---
license: apache-2.0
base_model: Qwen/Qwen3-8B
library_name: transformers
pipeline_tag: text-generation
language:
- en
---

# TabRankNaive

Qwen3-8B, fine-tuned for **single-call generative listwise table reranking**. Given a question and a list of candidate tables, it reads them all in one prompt and returns the full ranking in a single generation — no pairwise scoring, no cross-encoder passes.

![TabRank: question and candidate tables go into a single Qwen3-8B call, which reasons and emits a ranked list](tabrank_task.png)

This checkpoint is the **Answer-Only** variant: fine-tuned to output the ranking directly with no reasoning trace, so it's the fastest of the three objectives at inference.

Related: [TabRank](https://huggingface.co/AdarshSingh7647/TabRank) (same data, reasoning-conditioned, our best method) · [TabRankStandardSFT](https://huggingface.co/AdarshSingh7647/TabRankStandardSFT) (same data, standard CoT SFT).

## Input / output format

**Input** — a chat message with the question followed by each candidate table, labeled `### Table 1`, `### Table 2`, ...:

```text
Question: Which table shows 2022 quarterly revenue by region?

### Table 1
| Region | Q1 2022 | Q2 2022 | Q3 2022 | Q4 2022 |
|---|---|---|---|---|
| North America | 120 | 134 | 128 | 145 |
| Europe | 88 | 91 | 95 | 102 |

### Table 2
| Product | Units Sold | Year |
|---|---|---|
| Widget A | 4200 | 2021 |

### Table 3
| Region | Headcount |
|---|---|
| North America | 340 |
```

**Output** — a single JSON object with the ranked, one-indexed candidate positions, best first:

```json
{"ranked_tables": [1, 3, 2]}
```

Map the numbers back to your own table ids to get the reranked list — position `1` in the output is `### Table 1` from the input, etc.

## Evaluation

This card does not publish a verified results table for the Answer-Only objective on this training mix — earlier published numbers for this checkpoint could not be confirmed against source eval logs, so they're left out rather than risk repeating an error.

For a fully-verified comparison, see [TabRankStandardSFT](https://huggingface.co/AdarshSingh7647/TabRankStandardSFT) and [TabRank](https://huggingface.co/AdarshSingh7647/TabRank), which include in-distribution results (SQA, TAT-QA, HybridQA, TabFact, NQ-Tables) and out-of-distribution results on 7 benchmarks from the [IBM table-text-ir-evaluation suite](https://huggingface.co/collections/ibm-research/table-text-ir-evaluation). Source eval code and logs: [GitHub](https://github.com/AdarshSingh7647/TabRanker).

## Usage with vLLM

```python
from vllm import LLM, SamplingParams
from transformers import AutoTokenizer

repo = "AdarshSingh7647/TabRankNaive"
tok = AutoTokenizer.from_pretrained(repo)
llm = LLM(model=repo, dtype="bfloat16", max_model_len=32768)

system = ("You are a table relevance expert. Given a question and a set of candidate tables "
          "rank them from most to least useful for answering the question. Output exactly "
          "JSON with key ranked_tables.")
user = "Question: ...\n\n### Table 1\n...\n\n### Table 2\n...\n"

msgs = [{"role": "system", "content": system}, {"role": "user", "content": user}]
text = tok.apply_chat_template(msgs, tokenize=False, add_generation_prompt=True)
out = llm.generate([text], SamplingParams(temperature=0.6, top_p=0.95, max_tokens=2048))
print(out[0].outputs[0].text)
```

## Usage with Transformers

```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "AdarshSingh7647/TabRankNaive"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, torch_dtype=torch.bfloat16, device_map="auto")

system = ("You are a table relevance expert. Given a question and a set of candidate tables "
          "rank them from most to least useful for answering the question. Output exactly "
          "JSON with key ranked_tables.")
user = "Question: ...\n\n### Table 1\n...\n\n### Table 2\n...\n"

msgs = [{"role": "system", "content": system}, {"role": "user", "content": user}]
text = tok.apply_chat_template(msgs, tokenize=False, add_generation_prompt=True)
inputs = tok(text, return_tensors="pt").to(model.device)
out = model.generate(**inputs, max_new_tokens=2048, temperature=0.6, top_p=0.95, do_sample=True)
print(tok.decode(out[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))
```

Full training and eval code: [github.com/AdarshSingh7647/TabRanker](https://github.com/AdarshSingh7647/TabRanker).

## Model details

* Base model: Qwen3-8B
* Method: LoRA rank 16, merged into base weights
* Precision: bfloat16, ~16 GB
* Training data: NQ Tables + MultiTabQA

## Citation

```bibtex
@misc{singh2026tabrank,
      title={TabRank: Chain-of-Thought Distillation for Table Re-Rankers},
      author={Adarsh Singh and Kushal Raj Bhandari and Jianxi Gao and Soham Dan and Vivek Gupta},
      year={2026},
      eprint={2607.25182},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2607.25182}
}
```

The MultiTabQA data in this checkpoint's training mix comes from RAG over Tables:

```bibtex
@misc{zou2025ragtableshierarchicalmemory,
      title={RAG over Tables: Hierarchical Memory Index, Multi-Stage Retrieval, and Benchmarking},
      author={Jiaru Zou and Dongqi Fu and Sirui Chen and Xinrui He and Zihao Li and Yada Zhu and Jiawei Han and Jingrui He},
      year={2025},
      eprint={2504.01346},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2504.01346}
}
```