Drop overlength SFT tensors before backend training

[Kovbo]

Fixes SFT failures caused by examples longer than the configured training context length.

Changes:

• Drop overlength SFT trajectories inside tokenize_sft_batch.
• Emit a console warning when trajectories are dropped.
• Report num_dropped_trajectories from Unsloth and Megatron SFT metrics.
• Show dropped trajectory count in ART SFT progress output.
• Allow empty SFT batches as no-op batches, which can happen when all trajectories in a batch are dropped.