guoyande
guoyande
物理删除我们正在开发中。
目前不支持该功能，我们评估一下这个功能的必要性，如果价值较大，则进行排期开发。
string, int, long, float, double是标量类型的字段，vector是向量类型的字段。标量字段支持过滤，向量用于相似搜索。
重新导入一下数据吧，没有太好的方法
可以使用guoyande/gamma:scann镜像构建，镜像内bazel、gcc9、clang8均安装，需要手动安装一个python3.7/3.8, 然后pip install tensorflow~=2.5.0，最后编译scann 选项打开，正常编译gamma就可以。值的注意的是，scann需要下载github代码注意网络正常
文档没法看
I have the same problem. How did you solve it? Request “localhost:8000/v2/models/tensorrt_llm_bls/generate” interface can be successful. docker image: nvcr.io/nvidia/tritonserver:24.08-trtllm-python-py3 Perf: V100-SXM2-32GB Driver Version: 550.90.12 CUDA Version: 12.4
I solved it. The main reason is that the versions of "tensorrt-llm" and "tensorrtllm_backend" are different.
https://github.com/baidu/braft/pull/491 链接的时候没有链接brpc pthread, 已经提pr
https://forums.developer.nvidia.com/t/tensorrt-inference-api-that-open-clip-vit-l-14-is-slowing-down/309551/3 My successful convertion. "ViT-L/14" ---> .onnx ---> .trt But inference with the tensorrt framework is slower. Is this a normal phenomenon? The link above has some details.