View on GitHub
"```python\nfrom modal_training_gym.common.eval import EvalRowResult\n```\n\nOne model interaction: the prompt, the raw response, its parsed\nstructure (thinking / answer / tool calls), a score, and free-form\nmetadata.\n\nShared by eval rows (`EvalResult.rows`) and training rollout samples\n(`TrainingRolloutResult.samples`) — they were the same shape, so this is\nthe single canonical type for both.\n\n## Constructor\n\n```python\nEvalRowResult(**data)\n```\n\n| Parameter | Type | Default | Description |\n|-----------|------|---------|-------------|\n\n## Attributes\n\n| Attribute | Type | Default | Description |\n|-----------|------|---------|-------------|\n| `score` | `float` | | |\n| `prompt` | `str` | | |\n| `response` | `str` | | |\n| `parsed_response` | `ParsedResponse \\| None` | | |\n| `metadata` | `dict[str, Any]` | | |\n| `trace` | `list[modal_training_gym.common.sample.TraceSpan] \\| None` | | |\n\n## Related Tutorials\n\n- [Qwen3-4B haiku evaluation with verifiable rewards — serve, evaluate, train, compare](/tutorials/rl/000_rl_basics/)\n- [Multi-turn number-guessing RL with custom generate and reward functions](/tutorials/rl/002_multiturn/)\n- [On-policy distillation on math — Qwen3-8B teacher, Qwen3-4B student](/tutorials/rl/003_on_policy_distillation/)\n- [DAPO on math with Qwen3-4B](/tutorials/rl/005_dapo/)\n- [Cross-tokenizer agentic distillation on BFCL v3 multi-turn with live, execution-grounded rewards — DeepSeek V4 Flash teacher, Qwen3.6-35B-A3B student](/tutorials/rl/009_cross_tokenizer_distillation/)\n\n**Source:** [`modal_training_gym/common/eval.py`](https://github.com/modal-projects/training-gym/blob/main/modal_training_gym/common/eval.py)\n"
EvalRowResult
One model interaction: the prompt, the raw response, its parsed structure (thinking / answer / tool calls), a score, and free-form metadata.
from modal_training_gym.common.eval import EvalRowResultOne model interaction: the prompt, the raw response, its parsed structure (thinking / answer / tool calls), a score, and free-form metadata.
Shared by eval rows (EvalResult.rows) and training rollout samples
(TrainingRolloutResult.samples) — they were the same shape, so this is
the single canonical type for both.
Constructor
Section titled “Constructor”EvalRowResult(**data)| Parameter | Type | Default | Description |
|---|
Attributes
Section titled “Attributes”| Attribute | Type | Default | Description |
|---|---|---|---|
score | float | ||
prompt | str | ||
response | str | ||
parsed_response | ParsedResponse | None | ||
metadata | dict[str, Any] | ||
trace | list[modal_training_gym.common.sample.TraceSpan] | None |
Related Tutorials
Section titled “Related Tutorials”- Qwen3-4B haiku evaluation with verifiable rewards — serve, evaluate, train, compare
- Multi-turn number-guessing RL with custom generate and reward functions
- On-policy distillation on math — Qwen3-8B teacher, Qwen3-4B student
- DAPO on math with Qwen3-4B
- Cross-tokenizer agentic distillation on BFCL v3 multi-turn with live, execution-grounded rewards — DeepSeek V4 Flash teacher, Qwen3.6-35B-A3B student