Fix Hydra 1.1 upgrade issue for Fast ACUTEs

[EricMichaelSmith]

Patch description
When running Fast ACUTEs, the Hydra flags in base_fast_acute.yaml (such as mephisto.task.maximum_units_per_worker) were not being applied. This PR fixes this issue, and also prints out all actual values of Hydra flags for easier debugging of this kind of issue in the future.

Testing steps

• CI checks
• Running Fast ACUTEs to check, before and after, that the flags are now respected:

python parlai/crowdsourcing/tasks/acute_eval/fast_eval.py \
mephisto.blueprint.acute_eval_type=engaging \
mephisto.blueprint.config_path=${CONFIG_PATH} \
mephisto.blueprint.model_pairs=${MODEL_PAIRS} \
mephisto.blueprint.onboarding_path=${REPO_FOLDER}/parlai/crowdsourcing/tasks/acute_eval/task_config/onboarding.json \
mephisto.blueprint.root_dir=${ROOT_DIR} \
mephisto.blueprint.matchups_per_pair=500 \
mephisto.task.task_name=${TASK_NAME}