype=self.inv_freq.dtype
+        )
+        rotary_pos_emb_full = torch.outer(seq, self.inv_freq)
+        rotary_pos_emb = rotary_pos_emb_full[pos_ids].flatten(1)
+        rotary_pos_emb = rotary_pos_emb.to(hidden_states.device, hidden_states.dtype)
+
+        # create a cu_seqlens tensor to be used in the attention mask
+        cu_seqlens = torch.repeat_interleave(
+            grid_thw[:, 1] * grid_thw[:, 2], grid_thw[:, 0]
+        ).cumsum(dim=0, dtype=torch.int32)
+        cu_seqlens = F.pad(cu_seqlens, (1, 0), value=0)
+
+        # iterately apply the blocks to the hidden states
+        for block in self.blocks:
+            hidden_states = block(hidden_states, cu_seqlens, rotary_pos_emb)
+
+        # apply the final patch merger to the hidden states
+        hidden_states = self.merger(hidden_states, grid_thw)
+        return hidden_states
+
+
+class Qwen2VLForConditionalGeneration(nn.Module):
+    def __init__(self, prefix, config, weights):
+        super().__init__()
+        config.vision_config.quantize = None
+        config.vision_config.speculator = config.speculator
+        self.hidden_size = config.hidden_size
+        self.vision_start_token_id = config.vision_start_token_id
+        self.image_token_id = config.image_token_id
+        self.video_token_id = config.video_token_id
+        self.spatial_merge_size = config.vision_config.spatial_merge_size
+
+        self.visual = Qwen2VisionModel(
+            prefix="visual", config=config.vision_config, weights=weights
+        )
+        self.text_model = Qwen2Model(prefix=None, config=config, weights=weights)
+
+    def forward(
+        self,
+        input_ids: torch.Tensor,
+        position_ids: torch.Tensor,
+        cu_seqlen_prefill: Optional[torch.Tensor],
+        kv_cache: List[Tuple[torch.Tensor, torch.Tensor]],
+        block_tables: torch.Tensor,
+        slots: torch.Tensor,
+        seqlen: Seqlen,
+        max_s: int,
+        prefill_cache_indices: Optional[torch.Tensor],
+        lm_head_indices: Optional[torch.Tensor],
+        pixel_values: torch.FloatTensor = None,
+        image_grid_thw: Optional[torch.LongTensor] = None,
+        video_grid_thw: Optional[torch.LongTensor] = None,
+        pixel_attention_mask=None,
+        image_sizes: Optional[torch.LongTensor] = None,
+        adapter_data: Optional[torch.Tensor] = None,
+        cross_attention_states: Optional[torch.Tensor] = None,
+        image_indices=None,
+    ):
+
+        # make an attention_mask that is the same size as the input_ids
+        attention_mask = torch.ones_like(input_ids, dtype=torch.bool)
+
+        inputs_embeds = self.text_model.embed_tokens(input_ids)
+
+        # apply the visual model to the pixel values if they are provided
+        if pixel_values is not None and len(pixel_values) > 0:
+            if pixel_values is not None:
+                image_embeds = self.visual(pixel_values, grid_thw=image_grid_thw)
+                image_mask = (
+                    (input_ids == self.image_token_id)
+                    .unsqueeze(-1)
+                    .expand_as(inputs_embeds)
+                    .to(inputs_embeds.device)
+                )
+                image_embeds = image_embeds.to(
+                    inputs_embeds.device, inputs_embeds.dtype
+                )
+                # input embeddings are masked with image embeddings
+                inputs_embeds = inputs_embeds.masked_scatter(image_mask, image_embeds)
+
+        # handle the position_ids taking the multimodal inputs into account
+        mrope_position_deltas = []
+        if image_grid_thw is not None or video_grid_thw is not None:
+            total_input_ids = input_ids
+            position_ids = torch.ones(
+                3,
+                input_ids.shape[0],
+                input_ids.shape[1],
+                dtype=input_ids.dtype,
+                device=input_ids.device,
+            )
+            image_index, video_index = 0, 0
+
+            for i, input_ids in enumerate(total_input_ids):
+                if attention_mask is not None:
+                    input_ids = input_ids[attention_mask[i] == 1]
+                image_nums, video_nums = 0, 0
+                vision_start_indices = torch.argwhere(
+                    input_ids == self.vision_start_token_id
+                ).squeeze(1)
+                vision_tokens = input_ids[vision_start_indices + 1]
+                # determine the number of images and videos in the input
+                image_nums = (vision_tokens == self.image_token_id).sum()
+                video_nums = (vision_tokens == self.video_token_id).sum()
+                input_tokens = input_ids.tolist()
+                llm_pos_ids_list: list = []
+                st = 0
+                remain_images, remain_videos = image_nums, video_nums
+                # process each input based on it's token type and grid size
+                for _ in range(image_nums + video_nums):
+                    if self.image_token_id in input_tokens and remain_images > 0:
+                        ed_image = input_tokens.index(self.image_token_id, st)
+                    else:
+                        ed_image = len(input_tokens) + 1
+                    if self.video_token_id in input_tokens and remain_videos > 0:
+                        ed_video = input_tokens.index(self.video_token_id, st)
+                    else:
+                        ed_video = len(input_tokens) + 1
+                    if ed_image < ed_video:
+                        t, h, w = (
+                            image_grid_thw[image_index][0],
+                            image_grid_thw[image_index][1],
+                            image_grid_thw[image_index][2],
+                        )
+                        image_index += 1
+                        remain_images -= 1
+                        ed = ed_image
+                    else:
+                        t, h, w = (
+                            video_grid_thw[video_index][0],
+                            video_grid_thw[video_index][1],
+                            video_grid_thw[video_index][2],
+                        )
+                        video_index += 1
+                        remain_videos -= 1
+                        ed = ed_video
+                    llm_grid_t, llm_grid_h, llm_grid_w = (
+                        t.item(),
+                        h.item() // self.spatial_merge_size,
+                        w.item() // self.spatial_merge_size,
+                    )
+                    text_len = ed - st
+
+                    st_idx = (
+                        llm_pos_ids_list[-1].max() + 1
+                        if len(llm_pos_ids_list) > 0
+                        else 0
+                    )
+                    llm_pos_ids_list.append(
+                        torch.arange(text_len).view(1, -1).expand(3, -1) + st_idx
+                    )
+
+                    t_index = (
+                        torch.arange(llm_grid_t)
+                        .view(-1, 1)
+                        .expand(-1, llm_grid_h * llm_grid_w)
+                        .flatten()
+                    )
+                    h_index = (
+                        torch.arange(llm_grid_h)
+                        .view(1, -1, 1)
+                        .expand(llm_grid_t, -1, llm_grid_w)
+                        .flatten()
+                    )
+                    w_index = (
+                        torch.arange(llm_grid_w)
+                        .view(1, 1, -1)
+                        .expand(llm_grid_t, llm_grid_h, -1)
+                        .flatten()
+                    )
+                    llm_pos_ids_list.append(
+                        torch.stack([t_index, h_index, w_index]) + text_len + st_idx
+                    )
+                    st = ed + llm_grid_t * llm_grid_h * llm_grid_w
+
+                if st < len(input_tokens):
+                    st_idx = (
+                        llm_pos_ids_list[-1].max() + 1
+                        if len(llm_pos_ids_list) > 0
+                        else 0
+                    )
+                    text_len = len(input_tokens) - st
+                    llm_pos_ids_list.append(
+                        torch.arange(text_len).view(1, -1).expand(3, -1) + st_idx
+                    )
+
+                llm_positions = torch.cat(llm_pos_ids_list, dim=1).reshape(3, -1)
+                position_ids[..., i, attention_mask[i] == 1] = llm_positions.to(
+                    position_ids.device
+                )
+                mrope_position_deltas.append(
+                    llm_positions.max() + 1 - len(total_input_ids[i])
+                )
+            mrope_position_deltas = torch.tensor(
+                mrope_position_deltas, device=input_ids.device
+            ).unsqueeze(1)
+
+        # TODO: adjust model to accept 2D position_ids
+        outputs = self.text_model(
+            input_ids=input_ids,
+            position_ids=position_ids,
+            cu_seqlen_prefill=cu_seqlen_prefill,
+            kv_cache=kv_cache,
+            block_tables=block_tables,
+            slots=slots,
+            seqlen=seqlen,
+            max_s=max_s,
+            true_max_s=max_s,
+            prefill_cache_indices=prefill_cache_indices,
+            inputs_embeds=inputs_embeds,
+            attention_mask=attention_mask,
+        )
+
+        return outputs, None

server/text_generation_server/models/vlm_causal_lm.py

@@ -67,6 +67,8 @@ def image_text_replacement(processor, image_input, config, image_id: int) -> str
 
     elif config.model_type == "paligemma":
         return "<image>" * config.text_config.num_image_tokens
+    elif config.model_type == "qwen2_vl":
+        return "<image>"
     else:
         raise RuntimeError(f"Unknown config {config.model_type} for multimodal")
 
@@ -137,6 +139,7 @@ class VlmCausalLMBatch(FlashCausalLMBatch):
     pixel_values: Optional[List[torch.Tensor]]
     pixel_attention_mask: Optional[List[torch.Tensor]]
     image_sizes: Optional[List[Tuple[int, int]]]
+    image_grid_thw: Optional[torch.Tensor]
 
     @classmethod
     @tracer.start_as_current_span("concatenate")
@@ -145,6 +148,7 @@ class VlmCausalLMBatch(FlashCausalLMBatch):
         batch.pixel_values = None
         batch.pixel_attention_mask = None
         batch.image_sizes = None
+        batch.image_grid_thw = None
         return batch
 
     @tracer.start_as_current_span("filter")
@@ -153,6 +157,7 @@ class VlmCausalLMBatch(FlashCausalLMBatch):
         batch.pixel_values = None
         batch.pixel_attention_mask = None
         batch.image_sizes = None
+        batch.image_grid_thw = None
         return batch
 
     @classmethod
@@ -178,6 +183,10 @@ class VlmCausalLMBatch(FlashCausalLMBatch):
                     raise RuntimeError(f"Invalid chunk type {chunk_type}")
 
         if images:
+            # TODO: REMOVE (this is for debugging purposes)
+            images = images[0][0].resize(
+                (images[0][0].width * 2, images[0][0].height * 2)
+            )
             image_inputs = processor.image_processor(images, return_tensors="pt")
         else:
             image_inputs = None
@@ -237,10 +246,15 @@ class VlmCausalLMBatch(FlashCausalLMBatch):
                 batch.image_sizes = image_inputs["image_sizes"].to(device=device)
             else:
                 batch.image_sizes = None
+            if "image_grid_thw" in image_inputs:
+                batch.image_grid_thw = image_inputs["image_grid_thw"].to(device=device)
+            else:
+                batch.image_grid_thw = None
         else:
             batch.pixel_values = None
             batch.pixel_attention_mask = None
             batch.image_sizes = None
+            batch.image_grid_thw = None
         return batch
 
 
@@ -381,8 +395,9 @@ class VlmCausalLM(FlashCausalLM):
                     max_k=batch.max_current_length,
                 )
                 logits, speculative_logits = self.model.forward(
-                    input_ids=input_ids,
+                    # TODO: remove the unsqueeze(0)
-                    position_ids=position_ids,
+                    input_ids=input_ids.unsqueeze(0),
+                    position_ids=position_ids.unsqueeze(0),
                     cu_seqlen_prefill=cu_seqlen_prefill,
                     kv_cache=kv_cache,
                     block_tables=block_tables,
@@ -394,6 +409,7 @@ class VlmCausalLM(FlashCausalLM):
                     pixel_values=batch.pixel_values,
                     pixel_attention_mask=batch.pixel_attention_mask,
                     image_sizes=batch.image_sizes,
+                    image_grid_thw=batch.image_grid_thw,
                 )
                 if batch.prefill_cache_indices is not None:
                     batch.prefill_cache_indices = None
@@ -403,6 +419,8 @@ class VlmCausalLM(FlashCausalLM):
                     batch.pixel_attention_mask = None
                 if batch.image_sizes is not None:
                     batch.image_sizes = None
+                if batch.image_grid_thw is not None:
+                    batch.image_grid_thw = None
                 return logits, speculative_logits
 
         # Copy inputs to the static inputs of the cuda graph