WangxuP

以后是否支持在保存word的时候，将文本的标题样式也进行保存？

4

比如支持 正文、一级标题、二级标题 的自动识别

[BUG] RuntimeError: NPU out of memory. Tried to allocate 268.00 MiB

1

``` python3 -m fastchat.serve.cli --model-path /home/models/Qwen1.5-32B-Chat --device npu --gpus 0,1,2,3 (fast_chat) [root@localhost ~]# python3 -m fastchat.serve.cli --model-path /home/models/Qwen1.5-32B-Chat --device npu --gpus 0,1,2,3 /root/miniconda3/envs/fast_chat/lib/python3.8/site-packages/torch_npu/dynamo/__init__.py:18: UserWarning: Register eager implementation for the 'npu'...

How can I use multiple GPUs for inference.

6

**here is my GPUs info:** ``` GPU info: H800 * 8 CUDA: 11.8 nvidia-smi Mon Sep 4 10:39:54 2023 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 520.61.05 Driver Version: 520.61.05 CUDA Version: 11.8 |...

Under the duplicate directory, rasa SDK import actions will make an error #595

1

**Proposed changes**: - It optimizes the form of Rasa SDK reading the package under the whole project, **Status (please check what you already did)**: - [x] made PR ready for...

【ascend】 RuntimeError: [ERROR] HCCL error in: torch_npu/csrc/distributed/ProcessGroupHCCL.cpp:64

1

### Reminder - [X] I have read the README and searched the existing issues. ### Reproduction bash examples/lora_multi_npu/ds_zero0.sh ds_zero0.sh ``` #!/bin/bash NPROC_PER_NODE=4 NNODES=1 RANK=0 MASTER_ADDR=localhost MASTER_PORT=29500 ASCEND_RT_VISIBLE_DEVICES=0,1,2,3 torchrun \ --nproc_per_node...

调用/v1/chat/completions接口,用jmeter10并发进行压测，压测1分钟xinference就挂了，xinference==0.11.3

21

### Describe the bug 我们在压测xinference时候发现，V100 2卡，调用/v1/chat/completions接口，stream参数是True，模型用qwen-14b-chat，用jmeter10并发进行压测，压测1分钟xinference就挂了，如果stream是False,是可以的. ### 报错日志 ``` 2024-07-08 11:34:32,621 xinference.api.restful_api 8 INFO Disconnected from client (via refresh/close) Address(host='192.168.32.13', port=30733) during chat. INFO 07-08 11:34:32 async_llm_engine.py:158] Aborted request fcdb2432-3cda-11ef-af98-7e88271d2e8e....

consider adding an environment variable to automatically expand the deepseek_r1result instead of automatically collapsing it?

[Bug]: Grouped Retrieval Using the BM25 Algorithm

7

### Is there an existing issue for this? - [x] I have searched the existing issues ### Environment ```markdown - Milvus version: - Deployment mode(standalone or cluster): - MQ type(rocksmq,...

swift grpo: host pid is error

4

**What happened**: 使用swift进行grpo时候，在2张虚拟卡上进行训练，出现nccl错误，同时前面报host pid is error相关错误 **Environment**: - HAMi version: v2.6.1 - nvidia driver or other AI device driver version: 560.35.03 - Docker version from `docker version`: 20.10.16 - Docker...