whisper: Fix numpy version compatibility issue on gaudi

[lianhao]

Description

Issues

Fixes #1800

Type of change

List the type of change like below. Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

Dependencies

List the newly introduced 3rd party dependency if exists.

Tests

Describe the tests that you ran to verify your changes.

[eero-t]

Approved, but I wonder whether it would make more sense just to move version enforcement to asr/src/requirements.txt?

PS. There are couple of other Dockerfiles, which are using their own requirements.txt files instead of one from another component like is done here. Why the heck those do not specify proper numpy version in their requirements.txt files?

$ git grep 'numpy[=<]'
comps/animation/src/requirements.txt:numpy==1.23.5
comps/language_detection/src/requirements.txt:numpy==1.26.4
comps/llms/src/text-generation/Dockerfile.intel_hpu:    pip install --no-cache-dir --upgrade --force-reinstall pydantic numpy==1.25
comps/text2cypher/src/Dockerfile.intel_hpu:    pip install --no-cache-dir --upgrade --force-reinstall pydantic numpy==1.26.3
comps/third_parties/llama-vision/src/requirements.txt:numpy<2
comps/third_parties/llama-vision/src/requirements_tp.txt:numpy<2

(And why fixed 1.x versions are used on some other requirements.txt files, as can be seen above, instead of <2 ? Now that main numpy development is for 2.x versions, I think that would just mean getting security update versions...)

[lianhao]

This is fixed by #1789