[release/2.7] Upgrade numpy versions

[ethanwee1]

Relates to: ROCm/builder#82

Validation:
http://rocm-ci.amd.com/job/mainline-pytorch2.7-manylinux-wheels/80/

Ran the microbenchmark, import pandas, torch, numpy in
registry-sc-harbor.amd.com/framework/compute-rocm-dkms-no-npi-hipclang:16180_ubuntu24.04_py3.12_pytorch_lw_upgrade_numpy_versions_546fe04d

root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# pip list | egrep "numpy|pandas"
numpy                   2.1.2
pandas                  2.2.3
Python 3.12.3 (main, Feb  4 2025, 14:48:35) [GCC 13.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import pandas
>>> import numpy
>>> import torch
root@ubb4-rack-22:/var/lib/jenkins/pytorch-micro-benchmarking# python3 micro_benchmarking_pytorch.py --network resnet50
INFO: running forward and backward for warmup.
INFO: running the benchmark..
OK: finished running benchmark..
--------------------SUMMARY--------------------------
Microbenchmark for network : resnet50
Num devices: 1
Dtype: FP32
Mini batch size [img] : 64
Time per mini-batch : 0.10239996910095214
Throughput [img/sec] : 625.0001885928783

[rocm-repo-management-api]

Jenkins build for 546fe04dd3229a5a9fc9eee879d81b7895464303 commit finished as FAILURE
Links: Blue Ocean view / Build artifacts

[rocm-repo-management-api]

Jenkins build for 546fe04dd3229a5a9fc9eee879d81b7895464303 commit finished as FAILURE
Links: Blue Ocean view / Build artifacts