Add fal.ai inference endpoints

[gokayfem]

Adds fal.ai inference endpoints

[SBrandeis]

Cool stuff @gokayfem !

Would you mind adding tests for these new models? I think one per modality/task would be nice to ensure the inference client is compatible with FAL's APIs.

Tests are defined in HfInference.spec.ts.

To run them, you'll need to add a HF_FAL_AI_KEY to your environment and use the pnpmcommand:

export HF_FAL_AI_KEY="valid-fal-ai-key"
pnpm run test -t 'Fal AI'