INNE skips `BaseDetector.__init__`, so an out-of-range `contamination` is accepted silently

INNE.__init__ (pyod/models/inne.py:84-90) assigns self.contamination = contamination directly and never calls super().__init__, so the range check in BaseDetector.__init__ (base.py:62-66) is bypassed.

Evidence (v3.6.6):

>>> KNN(contamination=0.9)
ValueError: contamination must be in (0, 0.5], got: 0.900000
>>> INNE(contamination=0.9)
INNE(contamination=0.9, ...)      # accepted

Every other tabular detector rejects this at construction. INNE(contamination=0.9).fit(X) then labels 90% of the training set as outliers.

Proposed fix: super().__init__(contamination=contamination) and drop the direct assignment. One line plus a test. Happy to PR.

Found with a script that checks every detector for the sklearn parameter contract (get_params/clone/refit); drafted with Claude Code assistance and verified by hand.