This method forwards all its arguments to SiglipTokenizer's [`~PreTrainedTokenizer.batch_decode`]. Please refer to the docstring of t

Predict the sample from the previous timestep by reversing the SDE. This function propagates the flow process from the learned model outputs

r""" Args: prompt (`str` or `List[str]`, *optional*): prompt to be encoded prompt_2 (`str` or `L

r""" Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles custom timesteps. Any k

Args: sequence_length: length of sequence. packed_text_ids: 1-D int tensor, packed text token ids. packed

Encode a list of sentences to their respective embeddings. The sentences can be a list of strings or a string. Args: sent

Applies Rotary Position Embedding to the query and key tensors. Args: q (`torch.Tensor`): The query tensor. k (`torch.Tensor`): T

Load lines from a file. First tries to load from `path` directly, and if that doesn't exist, searches the `flow_grpo/assets` directory for a

r""" Encodes the prompt into text encoder hidden states. Args: prompt (`str` or `List[str]`, *optional*): prompt to be en

r""" Returns: image_features (`torch.FloatTensor` of shape `(batch_size, output_dim`): The image embeddings obtained by

加强版预处理函数： - 默认同时去除空白字符和标点符号 - 可通过参数控制是否去除标点符号 - 保留中文、英文、数字和#号

计算特定语言和级别(image/box)的指标

Encodes the given image into a feature representation using a pre-trained image encoder. Args: image (`PipelineImageInput`):

Creates and runs a batch of asynchronous evaluation tasks with proper error handling.

带重试机制的异步图像评估函数

Initialize distributed training

根据语言模式解析prompt（en：英文解析，cn/zh：中文解析）

Converts a PIL image to a base64 string for API calls.

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

响应后处理：提取识别文本，计算质量分、语义分 Args: response: LLMClient返回的原始响应 target: 参考文本（可选） Returns: result: 包含识别结果、评分的字典

Create an instance from an existing Pipeline

Generates a rectangle corresponding to a given point, using the Gurobi. Parameters: - points: List of tuples [(x1, y1), (x2, y2), ..

Generates a rectangle corresponding to a given point, using the Gurobi. Parameters: - points: List of tuples [(x1, y1), (x2, y2), ..

Creates the prompt for the Recognition task