标签: Transformer

ValueError: (‘The batch size of the image embeddings and the input points must be the same. ‘, ‘Got 5 and 1 respectively.’, ‘ if you want to pass multiple points for the same image, make sure

ValueError: ('The batch size of the image embeddings and the input points must be the same. ', 'Got 5 and 1 respectively.', ' if you want to pass multiple points for the same image, make sure

该报错通常出现在使用 Transformers 的 mask-generation pipeline 运行 SAM 模型、并把 crops_n_layers 设为大于 0 的场景。优先排查是否存在“pipeline 在裁剪模式下产生多个输出、与输入点 batch size 不一致”的已知 bug,而

UMT5 incredibly slow in generating

UMT5 incredibly slow in generating

在使用 Transformers 加载 google/umt5-small 并执行 generate 生成时,会出现“UMT5 incredibly slow in generating”,即同样参数规模下 UMT5 的生成耗时明显高于 mbart/t5/mt5。优先排查 UMT5 相对位置偏置计算

CUDA out of memory

CUDA out of memory

在 vLLM 上为 Gemma 4 多模态模型启用 encoder CUDA graph( cudagraph_mm_encoder: true )并开启视频输入时,捕获阶段会按 max_frames_per_batch 为每个预算分配整幅图像尺寸的 dense pixel buffer,导致出现