fix torch.mean(auc) and updating model init in utils.py

[ParthaPratimBanik]

Fixed issue #85

• The argument of torch.mean(), toch.std(), and torch.max() should be a tensor float type instead of list.
• eval_recall_at_k() return tensor float type value, it must convert to list type float value to append into the rec list.
  
  
  Fixed issue Inconsistency with BOND paper #83
• Remove heuristic selection of weight.
• Update the parameter name for detectors (DOMINANT, GAAN, and CONAD) from alpha to weight.

Issues #83, #85, and the above solutions are tested on torch 2.0.1+cu117.

[coveralls]

Pull Request Test Coverage Report for Build 5890522716

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 99.562%

---

Totals
Change from base Build 5709379320:	0.0%
Covered Lines:	2503
Relevant Lines:	2514

---

💛 - Coveralls

[kayzliu]

Thanks for your PR. It's over all looks good. But several minor things need to be changed.

[kayzliu]

LGTM.