centersl/MinerU
1
0
Fork 0
mirror of https://github.com/opendatalab/MinerU.git synced 2026-03-27 11:08:32 +07:00
Code Issues Packages Projects Releases Wiki Activity

Merge pull request #4409 from opendatalab/dev

Browse Source 
Dev
This commit is contained in:
Xiaomeng Zhao
2026-01-23 21:32:25 +08:00
committed by GitHub
parent ce620fedc9 cde5d02ffc
commit d05577666d
5 changed files with 23 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
README.md
View File

@@ -46,6 +46,14 @@
# Changelog
- 2026/01/23 2.7.2 Release
- Added support for domestic computing platforms Hygon, Enflame, and Moore Threads. Currently, the officially supported domestic computing platforms include:
- [Ascend](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/Ascend/)
- [T-Head](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/THead/)
- [METAX](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/METAX/)
- [Hygon](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/Hygon/)
- [Enflame](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/Enflame/)
- [MooreThreads](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/MooreThreads/)
- MinerU continues to ensure compatibility with domestic hardware platforms, supporting mainstream chip architectures. With secure and reliable technology, we empower researchers, government, and enterprises to reach new heights in document digitization!
- Cross-page table merging optimization, improving merge success rate and merge quality
- 2026/01/06 2.7.1 Release

6
README_zh-CN.md
View File

@@ -46,10 +46,14 @@
# 更新记录
- 2026/01/23 2.7.2 发布
- 新增国产算力平台适配:
- 新增国产算力平台海光、燧原、摩尔线程的适配支持，目前已由官方适配并支持的国产算力平台包括:
- [昇腾 Ascend](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/Ascend)
- [平头哥 T-Head](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/THead)
- [沐曦 METAX](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/METAX)
- [海光 DCU](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/Hygon/)
- [燧原 GCU](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/Enflame/)
- [摩尔线程 MUSA](https://opendatalab.github.io/MinerU/zh/usage/acceleration_cards/MooreThreads/)
- MinerU 持续兼容国产硬件平台，支持主流芯片架构。以安全可靠的技术，助力科研、政企用户迈向文档数字化新高度！
- 跨页表合并优化，提升合并成功率与合并效果
- 2026/01/06 2.7.1 发布

2
docs/zh/usage/acceleration_cards/Enflame.md
View File

@@ -4,7 +4,7 @@
os: Ubuntu 22.04.4 LTS
cpu: Intel x86-64
gcu: Enflame S60
driver: 1.7.0.9
driver: 1.7.0.9
docker: 28.0.1
```

15
mineru/utils/block_sort.py
View File

@@ -187,14 +187,17 @@ def model_init(model_name: str):
elif device_name.startswith("mps"):
bf_16_support = True
elif device_name.startswith("gcu"):
if torch.gcu.is_bf16_supported():
bf_16_support = True
if hasattr(torch, 'gcu') and torch.gcu.is_available():
if torch.gcu.is_bf16_supported():
bf_16_support = True
elif device_name.startswith("musa"):
if torch.musa.is_bf16_supported():
bf_16_support = True
if hasattr(torch, 'musa') and torch.musa.is_available():
if torch.musa.is_bf16_supported():
bf_16_support = True
elif device_name.startswith("npu"):
if torch.npu.is_bf16_supported():
bf_16_support = True
if hasattr(torch, 'npu') and torch.npu.is_available():
if torch.npu.is_bf16_supported():
bf_16_support = True
if model_name == 'layoutreader':
# 检测modelscope的缓存目录是否存在

6
mineru/utils/table_merge.py
View File

@@ -586,9 +586,3 @@ def merge_table(page_info_list):
for block in current_table_block["blocks"]:
block['lines'] = []
block[SplitFlag.LINES_DELETED] = True