mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2026-04-23 00:17:25 +08:00
luukunn
8496ec71a6
* first commit * step 9~10 * update multimodal * update multimodal * fix load tokenizer * add unit test * fix unit test & AdaptiveImageProcessor * Delete unused code
454 lines
19 KiB
Python
454 lines
19 KiB
Python
"""
|
|
# Copyright (c) 2025 PaddlePaddle Authors. 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.
|
|
"""
|
|
|
|
"""Unified multimodal processor for all VL model types.
|
|
|
|
Consolidates the four separate VL processor wrappers (QwenVLProcessor,
|
|
Qwen3VLProcessor, PaddleOCRVLProcessor, Ernie4_5_VLProcessor) into a
|
|
single class that dispatches per ``model_type``.
|
|
"""
|
|
|
|
from collections.abc import Mapping
|
|
from typing import Any, Dict, Optional
|
|
|
|
import numpy as np
|
|
|
|
from fastdeploy.input.base_processor import BaseTextProcessor
|
|
from fastdeploy.input.utils import IDS_TYPE_FLAG, process_stop_token_ids
|
|
from fastdeploy.utils import data_processor_logger
|
|
|
|
QWEN_VL = "qwen_vl"
|
|
QWEN3_VL = "qwen3_vl"
|
|
PADDLEOCR_VL = "paddleocr_vl"
|
|
ERNIE4_5_VL = "ernie4_5_vl"
|
|
|
|
_SUPPORTED_MODEL_TYPES = {QWEN_VL, QWEN3_VL, PADDLEOCR_VL, ERNIE4_5_VL}
|
|
|
|
_QWEN_EXPECTED_KWARGS = {
|
|
"video_max_frames": int,
|
|
"video_min_frames": int,
|
|
}
|
|
|
|
_ERNIE_EXPECTED_KWARGS = {
|
|
"spatial_conv_size": int,
|
|
"temporal_conv_size": int,
|
|
"image_min_pixels": int,
|
|
"image_max_pixels": int,
|
|
"video_min_pixels": int,
|
|
"video_max_pixels": int,
|
|
"video_target_frames": int,
|
|
"video_frames_sample": str,
|
|
"video_max_frames": int,
|
|
"video_min_frames": int,
|
|
"video_fps": int,
|
|
}
|
|
|
|
_DEFAULT_MM_LIMITS = {"image": 1, "video": 1, "audio": 1}
|
|
|
|
_SAMPLING_EPS = 1e-5
|
|
|
|
|
|
class MultiModalProcessor(BaseTextProcessor):
|
|
"""Unified multimodal processor for all supported VL model types.
|
|
|
|
Dispatches image-processor creation, config initialisation, and
|
|
encoding logic based on ``model_type``.
|
|
"""
|
|
|
|
def __init__(
|
|
self,
|
|
model_name_or_path: str,
|
|
model_type: str,
|
|
config=None,
|
|
limit_mm_per_prompt: Optional[Dict[str, Any]] = None,
|
|
mm_processor_kwargs: Optional[Dict[str, Any]] = None,
|
|
reasoning_parser_obj=None,
|
|
tool_parser_obj=None,
|
|
enable_processor_cache: bool = False,
|
|
):
|
|
if model_type not in _SUPPORTED_MODEL_TYPES:
|
|
raise ValueError(
|
|
f"Unsupported model_type '{model_type}'. " f"Must be one of {sorted(_SUPPORTED_MODEL_TYPES)}."
|
|
)
|
|
self.model_type = model_type
|
|
self.config = config
|
|
self.enable_processor_cache = enable_processor_cache
|
|
|
|
tokenizer_type = "ernie4_5" if model_type == ERNIE4_5_VL else "auto"
|
|
|
|
super().__init__(
|
|
model_name_or_path,
|
|
tokenizer_type=tokenizer_type,
|
|
reasoning_parser_obj=reasoning_parser_obj,
|
|
tool_parser_obj=tool_parser_obj,
|
|
)
|
|
|
|
data_processor_logger.info(f"model_name_or_path: {model_name_or_path}")
|
|
|
|
processor_kwargs = self._parse_processor_kwargs(mm_processor_kwargs)
|
|
self._init_mm_processor(processor_kwargs)
|
|
self._init_mm_config()
|
|
self.limit_mm_per_prompt = self._parse_limits(limit_mm_per_prompt)
|
|
|
|
def _load_tokenizer(self):
|
|
"""Load the appropriate tokenizer based on model_type."""
|
|
if self.tokenizer_type == "ernie4_5":
|
|
import os
|
|
|
|
from fastdeploy.input.ernie4_5_tokenizer import Ernie4_5Tokenizer
|
|
|
|
vocab_file_names = ["tokenizer.model", "spm.model", "ernie_token_100k.model"]
|
|
for name in vocab_file_names:
|
|
if os.path.exists(os.path.join(self.model_name_or_path, name)):
|
|
Ernie4_5Tokenizer.resource_files_names["vocab_file"] = name
|
|
break
|
|
tokenizer = Ernie4_5Tokenizer.from_pretrained(self.model_name_or_path)
|
|
else:
|
|
from paddleformers.transformers import AutoTokenizer
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(self.model_name_or_path, padding_side="left", use_fast=True)
|
|
return tokenizer
|
|
|
|
def _init_mm_processor(self, processor_kwargs: dict):
|
|
"""Create the model-type-specific internal DataProcessor."""
|
|
if self.model_type == QWEN_VL:
|
|
from fastdeploy.input.qwen_vl_processor.process import DataProcessor
|
|
|
|
tokens_per_second = getattr(getattr(self.config, "vision_config", None), "tokens_per_second", 2)
|
|
self.processor = DataProcessor(
|
|
model_path=self.model_name_or_path,
|
|
enable_processor_cache=self.enable_processor_cache,
|
|
tokens_per_second=tokens_per_second,
|
|
tokenizer=self.tokenizer,
|
|
**processor_kwargs,
|
|
)
|
|
elif self.model_type == QWEN3_VL:
|
|
from fastdeploy.input.qwen3_vl_processor.process import DataProcessor
|
|
|
|
self.processor = DataProcessor(
|
|
model_path=self.model_name_or_path,
|
|
enable_processor_cache=self.enable_processor_cache,
|
|
tokenizer=self.tokenizer,
|
|
**processor_kwargs,
|
|
)
|
|
elif self.model_type == PADDLEOCR_VL:
|
|
from fastdeploy.input.paddleocr_vl_processor.process import DataProcessor
|
|
|
|
tokens_per_second = getattr(getattr(self.config, "vision_config", None), "tokens_per_second", 2)
|
|
self.processor = DataProcessor(
|
|
model_path=self.model_name_or_path,
|
|
enable_processor_cache=self.enable_processor_cache,
|
|
tokens_per_second=tokens_per_second,
|
|
tokenizer=self.tokenizer,
|
|
**processor_kwargs,
|
|
)
|
|
elif self.model_type == ERNIE4_5_VL:
|
|
from fastdeploy.input.ernie4_5_vl_processor.process import DataProcessor
|
|
|
|
self.processor = DataProcessor(
|
|
tokenizer_name=self.model_name_or_path,
|
|
image_preprocessor_name=self.model_name_or_path,
|
|
enable_processor_cache=self.enable_processor_cache,
|
|
**processor_kwargs,
|
|
)
|
|
self.processor.eval()
|
|
|
|
def _init_mm_config(self):
|
|
"""Set model-type-specific multimodal configuration attributes."""
|
|
if self.model_type in (QWEN_VL, QWEN3_VL):
|
|
self.image_patch_id = self.processor.image_token_id
|
|
elif self.model_type == PADDLEOCR_VL:
|
|
self.image_patch_id = self.processor.image_patch_id
|
|
elif self.model_type == ERNIE4_5_VL:
|
|
self.image_patch_id = self.processor.image_patch_id
|
|
self.spatial_conv_size = self.processor.spatial_conv_size
|
|
|
|
def _parse_processor_kwargs(self, kwargs: Optional[dict]) -> dict:
|
|
"""Parse and validate multimodal processor kwargs."""
|
|
if not kwargs:
|
|
return {}
|
|
|
|
try:
|
|
if not isinstance(kwargs, dict):
|
|
raise ValueError("mm-processor-kwargs must be a dictionary")
|
|
|
|
data_processor_logger.info(f"Processing kwargs: {kwargs}")
|
|
|
|
if self.model_type == ERNIE4_5_VL:
|
|
expected_types = _ERNIE_EXPECTED_KWARGS
|
|
else:
|
|
expected_types = _QWEN_EXPECTED_KWARGS
|
|
|
|
for key, value in kwargs.items():
|
|
if key in expected_types and not isinstance(value, expected_types[key]):
|
|
raise ValueError(
|
|
f"Invalid type for {key}: expected "
|
|
f"{expected_types[key].__name__}, got {type(value).__name__}"
|
|
)
|
|
return kwargs
|
|
|
|
except Exception as e:
|
|
data_processor_logger.warning(f"Invalid mm-processor-kwargs format: {e}")
|
|
return {}
|
|
|
|
def _parse_limits(self, limits: Optional[dict]) -> dict:
|
|
"""Parse multimodal input limits, merging with defaults."""
|
|
if not limits:
|
|
return dict(_DEFAULT_MM_LIMITS)
|
|
|
|
try:
|
|
if not isinstance(limits, dict):
|
|
raise ValueError("limit-mm-per-prompt must be a dictionary")
|
|
data_processor_logger.info(f"_parse_limits:{limits}")
|
|
return {**_DEFAULT_MM_LIMITS, **limits}
|
|
except Exception as e:
|
|
data_processor_logger.warning(f"Invalid limit-mm-per-prompt format: {e}, using default limits")
|
|
return dict(_DEFAULT_MM_LIMITS)
|
|
|
|
def _check_mm_limits(self, item):
|
|
"""Validate multimodal inputs against configured limits."""
|
|
if isinstance(item, dict):
|
|
mm_data = item
|
|
else:
|
|
mm_data = {"image": [], "video": []}
|
|
for message in item:
|
|
if isinstance(message.get("content"), list):
|
|
for part in message["content"]:
|
|
part_type = part.get("type")
|
|
if part_type in ("image_url", "image"):
|
|
mm_data["image"].append(part)
|
|
elif part_type in ("video_url", "video"):
|
|
mm_data["video"].append(part)
|
|
|
|
for modality, data in mm_data.items():
|
|
if modality in self.limit_mm_per_prompt:
|
|
limit = self.limit_mm_per_prompt[modality]
|
|
if len(data) > limit:
|
|
raise ValueError(f"Too many {modality} items in prompt, " f"got {len(data)} but limit is {limit}")
|
|
|
|
def get_mm_max_tokens_per_item(self, seq_len: int) -> Optional[Mapping[str, int]]:
|
|
"""Return per-modality max token counts, if available."""
|
|
if self.model_type == ERNIE4_5_VL:
|
|
return self.processor.get_mm_max_tokens_per_item(seq_len)
|
|
return None
|
|
|
|
def process_request_dict(self, request, max_model_len=None):
|
|
"""Process a request dictionary into model inputs.
|
|
|
|
Unified template-method flow for all VL model types. Per-model
|
|
differences are handled by small conditional branches rather than
|
|
duplicating the entire pipeline.
|
|
"""
|
|
request = self._apply_default_parameters(request)
|
|
|
|
if not request.get("eos_token_ids"):
|
|
request["eos_token_ids"] = self.eos_token_ids
|
|
|
|
self._process_stop_tokens(request)
|
|
|
|
if self.model_type != PADDLEOCR_VL:
|
|
self._process_bad_words(request)
|
|
|
|
if self.model_type == ERNIE4_5_VL:
|
|
logits_processors_args = self._prepare_think_stop_sentence(
|
|
request.get("logits_processors_args") or {}, max_model_len
|
|
)
|
|
request["logits_processors_args"] = logits_processors_args
|
|
|
|
outputs = self._tokenize_request(request)
|
|
|
|
self._process_post_tokens(request, outputs)
|
|
|
|
if self.model_type in (QWEN_VL, QWEN3_VL):
|
|
request["enable_thinking"] = False
|
|
|
|
outputs = self.pack_outputs(outputs)
|
|
|
|
if self.model_type in (QWEN3_VL, ERNIE4_5_VL) and request.get("prompt_token_ids"):
|
|
pass # preserve existing prompt_token_ids
|
|
else:
|
|
request["prompt_token_ids"] = outputs["input_ids"].tolist()
|
|
request["prompt_token_ids_len"] = len(request["prompt_token_ids"])
|
|
request["multimodal_inputs"] = outputs
|
|
|
|
if max_model_len is not None and len(request["prompt_token_ids"]) > max_model_len:
|
|
request["prompt_token_ids"] = request["prompt_token_ids"][: max_model_len - 1]
|
|
|
|
if self.model_type == ERNIE4_5_VL:
|
|
logits_processors_args = self._update_thinking_prompt_state(
|
|
request["prompt_token_ids"], request.get("logits_processors_args") or {}
|
|
)
|
|
request["logits_processors_args"] = logits_processors_args
|
|
|
|
max_tokens = max_model_len - len(request["prompt_token_ids"])
|
|
if request.get("max_tokens") is None:
|
|
request["max_tokens"] = max(1, max_tokens)
|
|
else:
|
|
request["max_tokens"] = min(max_tokens, request["max_tokens"])
|
|
|
|
if self.model_type == ERNIE4_5_VL and request.get("reasoning_max_tokens") is None:
|
|
request["reasoning_max_tokens"] = max(int(request["max_tokens"] * 0.8), 1)
|
|
|
|
if self.model_type in (PADDLEOCR_VL, ERNIE4_5_VL):
|
|
if request.get("top_p") is not None and request.get("top_p") < _SAMPLING_EPS:
|
|
request["top_p"] = _SAMPLING_EPS
|
|
request["top_k"] = 1
|
|
|
|
if self.model_type != QWEN3_VL and self.reasoning_parser:
|
|
self._apply_reasoning_parser(request)
|
|
|
|
if self.model_type == ERNIE4_5_VL:
|
|
if request.get("response_max_tokens") is not None and request.get("enable_thinking") is False:
|
|
request["max_tokens"] = min(request["response_max_tokens"], request["max_tokens"])
|
|
|
|
data_processor_logger.info(f"Processed request {request}")
|
|
return request
|
|
|
|
def _process_stop_tokens(self, request):
|
|
"""Handle stop token processing based on model type."""
|
|
if self.model_type == QWEN3_VL:
|
|
stop_sequences = request.get("stop", [])
|
|
if stop_sequences:
|
|
stop_seqs, stop_seqs_len = self.update_stop_seq(stop_sequences)
|
|
request["stop_token_ids"] = stop_seqs
|
|
request["stop_seqs_len"] = stop_seqs_len
|
|
else:
|
|
process_stop_token_ids(request, self.update_stop_seq)
|
|
|
|
def _process_bad_words(self, request):
|
|
"""Process bad_words into token ids."""
|
|
bad_words = request.get("bad_words")
|
|
bad_words_token_ids = request.get("bad_words_token_ids")
|
|
if bad_words:
|
|
bad_words_token_ids = self.update_bad_words(bad_words, bad_words_token_ids)
|
|
request["bad_words_token_ids"] = bad_words_token_ids
|
|
|
|
def _tokenize_request(self, request):
|
|
"""Core tokenization dispatch: prompt_token_ids > prompt > messages."""
|
|
default_thinking = True if self.model_type == ERNIE4_5_VL else False
|
|
|
|
if request.get("prompt_token_ids") and self.model_type in (QWEN3_VL, ERNIE4_5_VL):
|
|
messages = request.get("messages")
|
|
if messages:
|
|
self._check_mm_limits(messages)
|
|
request.setdefault("enable_thinking", default_thinking)
|
|
return self.processor.prompt_token_ids2outputs(request)
|
|
|
|
elif request.get("prompt"):
|
|
multimodal_data = request.get("multimodal_data") or {}
|
|
self._check_mm_limits(multimodal_data)
|
|
images = multimodal_data.get("image", None)
|
|
videos = multimodal_data.get("video", None)
|
|
if self.model_type == ERNIE4_5_VL:
|
|
request["prompt_tokens"] = request.get("prompt")
|
|
request.setdefault("enable_thinking", default_thinking)
|
|
return self.processor.text2ids(request["prompt"], images, videos)
|
|
|
|
elif request.get("messages"):
|
|
messages = request["messages"]
|
|
self._check_mm_limits(messages)
|
|
chat_template_kwargs = request.get("chat_template_kwargs")
|
|
if chat_template_kwargs:
|
|
if isinstance(chat_template_kwargs, dict):
|
|
for k, v in chat_template_kwargs.items():
|
|
if k not in request or request[k] is None:
|
|
request[k] = v
|
|
else:
|
|
raise ValueError("Invalid input: chat_template_kwargs must be a dict")
|
|
request.setdefault("enable_thinking", default_thinking)
|
|
return self.processor.request2ids(request)
|
|
|
|
else:
|
|
raise ValueError(f"Request must contain 'prompt', or 'messages': {request}")
|
|
|
|
def _process_post_tokens(self, request, outputs):
|
|
"""Handle post-tokenization token appending."""
|
|
if self.model_type == PADDLEOCR_VL:
|
|
metadata = request.get("metadata")
|
|
if metadata and metadata.get("generated_token_ids"):
|
|
self._append_completion_tokens_qwen(outputs, metadata["generated_token_ids"])
|
|
else:
|
|
if request.get("completion_token_ids"):
|
|
self.append_completion_tokens(outputs, request["completion_token_ids"])
|
|
|
|
def _apply_reasoning_parser(self, request):
|
|
"""Apply reasoning parser and update model status dict."""
|
|
model_status = self.reasoning_parser.get_model_status(request["prompt_token_ids"])
|
|
parts = request["request_id"].split("_")
|
|
if len(parts) > 1:
|
|
real_req_id = parts[0]
|
|
index = int(parts[1])
|
|
n = request.get("n", 1)
|
|
for idx in range(index * n, (index + 1) * n):
|
|
self.model_status_dict[f"{real_req_id}_{idx}"] = model_status
|
|
else:
|
|
self.model_status_dict[request["request_id"]] = model_status
|
|
request["enable_thinking"] = model_status == "think_start"
|
|
|
|
def append_completion_tokens(self, multimodal_inputs, completion_token_ids):
|
|
"""Append completion tokens to existing multimodal outputs."""
|
|
if self.model_type == ERNIE4_5_VL:
|
|
self._append_completion_tokens_ernie(multimodal_inputs, completion_token_ids)
|
|
else:
|
|
self._append_completion_tokens_qwen(multimodal_inputs, completion_token_ids)
|
|
|
|
def _append_completion_tokens_qwen(self, multimodal_inputs, completion_token_ids):
|
|
"""Append completion tokens for qwen_vl / qwen3_vl / paddleocr_vl."""
|
|
num_tokens = len(completion_token_ids)
|
|
multimodal_inputs["input_ids"].extend(completion_token_ids)
|
|
multimodal_inputs["token_type_ids"].extend([0] * num_tokens)
|
|
|
|
pos_ids = self.processor._compute_text_positions(multimodal_inputs["cur_position"], num_tokens)
|
|
multimodal_inputs["position_ids"].append(pos_ids)
|
|
multimodal_inputs["cur_position"] += num_tokens
|
|
|
|
def _append_completion_tokens_ernie(self, multimodal_inputs, completion_token_ids):
|
|
"""Append completion tokens for ernie4_5_vl."""
|
|
num_tokens = len(completion_token_ids)
|
|
multimodal_inputs["input_ids"].extend(completion_token_ids)
|
|
multimodal_inputs["token_type_ids"].extend([IDS_TYPE_FLAG["text"]] * num_tokens)
|
|
|
|
start = multimodal_inputs["cur_position"]
|
|
for i in range(num_tokens):
|
|
multimodal_inputs["position_ids"].append([start + i] * 3)
|
|
multimodal_inputs["cur_position"] += num_tokens
|
|
|
|
def pack_outputs(self, outputs):
|
|
"""Convert intermediate processing outputs to final format."""
|
|
if not outputs["images"]:
|
|
outputs["images"] = None
|
|
outputs["grid_thw"] = None
|
|
outputs["image_type_ids"] = None
|
|
else:
|
|
outputs["images"] = np.vstack(outputs["images"])
|
|
outputs["grid_thw"] = np.vstack(outputs["grid_thw"])
|
|
outputs["image_type_ids"] = np.array(outputs["image_type_ids"])
|
|
|
|
outputs["input_ids"] = np.array(outputs["input_ids"], dtype=np.int64)
|
|
outputs["token_type_ids"] = np.array(outputs["token_type_ids"], dtype=np.int64)
|
|
outputs["mm_num_token_func"] = self.processor.mm_num_tokens
|
|
|
|
if self.model_type in (QWEN_VL, QWEN3_VL, PADDLEOCR_VL):
|
|
outputs["position_ids"] = np.concatenate(outputs["position_ids"], axis=1, dtype=np.int64)
|
|
outputs["image_patch_id"] = self.processor.image_token_id
|
|
outputs["video_patch_id"] = self.processor.video_token_id
|
|
outputs["position_ids"] = outputs["position_ids"].transpose(1, 0)
|
|
else:
|
|
outputs["position_ids"] = np.array(outputs["position_ids"], dtype=np.int64)
|
|
outputs["image_patch_id"] = self.image_patch_id
|
|
|
|
return outputs
|