license, license_name, license_link, language, base_model, datasets, pipeline_tag, library_name, tags

license	license_name	license_link	language	base_model	datasets	pipeline_tag	library_name	tags
other	qwen-research	https://huggingface.co/Qwen/Qwen2.5-VL-3B-Instruct/blob/main/LICENSE	en	Qwen/Qwen2.5-VL-3B-Instruct	maveryn/trace	image-text-to-text	transformers	multimodal visual-reasoning reinforcement-learning grpo rlvr trace

TRACE Qwen2.5-VL 3B

TRACE Qwen2.5-VL 3B is a GRPO checkpoint derived from Qwen/Qwen2.5-VL-3B-Instruct and trained on 64,000 grounded visual-reasoning examples spanning 1,000 tasks from maveryn/trace. This repository contains the merged step-500 inference checkpoint.

Paper · Project page · GitHub · Collection · Training configuration · Evaluation suite · Run artifacts

Training and provenance

Field	Value
Base model	Qwen/Qwen2.5-VL-3B-Instruct@66285546d2b821cf421d4f5eb2576359d3770cd3
Dataset	maveryn/trace@4e5b54361360296a855542b40cfd8b7f81b355fe
Method	GRPO, 500 steps, global/rollout batch 128, 8 responses per prompt
Optimization	AdamW, learning rate 1e-6, constant schedule, KL disabled
Reward	0.95 × exact JSON answer + 0.05 × valid JSON format; annotation reward 0
Reference run	8 NVIDIA H100 80GB GPUs
Released artifact	Merged step-500 inference weights

The released training profile reads prompt_answer, scores answer_gt, and does not use the advisory trace_supervision_mode column. The consumed fields, embedded image bytes, and row order in the current dataset release were verified identical to the original training input in the public equivalence receipt.

The canonical output ends with {"answer": ...}. Exact hashes, source revision, and run provenance are recorded in trace_training_provenance.json and .trace_model_revision.json. The repository does not include optimizer, scheduler, trainer, or FSDP state for continuation.

Evaluation

trace_eval_v1 evaluates 24 external benchmarks and 32,805 examples per model and decoding seed. Scores below are the unweighted macro mean of the 24 benchmark percentages, reported as mean ± sample standard deviation across seeds 42, 43, and 44.

Model	Overall score
Qwen2.5-VL-3B-Instruct	39.34 ± 0.63
TRACE Qwen2.5-VL 3B	42.85 ± 0.39
Paired improvement	+3.51 ± 0.25

Per-benchmark scores, model revisions, aggregation, and benchmark provenance are available in the public results.json and evaluation documentation.

Usage

from transformers import AutoProcessor, Qwen2_5_VLForConditionalGeneration

model_id = "maveryn/trace-qwen2.5-vl-3b"
revision = "2ec2374d5c219e6b12e26bda93d3b3adeb1e30c5"
image_url = "https://raw.githubusercontent.com/maveryn/trace/main/docs/assets/paper-domain-montage/trace-paper-domain-montage.png"

processor = AutoProcessor.from_pretrained(model_id, revision=revision)
model = Qwen2_5_VLForConditionalGeneration.from_pretrained(
    model_id,
    revision=revision,
    dtype="auto",
    device_map="auto",
)

messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": image_url},
            {"type": "text", "text": "How many visual domains are shown with example images? Respond with only a JSON object using the key \"answer\"."},
        ],
    }
]
inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    return_dict=True,
    return_tensors="pt",
).to(model.device)
output_ids = model.generate(**inputs, max_new_tokens=128)
generated_ids = [
    output[len(prompt) :] for prompt, output in zip(inputs.input_ids, output_ids)
]
print(processor.batch_decode(generated_ids, skip_special_tokens=True)[0])

Verified reference answer

For the published montage used above, the expected output is:

{"answer": 11}

The value is backed by the committed montage manifest (layout.panel_count). This is reference ground truth, not a claim about a particular decoding run; publish qualitative model outputs only with their recorded inference settings.

The pinned revision is the checkpoint used by the canonical evaluation; later repository heads may update documentation without changing the model weights.

Intended use and limitations

This checkpoint is intended for research on multimodal reasoning and verifiable-reward post-training. It can produce incorrect answers, unsupported reasoning, or unreliable grounding. It was trained on synthetic tasks and has not been validated for safety-critical, medical, legal, or autonomous decision-making uses. Results depend on the recorded prompts, decoding, parsers, and scorers.

License

This checkpoint is subject to the upstream Qwen Research License.