Feature Overview 
We want RHEL AI 1.5 to consume torch==2.6.0, as indicated in the formal "RHEL AI components version" document: https://docs.google.com/document/d/1rk5lgztANsY9SO4xYUwhT0K3Q0OSSHtntk5GBo0Gf4o/edit?usp=sharing __ 

However, in order to allow RHEL AI 1.5 to consume this version of PyTorch, we need to increase the upper bound of the `torch` dependency in our upstream repo (instructlab/instructlab) from `<2.6.0` to `<2.7.0`. Without increasing this upper bound in the upstream repo, the downstream RHEL AI build will not be able to consume `torch==2.6.0`.

Goals

• Update the top-level `requirements.txt` file to increase the upper bound of PyTorch 
• Ensure that the CI is testing against `torch==2.6.0`

Requirements:

N/A - this is just tracking the upstream updates in order to enable our downstream processes to build wheels and containers against torch==2.6.0. 

Done - Acceptance Criteria

• The `torch` entry in `requirements.txt` is updated from `torch<2.6.0` to `torch<2.7.0`
• The upstream CI is testing against `torch==2.6.0`
• All upstream CI checks are passing