夏悠然

WebUI界面图标不显示

6

## NNI 学生项目问题概述 / General Question of Student Program **请简要概述您的问题 / 观点 :** **Short summary about the question / idea :** **WebUI界面图标不显示 浏览器为Chrome 92.0.4515.107(64位， win7系统)**  **请提供 NNI 环境信息 :**...

在输出结果中添加数据流的信息，比如tensor的shape信息✨✨✨

国庆期间😴还在家造各种Pass的case，结果今天发现官方代码中已经提供了😂😂😂 具体的case请移步👉 [tests/python/relay/test_pass_*.py](https://github.com/apache/tvm/tree/main/tests/python/relay)

Suggest supporting thenlper/gte-small-zh

1

### System Info Question: **The reasoning results of TEI and sentence-transformers are inconsistent.** TEI version: 1.2 sentence-transformers: 2.3.1 embedding-model: thenlper/gte-small-zh ### Information - [X] Docker - [ ] The CLI...

support TEI

2

support TEI(Text Embeddings Inference), For detailed information, visit: [https://github.com/huggingface/text-embeddings-inference](https://github.com/huggingface/text-embeddings-inference).

docker build failed

5

When I run it locally, I follow the documentation： **step1:** ```bash cp langwatch/.env.example langwatch/.env ``` **step2:** ```bash docker-compose up --build ``` then docker build failed!!! error log: ```bash => ERROR...

knowledge-retrieval: rerank model can’t be closed

1

### Self Checks - [x] This is only for bug report, if you would like to ask a question, please head to [Discussions](https://github.com/langgenius/dify/discussions/categories/general). - [x] I have searched for existing...

[ToolParser] Add Qwen2.5 models tool parser

1

Add Qwen2.5 models tool parser, support model: Qwen2.5-1.5B-Instruct, Qwen2.5-7B-Instruct, Qwen2.5-32B-Instruct-GPTQ-Int4, etc.

GPU利用率不稳定问题

1

我使用旧版本的代码做embedding模型的预训练，在两张A100上进行测试，通过nvitop检测到cpu利用率100%了，但是GPU的利用率只有小部分时间是100%，其余时间都是0，我尝试调整dataloader_num_workers，但是并没有效果。我的脚本如下： ```bash CUDA_VISIBLE_DEVICES=0,1 torchrun --master_port 20036 --nproc_per_node 2 \ -m FlagEmbedding.baai_general_embedding.retromae_pretrain.run \ --output_dir /root/data1/bge-large-zh-v1.5-test \ --model_name_or_path /root/data1/huggingface/BAAI/bge-large-zh-v1.5 \ --train_data /root/data1/BAAI_DATA/PreTrain-Data \ --learning_rate 2e-5 \ --num_train_epochs 2 \ --per_device_train_batch_size 16 \...