[FX] Support weight quantization for operations where `weight_port_id` != 1

[siddhant-0707]

Changes

Updated the FX backend’s _get_input_scale_shape to use the FX insertion point shape and, when available, the actual weight tensor’s shape to compute the per‑channel scale shape. Adjusted statistics collector in _get_stat_collector so that the reduction and aggregation axes are derived using the same channel axes as used for scale shape computation.

Related tickets

Issue #3206

Tests

All tests run successfully

[anzr299]

Please check the failing pre-commit test.