Merge 2ad2ea283e into 37a239ca50

src/transformers/models/glm4v/modeling_glm4v.py

@@ -1053,6 +1053,7 @@ class Glm4vModel(Glm4vPreTrainedModel):
                 device=input_ids.device,
             )
             image_index, video_index = 0, 0
+            video_group_index = 0
             attention_mask = attention_mask.to(total_input_ids.device)
             for i, input_ids in enumerate(total_input_ids):
                 input_ids = input_ids[attention_mask[i] == 1]
@@ -1082,7 +1083,6 @@ class Glm4vModel(Glm4vPreTrainedModel):
 
                 llm_pos_ids_list = []
                 video_frame_num = 1
-
                 for modality_type, start_idx, end_idx in input_type_group:
                     st_idx = llm_pos_ids_list[-1].max() + 1 if len(llm_pos_ids_list) > 0 else 0
 
@@ -1126,7 +1126,11 @@ class Glm4vModel(Glm4vPreTrainedModel):
                             w_index = torch.arange(llm_grid_w).view(1, 1, -1).expand(1, llm_grid_h, -1).flatten()
                             llm_pos_ids_list.append(torch.stack([t_index, h_index, w_index]) + st_idx)
 
-                        video_index += 1
+                        video_group_index += 1
+
+                        if video_group_index >= video_grid_thw[video_index][0]:
+                            video_index += 1
+                            video_group_index = 0
 
                         video_frame_num += 1
 
@@ -1175,7 +1179,13 @@ class Glm4vModel(Glm4vPreTrainedModel):
                 The temporal, height and width of feature shape of each video in LLM.
         """
         pixel_values_videos = pixel_values_videos.type(self.visual.dtype)
-        video_embeds = self.visual(pixel_values_videos, grid_thw=video_grid_thw)
+        # reshape video_grid_thw -> [b, 3] -> [1, h, w] * frames
+        temp_frames_hw = []
+        for t, h, w in video_grid_thw:
+            repeated_row = torch.tensor([1, h.item(), w.item()]).unsqueeze(0).repeat(t, 1)
+            temp_frames_hw.append(repeated_row)
+        flattened_video_grid_thw = torch.cat(temp_frames_hw, dim=0)
+        video_embeds = self.visual(pixel_values_videos, grid_thw=flattened_video_grid_thw)
         split_sizes = (video_grid_thw.prod(-1) // self.visual.spatial_merge_size**2).tolist()
         video_embeds = torch.split(video_embeds, split_sizes)
         return video_embeds

src/transformers/models/glm4v/modular_glm4v.py

@@ -1087,6 +1087,7 @@ class Glm4vModel(Qwen2_5_VLModel):
                 device=input_ids.device,
             )
             image_index, video_index = 0, 0
+            video_group_index = 0
             attention_mask = attention_mask.to(total_input_ids.device)
             for i, input_ids in enumerate(total_input_ids):
                 input_ids = input_ids[attention_mask[i] == 1]
@@ -1116,7 +1117,6 @@ class Glm4vModel(Qwen2_5_VLModel):
 
                 llm_pos_ids_list = []
                 video_frame_num = 1
-
                 for modality_type, start_idx, end_idx in input_type_group:
                     st_idx = llm_pos_ids_list[-1].max() + 1 if len(llm_pos_ids_list) > 0 else 0
 
@@ -1160,7 +1160,11 @@ class Glm4vModel(Qwen2_5_VLModel):
                             w_index = torch.arange(llm_grid_w).view(1, 1, -1).expand(1, llm_grid_h, -1).flatten()
                             llm_pos_ids_list.append(torch.stack([t_index, h_index, w_index]) + st_idx)
 
-                        video_index += 1
+                        video_group_index += 1
+
+                        if video_group_index >= video_grid_thw[video_index][0]:
+                            video_index += 1
+                            video_group_index = 0
 
                         video_frame_num += 1
 
@@ -1196,6 +1200,30 @@ class Glm4vModel(Qwen2_5_VLModel):
 
             return position_ids, mrope_position_deltas
 
+    def get_video_features(
+        self, pixel_values_videos: torch.FloatTensor, video_grid_thw: Optional[torch.LongTensor] = None
+    ):
+        """
+        Encodes videos into continuous embeddings that can be forwarded to the language model.
+
+        Args:
+            pixel_values_videos (`torch.FloatTensor` of shape `(batch_size, num_channels, image_size, image_size)`):
+                The tensors corresponding to the input videos.
+            video_grid_thw (`torch.LongTensor` of shape `(num_videos, 3)`, *optional*):
+                The temporal, height and width of feature shape of each video in LLM.
+        """
+        pixel_values_videos = pixel_values_videos.type(self.visual.dtype)
+        # reshape video_grid_thw -> [b, 3] -> [1, h, w] * frames
+        temp_frames_hw = []
+        for t, h, w in video_grid_thw:
+            repeated_row = torch.tensor([1, h.item(), w.item()]).unsqueeze(0).repeat(t, 1)
+            temp_frames_hw.append(repeated_row)
+        flattened_video_grid_thw = torch.cat(temp_frames_hw, dim=0)
+        video_embeds = self.visual(pixel_values_videos, grid_thw=flattened_video_grid_thw)
+        split_sizes = (video_grid_thw.prod(-1) // self.visual.spatial_merge_size**2).tolist()
+        video_embeds = torch.split(video_embeds, split_sizes)
+        return video_embeds
+
     @auto_docstring
     @can_return_tuple
     def forward(
@@ -1687,32 +1715,38 @@ class Glm4vProcessor(Qwen2_5_VLProcessor):
             video_index = 0
             for i in range(len(text)):
                 while self.video_token in text[i]:
-                    num_frames = len(video_grid_thw)
+                    num_frames = video_grid_thw[video_index][0]
                     video_structure = ""
 
                     if hasattr(timestamps, "tolist"):
                         timestamps_list = timestamps.tolist()[0]
                     else:
                         timestamps_list = timestamps[0] if isinstance(timestamps[0], list) else timestamps
+
                     unique_timestamps = []
                     for idx in range(0, len(timestamps_list)):
                         unique_timestamps.append(timestamps_list[idx])
+
                     selected_timestamps = unique_timestamps[:num_frames]
                     while len(selected_timestamps) < num_frames:
                         selected_timestamps.append(selected_timestamps[-1] if selected_timestamps else 0)
+
                     for frame_idx in range(num_frames):
                         timestamp_sec = selected_timestamps[frame_idx]
                         frame_structure = f"<|begin_of_image|>{self.image_token}<|end_of_image|>{timestamp_sec}"
                         video_structure += frame_structure
+
                     text[i] = text[i].replace(self.video_token, video_structure, 1)
+                    num_image_tokens = (
+                        video_grid_thw[video_index].prod() // merge_length // video_grid_thw[video_index][0]
+                    )
+                    for frame_idx in range(num_frames):
+                        if self.image_token in text[i]:
+                            text[i] = text[i].replace(self.image_token, "<|placeholder|>" * num_image_tokens, 1)
+
                     video_index += 1
 
-                for frame_idx in range(len(video_grid_thw)):
-                    if self.image_token in text[i]:
-                        num_image_tokens = video_grid_thw[frame_idx].prod() // merge_length
-                        text[i] = text[i].replace(self.image_token, "<|placeholder|>" * num_image_tokens, 1)
                 text[i] = text[i].replace("<|placeholder|>", self.image_token)
-
         return_tensors = output_kwargs["text_kwargs"].pop("return_tensors", None)
         text_inputs = self.tokenizer(text, **output_kwargs["text_kwargs"])
         self._check_special_mm_tokens(text, text_inputs, modalities=["image", "video"])

src/transformers/models/glm4v/processing_glm4v.py

@@ -167,32 +167,38 @@ class Glm4vProcessor(ProcessorMixin):
             video_index = 0
             for i in range(len(text)):
                 while self.video_token in text[i]:
-                    num_frames = len(video_grid_thw)
+                    num_frames = video_grid_thw[video_index][0]
                     video_structure = ""
 
                     if hasattr(timestamps, "tolist"):
                         timestamps_list = timestamps.tolist()[0]
                     else:
                         timestamps_list = timestamps[0] if isinstance(timestamps[0], list) else timestamps
+
                     unique_timestamps = []
                     for idx in range(0, len(timestamps_list)):
                         unique_timestamps.append(timestamps_list[idx])
+
                     selected_timestamps = unique_timestamps[:num_frames]
                     while len(selected_timestamps) < num_frames:
                         selected_timestamps.append(selected_timestamps[-1] if selected_timestamps else 0)
+
                     for frame_idx in range(num_frames):
                         timestamp_sec = selected_timestamps[frame_idx]
                         frame_structure = f"<|begin_of_image|>{self.image_token}<|end_of_image|>{timestamp_sec}"
                         video_structure += frame_structure
+
                     text[i] = text[i].replace(self.video_token, video_structure, 1)
+                    num_image_tokens = (
+                        video_grid_thw[video_index].prod() // merge_length // video_grid_thw[video_index][0]
+                    )
+                    for frame_idx in range(num_frames):
+                        if self.image_token in text[i]:
+                            text[i] = text[i].replace(self.image_token, "<|placeholder|>" * num_image_tokens, 1)
+
                     video_index += 1
 
-                for frame_idx in range(len(video_grid_thw)):
-                    if self.image_token in text[i]:
-                        num_image_tokens = video_grid_thw[frame_idx].prod() // merge_length
-                        text[i] = text[i].replace(self.image_token, "<|placeholder|>" * num_image_tokens, 1)
                 text[i] = text[i].replace("<|placeholder|>", self.image_token)
-
         return_tensors = output_kwargs["text_kwargs"].pop("return_tensors", None)
         text_inputs = self.tokenizer(text, **output_kwargs["text_kwargs"])
         self._check_special_mm_tokens(text, text_inputs, modalities=["image", "video"])

src/transformers/models/glm4v/video_processing_glm4v.py

@@ -246,10 +246,6 @@ class Glm4vVideoProcessor(BaseVideoProcessor):
         processed_grids = reorder_videos(processed_grids, grouped_videos_index)
         pixel_values_videos = torch.cat(processed_videos, dim=0)
         video_grid_thw = torch.tensor(processed_grids)
-        total_frames = video_grid_thw[0][0].item()
-        h = video_grid_thw[0][1].item()
-        w = video_grid_thw[0][2].item()
-        video_grid_thw = [[1, h, w] for _ in range(total_frames)]
         data = {
             "pixel_values_videos": pixel_values_videos,
             "video_grid_thw": video_grid_thw,