Fix duplicate health server startup in ModelRunner

[ackizilkale]

Fix duplicate health server startup in ModelRunner

start_health_server_thread() was called twice on model load — once
unconditionally and once inside the
guard. The first call would bind the port successfully, causing the
second call to fail with OSError: [Errno 98] Address already in use.

Remove the unconditional call, keeping only the guarded one which
also correctly respects health_check_port=None to disable the server.

• Thread-2 (first call) bound port 8080 successfully and kept running — so health probes (/healthz/liveness, /healthz/readiness) still responded correctly
• Thread-3 (second call) crashed silently on startup and died

[github-actions]

Package	Line Rate	Health
clarifai	45%	❌
clarifai.cli	68%	➖
clarifai.cli.templates	46%	❌
clarifai.client	65%	➖
clarifai.client.auth	67%	➖
clarifai.constants	100%	✔
clarifai.datasets	100%	✔
clarifai.datasets.export	80%	✔
clarifai.datasets.upload	75%	➖
clarifai.datasets.upload.loaders	37%	❌
clarifai.models	100%	✔
clarifai.rag	0%	❌
clarifai.runners	53%	➖
clarifai.runners.models	64%	➖
clarifai.runners.pipeline_steps	39%	❌
clarifai.runners.pipelines	71%	➖
clarifai.runners.utils	62%	➖
clarifai.runners.utils.data_types	72%	➖
clarifai.schema	100%	✔
clarifai.urls	58%	➖
clarifai.utils	60%	➖
clarifai.utils.evaluation	16%	❌
clarifai.workflows	95%	✔
Summary	62% (10027 / 16299)	➖

Minimum allowed line rate is 50%