data-juicer 有支持arm cpu架构吗？

[qifengz]

Before Submitting

• I have asked the Data-Juicer Q&A Copilot (available on Doc Site, DingTalk, or Discord) about this feature, but still want to submit it as a formal request.

• I have searched the Data-Juicer issues and found no similar feature requests.

Description

目前arm架构下编译报错，如下部分日志：
inux2014_aarch64.manylinux_2_17_aarch64.whl #17 237.7 DEBUG No cache entry for: https://pypi.tuna.tsinghua.edu.cn/packages/b8/56/6033f5e86e8cc9bb629f0077ba71679508bdf54a9a5e112a3c0b91870332/cffi-2.0.0-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl #17 237.7 DEBUG Sending fresh HEAD request for: https://pypi.tuna.tsinghua.edu.cn/packages/b8/56/6033f5e86e8cc9bb629f0077ba71679508bdf54a9a5e112a3c0b91870332/cffi-2.0.0-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl #17 238.1 DEBUG No cache entry for: https://pypi.tuna.tsinghua.edu.cn/packages/ff/8a/9ce0424d456cd3580cc3a855b23a7ff86b81d5299fceb496a2f56f06c1c0/thinc-8.3.13-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl #17 238.1 DEBUG Sending fresh HEAD request for: https://pypi.tuna.tsinghua.edu.cn/packages/ff/8a/9ce0424d456cd3580cc3a855b23a7ff86b81d5299fceb496a2f56f06c1c0/thinc-8.3.13-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.whl #17 238.4 × No solution found when resolving dependencies: #17 238.4 ╰─▶ Because torchcodec==0.7.0 has no wheels with a matching platform tag #17 238.4 (e.g., manylinux_2_39_aarch64) and py-data-juicer[all]==1.5.1 depends #17 238.4 on torchcodec==0.7, we can conclude that py-data-juicer[all]==1.5.1 #17 238.4 cannot be used. #17 238.4 And because only py-data-juicer[all]==1.5.1 is available and you #17 238.4 require py-data-juicer[all], we can conclude that your requirements #17 238.4 are unsatisfiable. #17 238.4 #17 238.4 hint: Wheels are available for torchcodec (v0.7.0) on the following #17 238.4 platforms: manylinux_2_28_x86_64, macosx_11_0_arm64, win_amd64

大致意思是，torchcodec目前版本是0.7，要0.11.0以上才有 aarch64的包。另外decord依赖包也存在同样问题：

#17 280.9 decord==0.4.1 #17 280.9 decord==0.4.2 #17 280.9 decord==0.5.0 #17 280.9 decord==0.5.1 #17 280.9 decord==0.5.2 #17 280.9 decord==0.6.0 #17 280.9 and decord<=0.2.0 has no wheels with a matching Python ABI tag (e.g., #17 280.9 cp311), we can conclude that decord<=0.2.0 cannot be used. #17 280.9 And because decord>=0.3.0 has no wheels with a matching platform tag #17 280.9 (e.g., manylinux_2_39_aarch64) and py-data-juicer[all]==1.5.1 depends #17 280.9 on decord, we can conclude that py-data-juicer[all]==1.5.1 cannot be #17 280.9 used. #17 280.9 And because only py-data-juicer[all]==1.5.1 is available and you #17 280.9 require py-data-juicer[all], we can conclude that your requirements #17 280.9 are unsatisfiable. #17 280.9 #17 280.9 hint: Pre-releases are available for decord in the requested #17 280.9 range (e.g., 0.2.0b20190929), but pre-releases weren't enabled (try: #17 280.9 --prerelease=allow) #17 280.9 #17 280.9 hint: You require CPython 3.11 (cp311), but we only found wheels for #17 280.9 decord (v0.2.0) with the following Python ABI tags: cp35m, cp36m, #17 280.9 cp37m #17 280.9 #17 280.9 hint: Wheels are available for decord (v0.6.0) on the following #17 280.9 platforms: manylinux2010_x86_64, win_amd6

Use Case

用户aarch64架构环境下需要

Proposed Solution

升级相关依赖包以适配aarch64架构

Additional Context

No response

Are you willing to submit a PR for this feature?

• Yes I'd like to help by submitting a PR!