---
title: seta-sft-kimi-k2.5-nothink
canonical_url: "https://www.modelscope.cn/datasets/camel-ai/seta-sft-kimi-k2.5-nothink"
md_url: "https://www.modelscope.cn/datasets/camel-ai/seta-sft-kimi-k2.5-nothink.md"
repository: camel-ai/seta-sft-kimi-k2.5-nothink
last_updated: 2026-04-19
license: apache-2.0
storage_size: "250 MB"
downloads: 34
stars: 0
---

# seta-sft-kimi-k2.5-nothink

> seta-sft-kimi-k2.5-nothink - camel-ai 在 ModelScope 开源的数据集。Seta SFT — Kimi K2.5 (no-thinking)

camel-ai/seta-sft-kimi-k2.5-nothink 是 ModelScope 魔搭社区上的数据集，存储大小 250 MB，采用 apache-2.0 许可。

- **Repository**: camel-ai/seta-sft-kimi-k2.5-nothink
- **License**: apache-2.0
- **Storage size**: 250 MB
- **Downloads**: 34
- **Stars**: 0
- **Last updated**: 2026-04-19

Source: https://www.modelscope.cn/datasets/camel-ai/seta-sft-kimi-k2.5-nothink

---

# Seta SFT — Kimi K2.5 (no-thinking)

Supervised fine-tuning dataset distilled from 1488 successful
agent rollouts of **moonshot/kimi-k2.5** on the
[seta-env-v2](https://huggingface.co/datasets/camel-ai/seta-env-v2)
terminal-agent benchmark, tokenized with the **Qwen/Qwen3-8B** chat template
and ready for AREAL `FSDPLMEngine` SFT training.

## Schema

Each row preserves the full per-trial diagnostic record from the build
pipeline so consumers can inspect, filter, or re-tokenize without rerunning
the rollouts:

| column | type | meaning |
|---|---|---|
| `task_id`               | str       | task name (without `_t<i>_<hash>` suffix) |
| `trial_uid`             | str       | full trial dir name |
| `reward`                | float     | passed/total tests from `verifier/ctrf.json` |
| `model`                 | str       | rollout model id (e.g. `kimi-k2.5`) |
| `conv_json_path`        | str       | source conv json path on the build machine (provenance) |
| `provider_token_counts` | dict      | LLM API counts: `prompt_tokens`, `completion_tokens`, `total_tokens`, `cached_tokens` |
| `local_token_count`     | int       | `len(input_ids)` under the Qwen3-8B tokenizer |
| `n_assistant_tokens`    | int       | `sum(loss_mask)` |
| `n_messages`            | int       | reconstructed conversation length (incl. final assistant turn) |
| `raw_conv_json`         | str       | full original conv json, serialized |
| `chat_template_str`     | str       | exact rendered template that was tokenized |
| `input_ids`             | list[int] | tokenized full conversation |
| `loss_mask`             | list[int] | same length, `1` on every assistant span (incl. content, and `<tool_call>...</tool_call>`), `0` on system / user / tool-response / padding |

For AREAL training, the loader
[`scripts/areal_sft/seta_sft_dataset.py::get_seta_sft_dataset`](https://github.com/camel-ai/terminal_agent/blob/areal_dev/scripts/areal_sft/seta_sft_dataset.py)
projects each row to `(input_ids, loss_mask)` only — the trainer's
`pad_sequences_to_tensors` collator consumes nothing else.

## Statistics

| | value |
|---|---|
| rows                  | **1,488** |
| total tokens          | **13,177,096** |
| trainable tokens      | **5,271,764** (40.0%) |
| tokens / row (mean)   | 8856 |
| tokens / row (median) | 7434 |
| tokens / row (max)    | 24,767 |
| reward = 1.0 rows     | 1,112 (74.7%) |
| reward (mean)         | 0.932 |

## Provenance & processing

1. **Rollouts.** moonshot/kimi-k2.5 was run via the TITO (Token-In-Token-Out)
   agent on every task in `seta-env-v2` (1 trajectory per task, max 200
   iterations, 28 672 max total tokens, 4 096 max completion tokens,
   temperature 1.0, thinking enabled). Tasks that crashed or hit rate limits
   were re-run via `eval.py --resume` until convergence (4 resume passes).
2. **Merging.** Resume chain merged into one canonical dir via
   `seta_env.utils.collect_results --merge --collect-trials move` —
   1605 trial subdirs total, 18 fully-failed tasks, 1488 successfully
   produced both `verifier/ctrf.json` and a conversation snapshot in
   `CAMEL_LOG_DIR/`.
3. **Tokenization & loss-mask construction.**
   `seta_env.utils.sft_utils.build_sft_dataset` walks each trial dir, picks
   the largest `CAMEL_LOG_DIR/.../conv_*.json` (the most-complete snapshot),
   reconstructs `request.messages + response.choices[0].message`, applies
   the Qwen3-8B chat template, and builds the per-token loss mask via a
   token-stream scan that marks every `<|im_start|>assistant ... <|im_end|>\n`
   span as trainable. Boundary tokens (`<|im_start|>`, `<|im_end|>`,
   `<think>`, `</think>`, `<tool_call>`, trailing `\n`) are all attributed
   to the correct side of the mask — verified by per-trial inspection
   artifacts under the build script's `--debug` mode.
4. **Filtering.** Trials whose `verifier/ctrf.json` was missing entirely
   (errored or timed out before verification ran) were dropped — 117 of the
   1605 trial subdirs. Trials with any verified reward (including 0/N
   rollouts) were kept; downstream consumers can further filter on `reward`
   directly using the column on this dataset.
5. **Thinking handling.** Reasoning traces have been **pruned**: assistant turns contain only the visible content + tool calls, no `<think>...</think>` blocks. To match the inference-time distribution, the system prompt has been suffixed with `\n/no_think` (the Qwen3 directive that suppresses thinking).

## Companion datasets

- [camel-ai/seta-sft-kimi-k2.5-thinking](https://huggingface.co/datasets/camel-ai/seta-sft-kimi-k2.5-thinking) — same
  rollouts, thinking variant. Useful for ablations.

## Loading

For inspection / custom processing (full row):
```python
from datasets import load_dataset
ds = load_dataset("camel-ai/seta-sft-kimi-k2.5-nothink", split="train")
print(ds[0]["task_id"], ds[0]["reward"], ds[0]["local_token_count"])
print(ds[0]["chat_template_str"][:500])
```

For AREAL-trainer-ready loading (projected to `(input_ids, loss_mask)`):
```python
from seta_sft_dataset import get_seta_sft_dataset
ds = get_seta_sft_dataset("camel-ai/seta-sft-kimi-k2.5-nothink", split="train", max_length=16384)
# Dataset({features: ['input_ids', 'loss_mask'], num_rows: ...})
```

## License

Apache 2.0 (matches the upstream `camel-ai/seta-env-v2` license).