diff --git a/benchmark/mmmu/README.md b/benchmark/mmmu/README.md index 62a5a2f12..80db21921 100644 --- a/benchmark/mmmu/README.md +++ b/benchmark/mmmu/README.md @@ -27,6 +27,18 @@ python -m sglang.launch_server --model-path microsoft/Phi-4-multimodal-instruct python -m benchmark/mmmu/bench_sglang.py --concurrency 8 --lora-path vision ``` +You can use `--response-answer-regex` to specify how to extract the answer from the response string. E.g., +``` +python3 -m sglang.launch_server --model-path zai-org/GLM-4.1V-9B-Thinking --reasoning-parser glm45 + +python3 bench_sglang.py --response-answer-regex "<\|begin_of_box\|>(.*)<\|end_of_box\|>" --concurrency 64 +``` + +You can use `--extra-request-body` to specify additional OpenAI request parameters. E.g., +``` +python3 bench_sglang.py --extra-request-body '{"max_new_tokens": 128, "temperature": 0.01}' +``` + ### Evaluate hf ``` diff --git a/benchmark/mmmu/bench_sglang.py b/benchmark/mmmu/bench_sglang.py index 524beb7bc..372bfeed8 100644 --- a/benchmark/mmmu/bench_sglang.py +++ b/benchmark/mmmu/bench_sglang.py @@ -11,6 +11,7 @@ The eval output will be logged import argparse import asyncio +import re import sys import time import traceback @@ -145,7 +146,17 @@ async def eval_mmmu(args) -> None: _, response = await process_sample( client, sample, sampling_params, lora_path ) - process_result(response, sample, answer_dict, out_samples) + answer = ( + re.search(args.response_answer_regex, response) + if response is not None + else None + ) + process_result( + answer.group(1) if answer else response, + sample, + answer_dict, + out_samples, + ) else: semaphore = asyncio.Semaphore(args.concurrency) tasks = [ @@ -157,7 +168,17 @@ async def eval_mmmu(args) -> None: for coro in tqdm(asyncio.as_completed(tasks), total=len(tasks)): sample, response = await coro - process_result(response, sample, answer_dict, out_samples) + answer = ( + re.search(args.response_answer_regex, response) + if response is not None + else None + ) + process_result( + answer.group(1) if answer else response, + sample, + answer_dict, + out_samples, + ) if args.profile: print("Stopping profiler...") diff --git a/benchmark/mmmu/eval_utils.py b/benchmark/mmmu/eval_utils.py index 2ec669155..83f6dd7fb 100644 --- a/benchmark/mmmu/eval_utils.py +++ b/benchmark/mmmu/eval_utils.py @@ -35,6 +35,7 @@ class EvalArgs: profile: bool = False profile_number: int = 5 concurrency: int = 1 + response_answer_regex: str = "(.*)" lora_path: Optional[str] = None @staticmethod @@ -92,6 +93,12 @@ class EvalArgs: default=EvalArgs.concurrency, help="Number of concurrent requests to make during evaluation. Default is 1, which means no concurrency.", ) + parser.add_argument( + "--response-answer-regex", + type=str, + default=EvalArgs.response_answer_regex, + help="Specific regex to capture the answer from the response, string", + ) parser.add_argument( "--lora-path", type=str, diff --git a/docs/supported_models/multimodal_language_models.md b/docs/supported_models/multimodal_language_models.md index 66de3d8a1..a2adf99cb 100644 --- a/docs/supported_models/multimodal_language_models.md +++ b/docs/supported_models/multimodal_language_models.md @@ -39,3 +39,4 @@ in the GitHub search bar. | **Mistral-Small-3.1-24B** | `mistralai/Mistral-Small-3.1-24B-Instruct-2503` | `mistral` | Mistral 3.1 is a multimodal model that can generate text from text or images input. It also supports tool calling and structured output. | | **Phi-4-multimodal-instruct** | `microsoft/Phi-4-multimodal-instruct` | `phi-4-mm` | Phi-4-multimodal-instruct is the multimodal variant of the Phi-4-mini model, enhanced with LoRA for improved multimodal capabilities. It supports text, vision and audio modalities in SGLang. | | **MiMo-VL** (7B) | `XiaomiMiMo/MiMo-VL-7B-RL` | `mimo-vl` | Xiaomi's compact yet powerful vision-language model featuring a native resolution ViT encoder for fine-grained visual details, an MLP projector for cross-modal alignment, and the MiMo-7B language model optimized for complex reasoning tasks. | +| **GLM-4.5V** (106B) / **GLM-4.1V**(9B) | `zai-org/GLM-4.5V` | `glm-4v` | GLM-4.5V and GLM-4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning | diff --git a/python/sglang/lang/chat_template.py b/python/sglang/lang/chat_template.py index ef348d27e..80ea6d963 100644 --- a/python/sglang/lang/chat_template.py +++ b/python/sglang/lang/chat_template.py @@ -505,6 +505,22 @@ register_chat_template( ) ) +# Reference: https://huggingface.co/docs/transformers/main/model_doc/glm4_v#usage-example +register_chat_template( + ChatTemplate( + name="glm-4v", + default_system_prompt=None, + role_prefix_and_suffix={ + "system": ("<|system|>\n", "\n"), + "user": ("<|user|>\n", "\n"), + "assistant": ("<|assistant|>\n", "\n"), + }, + style=ChatTemplateStyle.PLAIN, + stop_str=["<|user|>", "<|endoftext|>", "<|observation|>"], + image_token="<|image|>", + ) +) + @register_chat_template_matching_function def match_deepseek(model_path: str): @@ -562,6 +578,8 @@ def match_chat_ml(model_path: str): return "chatml" if re.search(r"qwen.*vl", model_path, re.IGNORECASE): return "qwen2-vl" + if re.search(r"glm[-_]?4(\.\d+)?v", model_path, re.IGNORECASE): + return "glm-4v" if re.search(r"qwen.*(chat|instruct)", model_path, re.IGNORECASE) and not re.search( r"llava", model_path, re.IGNORECASE ): diff --git a/python/sglang/srt/configs/model_config.py b/python/sglang/srt/configs/model_config.py index da70ec740..1b96ae678 100644 --- a/python/sglang/srt/configs/model_config.py +++ b/python/sglang/srt/configs/model_config.py @@ -659,6 +659,8 @@ multimodal_model_archs = [ "DeepseekVL2ForCausalLM", "Gemma3ForConditionalGeneration", "Gemma3nForConditionalGeneration", + "Glm4vForConditionalGeneration", + "Glm4vMoeForConditionalGeneration", "Grok1VForCausalLM", "Grok1AForCausalLM", "LlavaLlamaForCausalLM", diff --git a/python/sglang/srt/function_call/ebnf_composer.py b/python/sglang/srt/function_call/ebnf_composer.py index 1db7da6d8..d41968ea7 100644 --- a/python/sglang/srt/function_call/ebnf_composer.py +++ b/python/sglang/srt/function_call/ebnf_composer.py @@ -316,6 +316,7 @@ class EBNFComposer: combined_args = "".join(rule_parts) arguments_rule = args_template.format(arg_rules=combined_args) + arguments_rule = arguments_rule or '""' # Add the function call rule and its arguments rule ebnf_lines.append( diff --git a/python/sglang/srt/function_call/glm4_moe_detector.py b/python/sglang/srt/function_call/glm4_moe_detector.py index 705bbcdb3..39822fb19 100644 --- a/python/sglang/srt/function_call/glm4_moe_detector.py +++ b/python/sglang/srt/function_call/glm4_moe_detector.py @@ -158,7 +158,7 @@ class Glm4MoeDetector(BaseFormatDetector): individual_call_end_token=self.eot_token, tool_call_separator="\\n", function_format="xml", - call_rule_fmt='"{name}" "\\n" {arguments_rule} "\\n"', + call_rule_fmt='"{name}" "\\n" ( {arguments_rule} "\\n" )?', key_value_rule_fmt='"{key}" "\\n" "" {valrule} ""', key_value_separator="\\n", ) diff --git a/python/sglang/srt/jinja_template_utils.py b/python/sglang/srt/jinja_template_utils.py index e23aa9226..be7d44097 100644 --- a/python/sglang/srt/jinja_template_utils.py +++ b/python/sglang/srt/jinja_template_utils.py @@ -102,6 +102,12 @@ def detect_jinja_template_content_format(chat_template: str) -> str: if _is_var_or_elems_access(loop_iter, "message", "content"): return "openai" # Found content iteration → openai format + # Also check for patterns like: {%- for item in msg.content -%} or {%- for item in m.content -%} + if _is_var_or_elems_access( + loop_iter, "msg", "content" + ) or _is_var_or_elems_access(loop_iter, "m", "content"): + return "openai" # Found content iteration → openai format (glm4v) + return "string" # No content loops found → string format except Exception as e: logger.debug(f"Error when parsing AST of Jinja template: {e}") diff --git a/python/sglang/srt/layers/rotary_embedding.py b/python/sglang/srt/layers/rotary_embedding.py index 252362201..c583a5d23 100644 --- a/python/sglang/srt/layers/rotary_embedding.py +++ b/python/sglang/srt/layers/rotary_embedding.py @@ -1,6 +1,7 @@ # Adapted from https://raw.githubusercontent.com/vllm-project/vllm/refs/tags/v0.6.6.post1/vllm/model_executor/layers/rotary_embedding.py """Rotary Positional Embeddings.""" +import itertools import math from typing import Any, Dict, List, Optional, Tuple, Union @@ -946,7 +947,37 @@ class MRotaryEmbedding(RotaryEmbedding): self.mrope_section = mrope_section if self.mrope_section: - assert sum(self.mrope_section) == rotary_dim // 2 + expected_sum = rotary_dim // 2 + actual_sum = sum(self.mrope_section) + if actual_sum != expected_sum: + print( + f"MRoPE section sum mismatch: expected {expected_sum}, got {actual_sum}. " + f"Adjusting mrope_section to match rotary_dim // 2 = {expected_sum}" + ) + # Auto-correct by scaling the mrope_section proportionally + if actual_sum > 0: + scale_factor = expected_sum / actual_sum + self.mrope_section = [ + max(1, int(section * scale_factor)) + for section in self.mrope_section + ] + # Ensure the sum exactly matches by adjusting the last element + current_sum = sum(self.mrope_section) + if current_sum != expected_sum: + self.mrope_section[-1] += expected_sum - current_sum + else: + # If all sections are 0, create a default distribution + self.mrope_section = [ + expected_sum // len(self.mrope_section) + ] * len(self.mrope_section) + # Handle remainder + remainder = expected_sum % len(self.mrope_section) + for i in range(remainder): + self.mrope_section[i] += 1 + + print( + f"Corrected mrope_section: {self.mrope_section} (sum={sum(self.mrope_section)})" + ) def forward( self, @@ -1153,6 +1184,204 @@ class MRotaryEmbedding(RotaryEmbedding): mrope_position_deltas = max_position_ids + 1 - s return position_ids, mrope_position_deltas + # Adapted from https://github.com/vllm-project/vllm/blob/3779eb8c81449b924a23457fc77e45a0e6171178/vllm/model_executor/layers/rotary_embedding.py#L1120 + @staticmethod + def get_rope_index_glm4v( + input_ids: torch.Tensor, + hf_config: Any, + image_grid_thw: Union[list[list[int]], torch.Tensor], + video_grid_thw: Union[list[list[int]], torch.Tensor], + attention_mask: torch.Tensor, + **kwargs, + ) -> tuple[torch.Tensor, torch.Tensor]: + """Get mrope input positions and delta value for GLM4V.""" + image_token_id = hf_config.image_token_id + video_start_token_id = hf_config.video_start_token_id + video_end_token_id = hf_config.video_end_token_id + spatial_merge_size = hf_config.vision_config.spatial_merge_size + + mrope_position_deltas = [] + if input_ids is not None and ( + image_grid_thw is not None or video_grid_thw is not None + ): + total_input_ids = input_ids + if attention_mask is None: + attention_mask = torch.ones_like(total_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 + 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] + input_tokens = input_ids.tolist() + + input_token_type = [] + video_check_flg = False + for token in input_tokens: + if token == video_start_token_id: + video_check_flg = True + elif token == video_end_token_id: + video_check_flg = False + + if token == image_token_id and not video_check_flg: + input_token_type.append("image") + elif token == image_token_id and video_check_flg: + input_token_type.append("video") + else: + input_token_type.append("text") + + input_type_group = [] + for key, group in itertools.groupby( + enumerate(input_token_type), lambda x: x[1] + ): + group = list(group) + start_index = group[0][0] + end_index = group[-1][0] + 1 + input_type_group.append((key, start_index, end_index)) + + 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 + ) + + if modality_type == "image": + t, h, w = ( + image_grid_thw[image_index][0], + image_grid_thw[image_index][1], + image_grid_thw[image_index][2], + ) + llm_grid_t, llm_grid_h, llm_grid_w = ( + t.item(), + h.item() // spatial_merge_size, + w.item() // spatial_merge_size, + ) + + 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]) + st_idx + ) + + image_index += 1 + video_frame_num = 1 + + elif modality_type == "video": + t, h, w = ( + video_frame_num, + video_grid_thw[video_index][1], + video_grid_thw[video_index][2], + ) + + llm_grid_t, llm_grid_h, llm_grid_w = ( + t, + h.item() // spatial_merge_size, + w.item() // spatial_merge_size, + ) + + for t_idx in range(llm_grid_t): + t_index = ( + torch.tensor(t_idx) + .view(-1, 1) + .expand(-1, llm_grid_h * llm_grid_w) + .flatten() + ) + + h_index = ( + torch.arange(llm_grid_h) + .view(1, -1, 1) + .expand(1, -1, llm_grid_w) + .flatten() + ) + 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_group_index += 1 + + if video_group_index >= video_grid_thw[video_index][0]: + video_index += 1 + video_group_index = 0 + + video_frame_num += 1 + + else: + text_len = end_idx - start_idx + llm_pos_ids_list.append( + torch.arange(text_len).view(1, -1).expand(3, -1) + st_idx + ) + + video_frame_num = 1 + + 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) + return position_ids, mrope_position_deltas + else: + if attention_mask is not None: + position_ids = attention_mask.long().cumsum(-1) - 1 + position_ids.masked_fill_(attention_mask == 0, 1) + position_ids = ( + position_ids.unsqueeze(0) + .expand(3, -1, -1) + .to(attention_mask.device) + ) + max_position_ids = position_ids.max(0, keepdim=False)[0].max( + -1, keepdim=True + )[0] + mrope_position_deltas = max_position_ids + 1 - attention_mask.shape[-1] + else: + position_ids = ( + torch.arange(input_ids.shape[1], device=input_ids.device) + .view(1, 1, -1) + .expand(3, input_ids.shape[0], -1) + ) + mrope_position_deltas = torch.zeros( + [input_ids.shape[0], 1], + device=input_ids.device, + dtype=input_ids.dtype, + ) + + return position_ids, mrope_position_deltas + @staticmethod def get_next_input_positions( mrope_position_delta: int, diff --git a/python/sglang/srt/models/glm4.py b/python/sglang/srt/models/glm4.py index 0ef11d6d0..7357e5f82 100644 --- a/python/sglang/srt/models/glm4.py +++ b/python/sglang/srt/models/glm4.py @@ -218,6 +218,12 @@ class Glm4Model(nn.Module): self.norm = RMSNorm(config.hidden_size, eps=config.rms_norm_eps) + def get_input_embeddings(self) -> nn.Embedding: + return self.embed_tokens + + def dtype(self) -> torch.dtype: + return next(self.parameters()).dtype + @torch.no_grad() def forward( self, diff --git a/python/sglang/srt/models/glm4v.py b/python/sglang/srt/models/glm4v.py new file mode 100644 index 000000000..fbd757849 --- /dev/null +++ b/python/sglang/srt/models/glm4v.py @@ -0,0 +1,589 @@ +import logging +from functools import lru_cache, partial +from typing import Iterable, List, Optional, Tuple + +import torch +import torch.nn as nn +import torch.nn.functional as F +from transformers.models.glm4v.configuration_glm4v import Glm4vConfig, Glm4vVisionConfig + +from sglang.srt.hf_transformers_utils import get_processor +from sglang.srt.layers.activation import SiluAndMul +from sglang.srt.layers.layernorm import RMSNorm +from sglang.srt.layers.linear import ( + ColumnParallelLinear, + MergedColumnParallelLinear, + RowParallelLinear, +) +from sglang.srt.layers.logits_processor import LogitsProcessor +from sglang.srt.layers.pooler import Pooler, PoolingType +from sglang.srt.layers.quantization.base_config import QuantizationConfig +from sglang.srt.layers.vocab_parallel_embedding import ParallelLMHead +from sglang.srt.managers.schedule_batch import MultimodalDataItem +from sglang.srt.model_loader.weight_utils import default_weight_loader +from sglang.srt.models.glm4 import Glm4Model +from sglang.srt.models.qwen2_5_vl import ( + Qwen2_5_VisionBlock, + Qwen2_5_VLForConditionalGeneration, +) +from sglang.srt.utils import add_prefix + +logger = logging.getLogger(__name__) + +cached_get_processor = lru_cache(get_processor) + + +class Glm4vRMSNorm(RMSNorm): + def forward(self, x: torch.Tensor) -> torch.Tensor: + original_shape = x.shape + x_2d = x.contiguous().reshape(-1, original_shape[-1]) + x_2d = super().forward(x_2d) + x = x_2d.reshape(original_shape) + return x + + +class Glm4vVisionMLP(nn.Module): + def __init__( + self, + in_features: int, + hidden_features: int, + bias: bool = False, + quant_config: Optional[QuantizationConfig] = None, + prefix: str = "", + ): + super().__init__() + self.gate_up_proj = MergedColumnParallelLinear( + input_size=in_features, + output_sizes=[hidden_features] * 2, + bias=bias, + quant_config=quant_config, + prefix=add_prefix("gate_up_proj", prefix), + ) + self.down_proj = RowParallelLinear( + hidden_features, + in_features, + bias=bias, + quant_config=quant_config, + prefix=add_prefix("down_proj", prefix), + ) + self.act_fn = SiluAndMul() + + def forward(self, x: torch.Tensor): + gate_up, _ = self.gate_up_proj(x) + x = self.act_fn(gate_up) + x, _ = self.down_proj(x) + return x + + +class Glm4vVisionBlock(Qwen2_5_VisionBlock): + def __init__( + self, + config: Glm4vVisionConfig, + norm_layer: Optional[nn.Module] = None, + quant_config: Optional[QuantizationConfig] = None, + prefix: str = "", + ) -> None: + super().__init__( + dim=config.hidden_size, + intermediate_dim=config.out_hidden_size, + num_heads=config.num_heads, + hidden_act=config.hidden_act, + norm_layer=norm_layer, + quant_config=quant_config, + prefix=prefix, + ) + self.norm1 = Glm4vRMSNorm(config.hidden_size, eps=config.rms_norm_eps) + self.norm2 = Glm4vRMSNorm(config.hidden_size, eps=config.rms_norm_eps) + + self.mlp = Glm4vVisionMLP( + config.hidden_size, + config.out_hidden_size, + bias=False, + quant_config=quant_config, + prefix=add_prefix("mlp", prefix), + ) + + +class Glm4vVisionPatchEmbed(nn.Module): + def __init__( + self, + patch_size: int = 14, + temporal_patch_size: int = 2, + in_channels: int = 3, + hidden_size: int = 1536, + ) -> None: + super().__init__() + self.patch_size = patch_size + self.temporal_patch_size = temporal_patch_size + self.hidden_size = hidden_size + self.in_channels = in_channels + + kernel_size = (temporal_patch_size, patch_size, patch_size) + self.proj = nn.Conv3d( + in_channels, + hidden_size, + kernel_size=kernel_size, + stride=kernel_size, + bias=True, + ) + + def forward(self, x: torch.Tensor) -> torch.Tensor: + x = x.view( + -1, + self.in_channels, + self.temporal_patch_size, + self.patch_size, + self.patch_size, + ) + x = self.proj(x).view(-1, self.hidden_size) + return x + + +class Glm4vPatchMerger(nn.Module): + def __init__( + self, + d_model: int, + context_dim: int, + quant_config: Optional[QuantizationConfig] = None, + bias: bool = False, + prefix: str = "", + ) -> None: + super().__init__() + self.hidden_size = d_model + self.proj = ColumnParallelLinear( + self.hidden_size, + self.hidden_size, + bias=bias, + quant_config=quant_config, + prefix=add_prefix("proj", prefix), + gather_output=True, + ) + self.post_projection_norm = nn.LayerNorm(self.hidden_size) + self.gate_up_proj = MergedColumnParallelLinear( + input_size=self.hidden_size, + output_sizes=[context_dim] * 2, + bias=bias, + quant_config=quant_config, + prefix=add_prefix("gate_up_proj", prefix), + ) + self.down_proj = RowParallelLinear( + context_dim, + self.hidden_size, + bias=bias, + quant_config=quant_config, + prefix=add_prefix("down_proj", prefix), + ) + self.extra_activation_func = nn.GELU() + + def forward(self, x: torch.Tensor): + x, _ = self.proj(x) + x = self.extra_activation_func(self.post_projection_norm(x)) + gate_up, _ = self.gate_up_proj(x) + gate, up = gate_up.chunk(2, dim=-1) + x = F.silu(gate) * up + x, _ = self.down_proj(x) + return x + + +class Glm4vVisionEmbeddings(nn.Module): + def __init__(self, config: Glm4vVisionConfig): + super().__init__() + self.config = config + self.embed_dim = config.hidden_size + self.image_size = config.image_size + self.patch_size = config.patch_size + + self.num_patches = (self.image_size // self.patch_size) ** 2 + self.num_positions = self.num_patches + self.position_embedding = nn.Embedding(self.num_positions, self.embed_dim) + self.register_buffer( + "position_ids", + torch.arange(self.num_positions).expand((1, -1)), + persistent=False, + ) + + def forward( + self, embeddings, lengths, image_shapes, h_coords, w_coords + ) -> torch.Tensor: + pos_embed_weight = self.position_embedding.weight + hidden_size = pos_embed_weight.shape[1] + total_seq = h_coords.shape[0] + device = pos_embed_weight.device + + # Move coordinates to correct device + h_coords, w_coords = h_coords.to(device), w_coords.to(device) + + # Handle empty sequence case + if total_seq == 0: + adapted_pos_embed = torch.empty( + 0, hidden_size, device=device, dtype=pos_embed_weight.dtype + ) + else: + # Convert inputs to tensors if needed + if isinstance(lengths, list): + lengths = torch.tensor(lengths, device=device, dtype=torch.long) + if not isinstance(image_shapes, torch.Tensor): + image_shapes = torch.tensor( + image_shapes, device=device, dtype=torch.long + ) + + # Prepare 2D position embedding + orig_size_sq = pos_embed_weight.shape[0] + orig_size = int(orig_size_sq**0.5) + pos_embed_2d = ( + pos_embed_weight.view(orig_size, orig_size, hidden_size) + .permute(2, 0, 1) + .unsqueeze(0) + .to(device=device, dtype=torch.float32) + ) + + # Calculate target dimensions for each patch + target_h = torch.cat( + [image_shapes[i, 1].repeat(lengths[i]) for i in range(len(lengths))] + ).to(device=device, dtype=torch.float32) + target_w = torch.cat( + [image_shapes[i, 2].repeat(lengths[i]) for i in range(len(lengths))] + ).to(device=device, dtype=torch.float32) + + # Normalize coordinates to [-1, 1] range for grid_sample + h_coords = h_coords.to(device=device, dtype=torch.float32) + w_coords = w_coords.to(device=device, dtype=torch.float32) + norm_w = ((w_coords + 0.5) / target_w) * 2 - 1 + norm_h = ((h_coords + 0.5) / target_h) * 2 - 1 + + # Create sampling grid + grid = torch.stack((norm_w, norm_h), dim=-1).unsqueeze(0).unsqueeze(2) + + # Perform bicubic interpolation + interpolated_embed_fp32 = F.grid_sample( + pos_embed_2d, + grid, + mode="bicubic", + align_corners=False, + padding_mode="border", + ) + + # Reshape and convert back to original dtype + adapted_pos_embed_fp32 = ( + interpolated_embed_fp32.squeeze(0).squeeze(-1).permute(1, 0) + ) + adapted_pos_embed = adapted_pos_embed_fp32.to(pos_embed_weight.dtype).to( + embeddings.device + ) + + # Add adapted position encoding to embeddings + embeddings = embeddings + adapted_pos_embed + return embeddings + + +class Glm4vVisionRotaryEmbedding(nn.Module): + def __init__(self, dim: int, theta: float = 10000.0) -> None: + super().__init__() + self.dim = dim + self.theta = theta + inv_freq = 1.0 / (theta ** (torch.arange(0, dim, 2, dtype=torch.float) / dim)) + self.register_buffer("inv_freq", inv_freq, persistent=False) + self._seq_len_cached = 0 + self._freqs_cached = None + + def update_freqs_cache(self, seqlen: int) -> None: + if seqlen > self._seq_len_cached: + seqlen *= 2 + self._seq_len_cached = seqlen + self.inv_freq = 1.0 / ( + self.theta + ** ( + torch.arange( + 0, + self.dim, + 2, + dtype=torch.float, + device=self.inv_freq.device, + ) + / self.dim + ) + ) + seq = torch.arange( + seqlen, device=self.inv_freq.device, dtype=self.inv_freq.dtype + ) + freqs = torch.outer(seq, self.inv_freq) + self._freqs_cached = freqs + + def forward(self, seqlen: int) -> torch.Tensor: + self.update_freqs_cache(seqlen) + return self._freqs_cached[:seqlen] + + +class Glm4vVisionModel(nn.Module): + def __init__( + self, + vision_config: Glm4vVisionConfig, + norm_eps: float = 1e-6, + quant_config: Optional[QuantizationConfig] = None, + prefix: str = "", + ) -> None: + super().__init__() + + patch_size = vision_config.patch_size + temporal_patch_size = vision_config.temporal_patch_size + in_channels = vision_config.in_channels + depth = vision_config.depth + self.hidden_size = vision_config.hidden_size + self.num_heads = vision_config.num_heads + + self.patch_size = vision_config.patch_size + self.spatial_merge_size = vision_config.spatial_merge_size + self.out_hidden_size = vision_config.out_hidden_size + + self.patch_embed = Glm4vVisionPatchEmbed( + patch_size=patch_size, + temporal_patch_size=temporal_patch_size, + in_channels=in_channels, + hidden_size=self.hidden_size, + ) + + norm_layer = partial(Glm4vRMSNorm, eps=norm_eps) + head_dim = self.hidden_size // self.num_heads + self.rotary_pos_emb = Glm4vVisionRotaryEmbedding(head_dim // 2) + + self.blocks = nn.ModuleList( + [ + Glm4vVisionBlock( + config=vision_config, + norm_layer=norm_layer, + quant_config=quant_config, + prefix=add_prefix(f"blocks.{layer_idx}", prefix), + ) + for layer_idx in range(depth) + ] + ) + + self.merger = Glm4vPatchMerger( + d_model=vision_config.out_hidden_size, + context_dim=vision_config.intermediate_size, + quant_config=quant_config, + bias=False, + prefix=add_prefix("merger", prefix), + ) + + self.embeddings = Glm4vVisionEmbeddings(vision_config) + + self.post_conv_layernorm = Glm4vRMSNorm( + vision_config.hidden_size, eps=vision_config.rms_norm_eps + ) + self.downsample = nn.Conv2d( + in_channels=vision_config.hidden_size, + out_channels=vision_config.out_hidden_size, + kernel_size=vision_config.spatial_merge_size, + stride=vision_config.spatial_merge_size, + ) + self.post_layernorm = Glm4vRMSNorm( + vision_config.hidden_size, eps=vision_config.rms_norm_eps + ) + + @property + def dtype(self) -> torch.dtype: + return self.patch_embed.proj.weight.dtype + + @property + def device(self) -> torch.device: + return self.patch_embed.proj.weight.device + + def rot_pos_emb(self, grid_thw: torch.Tensor) -> torch.Tensor: + pos_ids = [] + for t, h, w in grid_thw: + hpos_ids = torch.arange(h).unsqueeze(1).expand(-1, w) + wpos_ids = torch.arange(w).unsqueeze(0).expand(h, -1) + hpos_ids = ( + hpos_ids.reshape( + h // self.spatial_merge_size, + self.spatial_merge_size, + w // self.spatial_merge_size, + self.spatial_merge_size, + ) + .permute(0, 2, 1, 3) + .flatten() + ) + wpos_ids = ( + wpos_ids.reshape( + h // self.spatial_merge_size, + self.spatial_merge_size, + w // self.spatial_merge_size, + self.spatial_merge_size, + ) + .permute(0, 2, 1, 3) + .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() + rotary_pos_emb_full = self.rotary_pos_emb(max_grid_size) + rotary_pos_emb = rotary_pos_emb_full[pos_ids].flatten(1) + return rotary_pos_emb, pos_ids + + def forward(self, x: torch.Tensor, grid_thw: torch.Tensor) -> torch.Tensor: + # patchify + x = x.to(device=self.device, dtype=self.dtype) + x = self.patch_embed(x) + x = self.post_conv_layernorm(x) + + # compute position embedding + rotary_pos_emb, image_type_ids = self.rot_pos_emb(grid_thw) + # compute cu_seqlens + 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), "constant", 0) + + seqlens = (cu_seqlens[1:] - cu_seqlens[:-1]).tolist() + x = self.embeddings( + x, seqlens, grid_thw, image_type_ids[:, 0], image_type_ids[:, 1] + ) + + emb = torch.cat((rotary_pos_emb, rotary_pos_emb), dim=-1) + rotary_pos_emb_tuple = (emb.cos(), emb.sin()) + + # x.shape: (s, b, d) where b=1 for vision processing + # transformers + x = x.unsqueeze(1) + for blk in self.blocks: + x = blk(x, cu_seqlens=cu_seqlens, position_embeddings=rotary_pos_emb_tuple) + + # adapter + x = self.post_layernorm(x) + x = x.view(-1, self.spatial_merge_size, self.spatial_merge_size, x.shape[-1]) + x = x.permute(0, 3, 1, 2) + x = self.downsample(x).view(-1, self.out_hidden_size) + x = self.merger(x) + + return x + + +class Glm4vForConditionalGeneration(Qwen2_5_VLForConditionalGeneration): + def __init__( + self, + config: Glm4vConfig, + quant_config: Optional[QuantizationConfig] = None, + prefix: str = "", + ) -> None: + nn.Module.__init__(self) + + self.config = config + + self.model = Glm4Model( + config, + quant_config, + prefix=add_prefix("model", prefix), + ) + self.visual = Glm4vVisionModel( + config.vision_config, + norm_eps=getattr(config, "rms_norm_eps", 1e-5), + quant_config=quant_config, + prefix=add_prefix("visual", prefix), + ) + + if config.tie_word_embeddings: + self.lm_head = self.model.embed_tokens + else: + self.lm_head = ParallelLMHead( + config.vocab_size, + config.hidden_size, + quant_config=quant_config, + prefix=add_prefix("lm_head", prefix), + ) + + self.logits_processor = LogitsProcessor(config) + self.pooler = Pooler(pooling_type=PoolingType.LAST, normalize=True) + self.is_mrope_enabled = "mrope_section" in self.config.rope_scaling + + def get_image_feature(self, items: List[MultimodalDataItem]) -> torch.Tensor: + pixel_values = torch.cat( + [item.feature.squeeze(0) for item in items], dim=0 + ).type(self.visual.dtype) + image_grid_thw = torch.concat([item.image_grid_thw for item in items], dim=0) + # For multi-image, pixel_values is [num_of_images, L, C] shape + # assert pixel_values.dim() == 2, pixel_values.dim() + assert image_grid_thw.dim() == 2, image_grid_thw.dim() + image_embeds = self.visual(pixel_values, grid_thw=image_grid_thw) + split_sizes = ( + image_grid_thw.prod(-1) // self.visual.spatial_merge_size**2 + ).tolist() + image_embeds = torch.split(image_embeds, split_sizes) + return torch.cat(image_embeds) + + def get_video_feature(self, items: List[MultimodalDataItem]) -> torch.Tensor: + pixel_values_videos = torch.cat( + [item.feature.squeeze(0) for item in items], dim=0 + ).type(self.visual.dtype) + video_grid_thw = torch.concat([item.video_grid_thw for item in items], dim=0) + # For multi-video, pixel_values_videos is [num_of_videos, L, C] shape + # assert pixel_values_videos.dim() == 2, pixel_values_videos.dim() + assert video_grid_thw.dim() == 2, video_grid_thw.dim() + + # 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 torch.cat(video_embeds) + + def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]): + stacked_params_mapping = [ + # (param_name, shard_name, shard_id) + (".qkv_proj", ".q_proj", "q"), + (".qkv_proj", ".k_proj", "k"), + (".qkv_proj", ".v_proj", "v"), + (".gate_up_proj", ".up_proj", 1), + (".gate_up_proj", ".gate_proj", 0), + ] + params_dict = dict(self.named_parameters(remove_duplicate=False)) + for name, loaded_weight in weights: + if "language_model." in name: + name = name.replace("language_model.", "") + if "model.visual." in name: + name = name.replace("model.visual.", "visual.") + + if "rotary_emb.inv_freq" in name: + continue + + for param_name, weight_name, shard_id in stacked_params_mapping: + if weight_name not in name: + continue + name = name.replace(weight_name, param_name) + + # Skip loading extra bias for GPTQ models. + if name.endswith(".bias") and name not in params_dict: + continue + param = params_dict[name] + weight_loader = param.weight_loader + weight_loader(param, loaded_weight, shard_id) + break + else: + if "visual" in name: + # adapt to VisionAttention + name = name.replace(r"attn.qkv.", r"attn.qkv_proj.") + + try: + # Skip loading extra bias for GPTQ models. + if name.endswith(".bias") and name not in params_dict: + continue + param = params_dict[name] + except KeyError: + print(params_dict.keys()) + raise + + weight_loader = getattr(param, "weight_loader", default_weight_loader) + weight_loader(param, loaded_weight) + + +EntryClass = [Glm4vForConditionalGeneration] diff --git a/python/sglang/srt/models/glm4v_moe.py b/python/sglang/srt/models/glm4v_moe.py new file mode 100644 index 000000000..140b6e135 --- /dev/null +++ b/python/sglang/srt/models/glm4v_moe.py @@ -0,0 +1,400 @@ +import logging +from functools import lru_cache +from typing import Iterable, Optional, Tuple + +import torch +import torch.nn as nn +from transformers.models.glm4v_moe.configuration_glm4v_moe import Glm4vMoeConfig + +from sglang.srt.distributed import ( + get_moe_expert_parallel_world_size, + get_tensor_model_parallel_rank, + get_tensor_model_parallel_world_size, + parallel_state, + tensor_model_parallel_all_reduce, +) +from sglang.srt.hf_transformers_utils import get_processor +from sglang.srt.layers.dp_attention import ( + get_attention_tp_rank, + get_attention_tp_size, + get_local_attention_dp_size, +) +from sglang.srt.layers.logits_processor import LogitsProcessor +from sglang.srt.layers.moe.ep_moe.layer import get_moe_impl_class +from sglang.srt.layers.pooler import Pooler, PoolingType +from sglang.srt.layers.quantization.base_config import QuantizationConfig +from sglang.srt.layers.vocab_parallel_embedding import ParallelLMHead +from sglang.srt.managers.schedule_batch import global_server_args_dict +from sglang.srt.model_loader.weight_utils import default_weight_loader +from sglang.srt.models.glm4_moe import Glm4MoeModel +from sglang.srt.models.glm4v import Glm4vForConditionalGeneration, Glm4vVisionModel +from sglang.srt.utils import add_prefix, is_cuda, log_info_on_rank0 + +_is_cuda = is_cuda() + +logger = logging.getLogger(__name__) + +cached_get_processor = lru_cache(get_processor) + + +class Glm4vMoeForConditionalGeneration(Glm4vForConditionalGeneration): + def __init__( + self, + config: Glm4vMoeConfig, + quant_config: Optional[QuantizationConfig] = None, + prefix: str = "", + ) -> None: + nn.Module.__init__(self) + + config.moe_layer_freq = 1 + self.config = config + self.tp_size = get_tensor_model_parallel_world_size() + self.dp_size = get_local_attention_dp_size() + self.quant_config = quant_config + self.determine_num_fused_shared_experts("Glm4MoeForCausalLM") + self.num_fused_shared_experts = ( + 0 + if global_server_args_dict["disable_shared_experts_fusion"] + else config.n_shared_experts + ) + + self.model = Glm4MoeModel( + config, + quant_config, + prefix=add_prefix("language_model", prefix), + ) + self.visual = Glm4vVisionModel( + config.vision_config, + norm_eps=getattr(config, "rms_norm_eps", 1e-5), + quant_config=quant_config, + prefix=add_prefix("visual", prefix), + ) + + self.lm_head = ParallelLMHead( + config.vocab_size, + config.hidden_size, + quant_config=quant_config, + prefix=add_prefix("lm_head", prefix), + use_attn_tp_group=global_server_args_dict["enable_dp_lm_head"], + ) + self.logits_processor = LogitsProcessor(config) + self.pooler = Pooler(pooling_type=PoolingType.LAST, normalize=True) + self.is_mrope_enabled = "mrope_section" in self.config.rope_scaling + + def determine_num_fused_shared_experts( + self, architecture: str = "Glm4MoeForCausalLM" + ): + self.num_fused_shared_experts = 0 + if global_server_args_dict["disable_shared_experts_fusion"]: + return + + # Only Deepseek V3/R1 can use shared experts fusion optimization now. + disable_reason = None + if ( + not _is_cuda + or torch.cuda.get_device_capability("cuda") < (8, 0) + or self.config.architectures[0] != architecture + or self.config.n_shared_experts != 1 + ): + disable_reason = "Only GLM-4.5 on NV-platform with capability >= 80 can use shared experts fusion optimization." + elif get_moe_expert_parallel_world_size() > 1: + disable_reason = "Deepseek and GLM-4.5 can not use shared experts fusion optimization under expert parallelism." + + if disable_reason is not None: + global_server_args_dict["disable_shared_experts_fusion"] = True + self.num_fused_shared_experts = 0 + log_info_on_rank0( + logger, + f"{disable_reason} Shared experts fusion optimization is disabled.", + ) + return + + self.num_fused_shared_experts = self.config.n_shared_experts + + def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]], is_nextn=False): + + if is_nextn: + if hasattr(self.config, "num_nextn_predict_layers"): + num_nextn_layers = self.config.num_nextn_predict_layers + assert num_nextn_layers == 1, "Only 1 nextn layer is supported" + # compatible with old design + nextn_layer_id = ( + 0 + if self.config.num_hidden_layers == 1 + else self.config.num_hidden_layers + ) + else: + raise ValueError("num_nextn_predict_layers is not in the config") + + stacked_params_mapping = [ + # (param_name, shard_name, shard_id) + ("qkv_proj", "q_proj", "q"), + ("qkv_proj", "k_proj", "k"), + ("qkv_proj", "v_proj", "v"), + ("gate_up_proj", "gate_proj", 0), + ("gate_up_proj", "up_proj", 1), + ] + if self.num_fused_shared_experts > 0: + assert self.num_fused_shared_experts == 1 + weights_list = list(weights) + weights_dict = dict(weights_list) + if self.quant_config is not None: + if self.quant_config.get_name() == "w8a8_int8": + suffix_list = [ + "down_proj.weight", + "down_proj.weight_scale", + "gate_proj.weight", + "gate_proj.weight_scale", + "up_proj.weight", + "up_proj.weight_scale", + ] + elif ( + self.quant_config.get_name() == "fp8" + or self.quant_config.get_name() == "blockwise_int8" + or self.quant_config.get_name() == "compressed_tensors" + ): + suffix_list = [ + "down_proj.weight", + "down_proj.weight_scale", + "gate_proj.weight", + "gate_proj.weight_scale", + "up_proj.weight", + "up_proj.weight_scale", + ] + elif self.quant_config.get_name() == "awq": + suffix_list = [ + "down_proj.qweight", + "down_proj.qzeros", + "down_proj.scales", + "gate_proj.qweight", + "gate_proj.qzeros", + "gate_proj.scales", + "up_proj.qweight", + "up_proj.qzeros", + "up_proj.scales", + ] + elif self.quant_config.get_name() == "modelopt_fp4": + suffix_list = [ + "down_proj.weight", + "down_proj.weight_scale", + "down_proj.weight_scale_2", + "down_proj.input_scale", + "gate_proj.weight", + "gate_proj.weight_scale", + "gate_proj.weight_scale_2", + "gate_proj.input_scale", + "up_proj.weight", + "up_proj.weight_scale", + "up_proj.weight_scale_2", + "up_proj.input_scale", + ] + else: + raise ValueError( + f"Unsupported shared expert fusion for quantization: {self.quant_config.get_name()}." + ) + else: + suffix_list = [ + "down_proj.weight", + "gate_proj.weight", + "up_proj.weight", + ] + names_to_remove = [] + + moe_layers = ( + range( + self.config.first_k_dense_replace, + self.config.num_hidden_layers, + self.config.moe_layer_freq, + ) + if not is_nextn + else [nextn_layer_id] + ) + + for moe_layer in moe_layers: + for suffix in suffix_list: + shared_expert_weight_name = ( + f"model.layers.{moe_layer}.mlp.shared_experts.{suffix}" + ) + # online fp8 quantization does not load weight_scale + if shared_expert_weight_name not in weights_dict: + continue + weights_list.append( + ( + f"model.layers.{moe_layer}." + f"mlp.experts." + f"{self.config.n_routed_experts + 0}" + f".{suffix}", + weights_dict[shared_expert_weight_name], + ) + ) + names_to_remove += [shared_expert_weight_name] + weights = [w for w in weights_list if w[0] not in names_to_remove] + + # Params for weights, fp8 weight scales, fp8 activation scales + # (param_name, weight_name, expert_id, shard_id) + expert_params_mapping = get_moe_impl_class().make_expert_params_mapping( + ckpt_gate_proj_name="gate_proj", + ckpt_down_proj_name="down_proj", + ckpt_up_proj_name="up_proj", + num_experts=self.config.n_routed_experts + self.num_fused_shared_experts, + ) + + # Fuse q_a_proj and kv_a_proj_with_mqa along output dimension when q_lora_rank is not None + fuse_qkv_a_proj = hasattr(self.config, "q_lora_rank") and ( + self.config.q_lora_rank is not None + ) + cached_a_proj = {} if fuse_qkv_a_proj else None + + if is_nextn: + nextn_layer_prefix = f"model.layers.{nextn_layer_id}" + nextn_spec_weight_names = [ + "shared_head.norm", + "eh_proj", + "enorm", + "hnorm", + ] + + params_dict = dict(self.named_parameters()) + weight_names = [] + for name, loaded_weight in weights: + weight_names.append(name) + + if not is_nextn: + if hasattr(self.config, "num_nextn_predict_layers"): + num_nextn_layers = self.config.num_nextn_predict_layers + if num_nextn_layers > 0 and name.startswith("model.layers"): + name_list = name.split(".") + if ( + len(name_list) >= 3 + and int(name_list[2]) >= self.config.num_hidden_layers + ): + continue + else: + if not name.startswith(nextn_layer_prefix): + continue + + # Use shared head and embed weights from target model + if "shared_head.head" in name or "embed_tokens" in name: + continue + + is_decoder = True + # For nextn specific weights + for weight_name in nextn_spec_weight_names: + if weight_name in name: + name = name.replace(nextn_layer_prefix, "model") + is_decoder = False + break + # For decoder layer weights + if is_decoder: + name = name.replace(nextn_layer_prefix, "model.decoder") + + if "language_model." in name: + name = name.replace("language_model.", "") + if "model.visual." in name: + name = name.replace("model.visual.", "visual.") + if "rotary_emb.inv_freq" in name: + continue + for param_name, weight_name, shard_id in stacked_params_mapping: + # Skip non-stacked layers and experts (experts handled below). + if weight_name not in name: + continue + # We have mlp.experts[0].gate_proj in the checkpoint. + # Since we handle the experts below in expert_params_mapping, + # we need to skip here BEFORE we update the name, otherwise + # name will be updated to mlp.experts[0].gate_up_proj, which + # will then be updated below in expert_params_mapping + # for mlp.experts[0].gate_gate_up_proj, which breaks load. + if ("mlp.experts." in name) and name not in params_dict: + continue + name = name.replace(weight_name, param_name) + # Skip loading extra bias for GPTQ models. + if name.endswith(".bias") and name not in params_dict: + continue + param = params_dict[name] + + weight_loader = param.weight_loader + weight_loader(param, loaded_weight, shard_id) + break + else: + for mapping in expert_params_mapping: + param_name, weight_name, expert_id, shard_id = mapping + if weight_name not in name: + continue + name = name.replace(weight_name, param_name) + param = params_dict[name] + weight_loader = param.weight_loader + weight_loader( + param, + loaded_weight, + name, + shard_id=shard_id, + expert_id=expert_id, + ) + break + else: + if "visual" in name: + # adapt to VisionAttention + name = name.replace(r"attn.qkv.", r"attn.qkv_proj.") + + # Skip loading extra bias for GPTQ models. + if name.endswith(".bias") and name not in params_dict: + continue + if fuse_qkv_a_proj and ( + "q_a_proj" in name or "kv_a_proj_with_mqa" in name + ): + cached_a_proj[name] = loaded_weight + q_a_proj_name = ( + name + if "q_a_proj" in name + else name.replace("kv_a_proj_with_mqa", "q_a_proj") + ) + kv_a_proj_name = ( + name + if "kv_a_proj_with_mqa" in name + else name.replace("q_a_proj", "kv_a_proj_with_mqa") + ) + + # When both q_a_proj and kv_a_proj_with_mqa has been cached, load the fused weight to parameter + if ( + q_a_proj_name in cached_a_proj + and kv_a_proj_name in cached_a_proj + ): + q_a_proj_weight = cached_a_proj[q_a_proj_name] + kv_a_proj_weight = cached_a_proj[kv_a_proj_name] + fused_weight = torch.cat( + [q_a_proj_weight, kv_a_proj_weight], dim=0 + ) + param_name = ( + name.replace("q_a_proj", "fused_qkv_a_proj_with_mqa") + if "q_a_proj" in name + else name.replace( + "kv_a_proj_with_mqa", "fused_qkv_a_proj_with_mqa" + ) + ) + param = params_dict[param_name] + + weight_loader = getattr( + param, "weight_loader", default_weight_loader + ) + weight_loader(param, fused_weight) + cached_a_proj.pop(q_a_proj_name) + cached_a_proj.pop(kv_a_proj_name) + else: + if ( + "k_scale" in name or "v_scale" in name + ) and name not in params_dict: + # modelopt attn kv scale is named differently + if any(scale in name for scale in ["k_scale", "v_scale"]): + name = name.replace("_proj", "attn_mqa") + else: + logger.warning( + f"Unknown scale found in checkpoint: {name}" + ) + param = params_dict[name] + weight_loader = getattr( + param, "weight_loader", default_weight_loader + ) + weight_loader(param, loaded_weight) + + +EntryClass = [Glm4vMoeForConditionalGeneration] diff --git a/python/sglang/srt/multimodal/processors/base_processor.py b/python/sglang/srt/multimodal/processors/base_processor.py index 760d3c26f..db80184c3 100644 --- a/python/sglang/srt/multimodal/processors/base_processor.py +++ b/python/sglang/srt/multimodal/processors/base_processor.py @@ -22,13 +22,19 @@ class BaseMultiModalProcessorOutput: input_text: str # frames loaded from image, in given order - images: Optional[list[Union[Image.Image, dict]]] = None + images: Optional[list[Union[Image.Image, dict]]] = dataclasses.field( + default_factory=list + ) # videos - videos: Optional[list[Union[torch.Tensor, dict]]] = None + videos: Optional[list[Union[torch.Tensor, dict]]] = dataclasses.field( + default_factory=list + ) # audios - audios: Optional[list[Union[np.ndarray, dict]]] = None + audios: Optional[list[Union[np.ndarray, dict]]] = dataclasses.field( + default_factory=list + ) def organize_results(self) -> List[Tuple[Modality, Any]]: """ diff --git a/python/sglang/srt/multimodal/processors/glm4v.py b/python/sglang/srt/multimodal/processors/glm4v.py new file mode 100644 index 000000000..58c55c0f8 --- /dev/null +++ b/python/sglang/srt/multimodal/processors/glm4v.py @@ -0,0 +1,132 @@ +import re +from typing import List, Union + +from decord import VideoReader +from transformers.video_utils import VideoMetadata + +from sglang.srt.layers.rotary_embedding import MRotaryEmbedding +from sglang.srt.models.glm4v import Glm4vForConditionalGeneration +from sglang.srt.models.glm4v_moe import Glm4vMoeForConditionalGeneration +from sglang.srt.multimodal.processors.base_processor import ( + BaseMultimodalProcessor as SGLangBaseProcessor, +) +from sglang.srt.multimodal.processors.base_processor import ( + BaseMultiModalProcessorOutput, + MultimodalSpecialTokens, +) + + +class Glm4vImageProcessor(SGLangBaseProcessor): + models = [Glm4vForConditionalGeneration, Glm4vMoeForConditionalGeneration] + + def __init__(self, hf_config, server_args, _processor, *args, **kwargs): + super().__init__(hf_config, server_args, _processor, *args, **kwargs) + + # GLM-4.1V and GLM-4.5V specific tokens + self.IMAGE_TOKEN = "<|image|>" + self.VIDEO_TOKEN = "<|video|>" + self.IMAGE_START_TOKEN = "<|begin_of_image|>" + self.IMAGE_END_TOKEN = "<|end_of_image|>" + self.VIDEO_START_TOKEN = "<|begin_of_video|>" + self.VIDEO_END_TOKEN = "<|end_of_video|>" + + # Token IDs + self.IM_TOKEN_ID = hf_config.image_token_id + self.VIDEO_TOKEN_ID = hf_config.video_token_id + self.IMAGE_START_TOKEN_ID = hf_config.image_start_token_id + self.IMAGE_END_TOKEN_ID = hf_config.image_end_token_id + self.VIDEO_START_TOKEN_ID = hf_config.video_start_token_id + self.VIDEO_END_TOKEN_ID = hf_config.video_end_token_id + + # Vision config + self.IMAGE_FACTOR = 28 + self.MIN_PIXELS = 112 * 112 + self.MAX_PIXELS = 30000 * 28 * 28 * 2 + + self.mm_tokens = MultimodalSpecialTokens( + image_token=self.IMAGE_TOKEN, + image_token_id=self.IM_TOKEN_ID, + video_token=self.VIDEO_TOKEN, + # Note: For GLM4v videos, it uses the video token before tokenization but uses image token after tokenization + video_token_id=self.IM_TOKEN_ID, + ).build(_processor) + + # adapted from https://github.com/huggingface/transformers/blob/369c99d0cea403b77bd0aef818527106453fd9fc/src/transformers/video_utils.py#L312 + async def preprocess_video(self, vr: VideoReader): + """ + Preprocess video using VideoReader from Decord backend. + + Args: + vr (VideoReader): VideoReader object from decord + + Returns: + tuple: A tuple containing processed frames and metadata + """ + video_fps = vr.get_avg_fps() + total_num_frames = len(vr) + duration = total_num_frames / video_fps if video_fps else 0 + + metadata = VideoMetadata( + total_num_frames=int(total_num_frames), + fps=float(video_fps), + duration=float(duration), + video_backend="decord", + ) + + # Extract all frames + indices = list(range(total_num_frames)) + frames = vr.get_batch(indices).asnumpy() + metadata.frames_indices = indices + + return frames, metadata + + async def process_mm_data_async( + self, + image_data: List[Union[str, bytes]], + input_text, + request_obj, + *args, + **kwargs, + ): + base_output = self.load_mm_data( + prompt=input_text, + image_data=image_data, + video_data=request_obj.video_data, + multimodal_tokens=self.mm_tokens, + ) + + video_metadata = None + + if base_output.videos: + videos_processed = [ + await self.preprocess_video(video) for video in base_output.videos + ] + base_output.videos, video_metadata = map(list, zip(*videos_processed)) + # transformer requires the video inputs to be under this format + base_output.videos = [base_output.videos] + video_metadata = [video_metadata] + + mm_items, input_ids, ret = self.process_and_combine_mm_data( + base_output, self.mm_tokens, video_metadata=video_metadata + ) + + input_ids = input_ids.flatten() + mrope_positions, mrope_position_delta = MRotaryEmbedding.get_rope_index_glm4v( + input_ids=input_ids.unsqueeze(0), + hf_config=self.hf_config, + image_grid_thw=getattr(ret, "image_grid_thw", None), + video_grid_thw=getattr(ret, "video_grid_thw", None), + attention_mask=getattr(ret, "attention_mask", None), + ) + mrope_positions = mrope_positions.squeeze(1) + + mm_inputs = { + "input_ids": input_ids.tolist(), + "mm_items": mm_items, + "im_token_id": self.mm_tokens.image_token_id, + "video_token_id": self.mm_tokens.video_token_id, + "mrope_positions": mrope_positions, + "mrope_position_delta": mrope_position_delta, + } + + return mm_inputs diff --git a/python/sglang/srt/utils.py b/python/sglang/srt/utils.py index edf441945..a234e7547 100644 --- a/python/sglang/srt/utils.py +++ b/python/sglang/srt/utils.py @@ -815,7 +815,7 @@ def load_video(video_file: Union[str, bytes], use_gpu: bool = True): vr = VideoReader(tmp_file.name, ctx=ctx) elif video_file.startswith("data:"): _, encoded = video_file.split(",", 1) - video_bytes = base64.b64decode(encoded) + video_bytes = pybase64.b64decode(encoded) tmp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") tmp_file.write(video_bytes) tmp_file.close() @@ -823,7 +823,7 @@ def load_video(video_file: Union[str, bytes], use_gpu: bool = True): elif os.path.isfile(video_file): vr = VideoReader(video_file, ctx=ctx) else: - video_bytes = base64.b64decode(video_file) + video_bytes = pybase64.b64decode(video_file) tmp_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") tmp_file.write(video_bytes) tmp_file.close() diff --git a/test/srt/openai_server/function_call/test_openai_function_calling.py b/test/srt/openai_server/function_call/test_openai_function_calling.py index 4efc04386..291ef98b7 100644 --- a/test/srt/openai_server/function_call/test_openai_function_calling.py +++ b/test/srt/openai_server/function_call/test_openai_function_calling.py @@ -948,5 +948,6 @@ class TestOpenAIPythonicFunctionCalling(CustomTestCase): # def test_function_calling_multiturn(self): # self._test_function_calling_multiturn() + if __name__ == "__main__": unittest.main() diff --git a/test/srt/test_function_call_parser.py b/test/srt/test_function_call_parser.py index afbba82e3..0c8cabfa6 100644 --- a/test/srt/test_function_call_parser.py +++ b/test/srt/test_function_call_parser.py @@ -497,6 +497,17 @@ class TestEBNFGeneration(unittest.TestCase): }, ), ), + Tool( + type="function", + function=Function( + name="empty_param_func", + description="Function with empty parameters", + parameters={ + "properties": {}, + "required": [], + }, + ), + ), ] self.tokenizer = get_tokenizer(DEFAULT_SMALL_MODEL_NAME_FOR_TEST) @@ -630,16 +641,21 @@ class TestEBNFGeneration(unittest.TestCase): self.assertIsNotNone(ebnf) # Check that the EBNF contains expected patterns for XML format self.assertIn('"" function_call ""', ebnf) - self.assertIn('"get_weather" "\\n" arguments_get_weather', ebnf) + self.assertIn('"get_weather" "\\n" ( arguments_get_weather "\\n" )?', ebnf) self.assertIn( '"location" "\\n" "" xml_text "" ( "\\n" ( "unit" "\\n" "" ("celsius" | "fahrenheit") "" ) )?', ebnf, ) - self.assertIn('"search" "\\n" arguments_search', ebnf) + self.assertIn('"search" "\\n" ( arguments_search "\\n" )?', ebnf) self.assertIn( '"query" "\\n" "" xml_text ""', ebnf, ) + self.assertIn( + '"empty_param_func" "\\n" ( arguments_empty_param_func "\\n" )?', ebnf + ) + self.assertIn('arguments_empty_param_func ::= ""', ebnf) + # Validate that the EBNF can be compiled by GrammarCompiler try: ctx = self.grammar_compiler.compile_grammar(ebnf) diff --git a/test/srt/test_jinja_template_utils.py b/test/srt/test_jinja_template_utils.py index 7764659d2..a861ac824 100644 --- a/test/srt/test_jinja_template_utils.py +++ b/test/srt/test_jinja_template_utils.py @@ -60,6 +60,86 @@ class TestTemplateContentFormatDetection(CustomTestCase): result = detect_jinja_template_content_format("") self.assertEqual(result, "string") + def test_detect_msg_content_pattern(self): + """Test detection of template with msg.content pattern (should be 'openai' format).""" + msg_content_pattern = """ +[gMASK] +{%- for msg in messages %} + {%- if msg.role == 'system' %} +<|system|> +{{ msg.content }} + {%- elif msg.role == 'user' %} +<|user|>{{ '\n' }} + {%- if msg.content is string %} +{{ msg.content }} + {%- else %} + {%- for item in msg.content %} + {%- if item.type == 'video' or 'video' in item %} +<|begin_of_video|><|video|><|end_of_video|> + {%- elif item.type == 'image' or 'image' in item %} +<|begin_of_image|><|image|><|end_of_image|> + {%- elif item.type == 'text' %} +{{ item.text }} + {%- endif %} + {%- endfor %} + {%- endif %} + {%- elif msg.role == 'assistant' %} + {%- if msg.metadata %} +<|assistant|>{{ msg.metadata }} +{{ msg.content }} + {%- else %} +<|assistant|> +{{ msg.content }} + {%- endif %} + {%- endif %} +{%- endfor %} +{% if add_generation_prompt %}<|assistant|> +{% endif %} + """ + + result = detect_jinja_template_content_format(msg_content_pattern) + self.assertEqual(result, "openai") + + def test_detect_m_content_pattern(self): + """Test detection of template with m.content pattern (should be 'openai' format).""" + msg_content_pattern = """ +[gMASK] +{%- for m in messages %} + {%- if m.role == 'system' %} +<|system|> +{{ m.content }} + {%- elif m.role == 'user' %} +<|user|>{{ '\n' }} + {%- if m.content is string %} +{{ m.content }} + {%- else %} + {%- for item in m.content %} + {%- if item.type == 'video' or 'video' in item %} +<|begin_of_video|><|video|><|end_of_video|> + {%- elif item.type == 'image' or 'image' in item %} +<|begin_of_image|><|image|><|end_of_image|> + {%- elif item.type == 'text' %} +{{ item.text }} + {%- endif %} + {%- endfor %} + {%- endif %} + {%- elif m.role == 'assistant' %} + {%- if m.metadata %} +<|assistant|>{{ m.metadata }} +{{ m.content }} + {%- else %} +<|assistant|> +{{ m.content }} + {%- endif %} + {%- endif %} +{%- endfor %} +{% if add_generation_prompt %}<|assistant|> +{% endif %} + """ + + result = detect_jinja_template_content_format(msg_content_pattern) + self.assertEqual(result, "openai") + def test_process_content_openai_format(self): """Test content processing for openai format.""" msg_dict = { diff --git a/test/srt/test_vision_openai_server_b.py b/test/srt/test_vision_openai_server_b.py index f954aee48..95941149d 100644 --- a/test/srt/test_vision_openai_server_b.py +++ b/test/srt/test_vision_openai_server_b.py @@ -348,6 +348,33 @@ class TestVILAServer(TestOpenAIVisionServer): cls.base_url += "/v1" +# Skip for ci test +# class TestGLM41VServer(TestOpenAIVisionServer): +# @classmethod +# def setUpClass(cls): +# cls.model = "zai-org/GLM-4.1V-9B-Thinking" +# cls.base_url = DEFAULT_URL_FOR_TEST +# cls.api_key = "sk-123456" +# cls.process = popen_launch_server( +# cls.model, +# cls.base_url, +# timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH, +# other_args=[ +# "--trust-remote-code", +# "--mem-fraction-static", +# "0.68", +# "--cuda-graph-max-bs", +# "4", +# "--reasoning-parser", +# "glm45", +# ], +# ) +# cls.base_url += "/v1" + +# def test_video_chat_completion(self): +# self._test_video_chat_completion() + + if __name__ == "__main__": del TestOpenAIVisionServer unittest.main() diff --git a/test/srt/test_vision_openai_server_common.py b/test/srt/test_vision_openai_server_common.py index a8c0aac38..7e30b3de2 100644 --- a/test/srt/test_vision_openai_server_common.py +++ b/test/srt/test_vision_openai_server_common.py @@ -96,8 +96,13 @@ class TestOpenAIVisionServer(CustomTestCase): ), f"text: {text}, should contain cab, taxi, SUV, vehicle or car" # MiniCPMO fails to recognize `iron`, but `hanging` assert ( - "iron" in text or "hang" in text or "cloth" in text or "holding" in text - ), f"text: {text}, should contain iron, hang, cloth or holding" + "iron" in text + or "hang" in text + or "cloth" in text + or "coat" in text + or "holding" in text + or "outfit" in text + ), f"text: {text}, should contain iron, hang, cloth, coat or holding or outfit" assert response.id assert response.created assert response.usage.prompt_tokens > 0 @@ -193,11 +198,15 @@ class TestOpenAIVisionServer(CustomTestCase): print(f"Multi images response:\n{text}") print("-" * 30) assert ( - "man" in text or "cab" in text or "SUV" in text or "taxi" in text - ), f"text: {text}, should contain man, cab, SUV or taxi" + "man" in text + or "cab" in text + or "SUV" in text + or "taxi" in text + or "car" in text + ), f"text: {text}, should contain man, cab, SUV, taxi or car" assert ( - "logo" in text or '"S"' in text or "SG" in text - ), f"text: {text}, should contain logo, S or SG" + "logo" in text or '"S"' in text or "SG" in text or "graphic" in text + ), f"text: {text}, should contain logo, S or SG or graphic" assert response.id assert response.created assert response.usage.prompt_tokens > 0 @@ -320,11 +329,12 @@ class TestOpenAIVisionServer(CustomTestCase): or "individual" in video_response or "speaker" in video_response or "Steve" in video_response + or "hand" in video_response ), f""" ====================== video_response ===================== {video_response} =========================================================== - should contain 'man' or 'person' or 'individual' or 'speaker' + should contain 'man' or 'person' or 'individual' or 'speaker' or 'hand' """ assert ( "present" in video_response @@ -375,7 +385,8 @@ class TestOpenAIVisionServer(CustomTestCase): or "person" in video_response or "individual" in video_response or "speaker" in video_response - ), f"video_response: {video_response}, should either have 'man' in video_response, or 'person' in video_response, or 'individual' in video_response or 'speaker' in video_response" + or "hand" in video_response + ), f"video_response: {video_response}, should either have 'man' in video_response, or 'person' in video_response, or 'individual' in video_response, or 'speaker' in video_response or 'hand' in video_response" assert ( "present" in video_response or "examine" in video_response