Source symbols: scripts/chat_eval.py

Deterministic, value-free symbol metadata extracted from public source. Implementations and literal values are not copied into this page.

• Revision: master@92d63d4e8bb4df75c3b71618f31ddde2378b2bcd · 2026-07-04T18:18:25Z
• Declarations: 3
• Referenced calls: 43

Declared symbols

• run_generative_eval — def at line 28; references get_dist_info, get_device, len, min, range, render_for_completion, generate_batch, decode, evaluate, any, int, print, tensor, all_reduce, item, print0, get_bos_token_id, ceil_div
• run_categorical_eval — def at line 87; references get_dist_info, get_device, get_bos_token_id, len, min, ceil_div
• run_chat_eval — def at line 156; references partial

File-level call references

• get_dist_info — line 30
• get_device — line 31
• len — line 33
• min — line 33
• range — line 37
• render_for_completion — line 41
• generate_batch — line 43
• decode — line 52
• evaluate — line 54
• any — line 55
• int — line 59
• print — line 62
• tensor — line 69
• all_reduce — line 71
• item — line 73
• print0 — line 76
• get_bos_token_id — line 91
• ceil_div — line 96
• max — line 107
• no_grad — line 113
• model — line 114
• enumerate — line 120
• encode — line 126
• append — line 129
• argmax — line 134
• partial — line 162
• task_module — line 167
• run_generative_eval — line 170
• run_categorical_eval — line 172
• ValueError — line 174
• ArgumentParser — line 181
• add_argument — line 182
• parse_args — line 193
• autodetect_device_type — line 195
• compute_init — line 196
• load_model — line 198
• Engine — line 199
• split — line 210
• run_chat_eval — line 215
• all — line 230
• items — line 233
• get — line 234
• compute_cleanup — line 240

Source: scripts/chat_eval.py