Add downstream task evaluation script for llama3_native_te recipe

Adds eval_downstream.py that runs lm-eval benchmarks (arc_challenge,
arc_easy, boolq, copa, hellaswag, piqa, winogrande) on trained Lingua
1B checkpoints. Supports safetensors, distributed FSDP2 (DCP), and DDP
checkpoint formats. Self-contained checkpoint loading avoids TE version
compatibility issues with checkpoint.py imports.

Made-with: Cursor

Author: savitha-eng