Model Card for T5-GenQ-T-v1

🤖 ✨ 🔍 Generate precise, realistic user-focused search queries from product text 🛒 🚀 📊

Model Description

Model Name: Fine-Tuned Query-Generation Model
Model type: Text-to-Text Transformer
Finetuned from model: BeIR/query-gen-msmarco-t5-base-v1
Dataset: smartcat/Amazon-2023-GenQ
Primary Use Case: Generating accurate and relevant search queries from item descriptions
Repository: smartcat-labs/product2query

Model variations

Model	ROUGE-1	ROUGE-2	ROUGE-L	ROUGE-Lsum
T5-GenQ-T-v1	75.2151	54.8735	74.5142	74.5262
T5-GenQ-TD-v1	78.2570	58.9586	77.5308	77.5466
T5-GenQ-TDE-v1	76.9075	57.0980	76.1464	76.1502
T5-GenQ-TDC-v1 (best)	80.0754	61.5974	79.3557	79.3427

Uses

This model is designed to improve e-commerce search functionality by generating user-friendly search queries based on product descriptions. It is particularly suited for applications where product descriptions are the primary input, and the goal is to create concise, descriptive queries that align with user search intent.

Examples of Use:

Generating search queries for product indexing.

Enhancing product discoverability in e-commerce search engines.

Automating query generation for catalog management.

Comparison of ROUGE scores:

Model	ROUGE-1	ROUGE-2	ROUGE-L	ROUGE-Lsum
T5-GenQ-T-v1	73.11	52.27	72.51	72.51
query-gen-msmarco-t5-base-v1	40.34	19.52	39.21	39.21

Note: This evaluation is done after training, based on the test split of the smartcat/Amazon-2023-GenQ dataset.

Examples

Expand to see table with examples

Input Text	Target Query	Before Fine-tuning	After Fine-tuning
PANDORA Jewelry Crossover Pave Triple Band Ring for Women - Sterling Silver with Cubic Zirconia	PANDORA Crossover Triple Band Ring	what is pandora jewelry	Pandora crossover ring
SAYOYO Baby Sneakers Leather Baby Shoes Crib Shoes Toddler Soft Sole Sneakers	SAYOYO Baby Sneakers	what kind of shoes are baby sneakers	baby leather sneakers
5 PCS Strap Replacement Compatible with Xiaomi Mi Band 3/4, Bands Xiaomi Mi Band 4 Smart Watch Wristbands Replacement Accessories Strap Bracelets for Mi Fit 3 Straps	Replacement Straps for Xiaomi Mi Band 3/4p	what is the strap on a xiaomi smartwatch	Xiaomi Mi Fit 3 replacement bands
Backpacker Ladies' Solid Flannel Shirt	ladies flannel shirt	what kind of shirt is a backpacker	women's flannel shirt

How to Get Started with the Model

Use the code below to get started with the model.

from transformers import AutoModelForSeq2SeqLM, AutoTokenizer

model = AutoModelForSeq2SeqLM.from_pretrained("smartcat/T5-GenQ-T-v1")
tokenizer = AutoTokenizer.from_pretrained("smartcat/T5-GenQ-T-v1")

description = "Silver-colored cuff with embossed braid pattern. Made of brass, flexible to fit wrist."

inputs = tokenizer(description, return_tensors="pt", padding=True, truncation=True)
generated_ids = model.generate(inputs["input_ids"], max_length=30, num_beams=4, early_stopping=True)

generated_text = tokenizer.decode(generated_ids[0], skip_special_tokens=True)

Training Details

Training Data

The model was trained on the smartcat/Amazon-2023-GenQ dataset, which consists of user-like queries generated from product descriptions. The dataset was created using Claude Haiku 3, incorporating key product attributes such as the title, description, and images to ensure relevant and realistic queries. For more information, read the Dataset Card. 😊

Preprocessing

Trained on only the product titles
Tokenized using T5’s default tokenizer with truncation to handle long text.

Training Hyperparameters

max_input_length: 512
max_target_length: 30
batch_size: 48
num_train_epochs: 8
evaluation_strategy: epoch
save_strategy: epoch
learning_rate: 5.6e-05
weight_decay: 0.01
predict_with_generate: true
load_best_model_at_end: true
metric_for_best_model: eval_rougeL
greater_is_better: true
logging_strategy: epoch

Train time: 2.43 hrs

Hardware

A6000 GPU:

Memory Size: 48 GB
Memory Type: GDDR6
CUDA: 8.6

Metrics

ROUGE, or Recall-Oriented Understudy for Gisting Evaluation, is a set of metrics used for evaluating automatic summarization and machine translation in NLP. The metrics compare an automatically produced summary or translation against a reference or a set of references (human-produced) summary or translation. ROUGE metrics range between 0 and 1, with higher scores indicating higher similarity between the automatically produced summary and the reference.

In our evaluation, ROUGE scores are scaled to resemble percentages for better interpretability. The metric used in the training was ROUGE-L.

Epoch	Step	Loss	Grad Norm	Learning Rate	Eval Loss	ROUGE-1	ROUGE-2	ROUGE-L	ROUGE-Lsum
1.0	4285	0.9465	6.7834	4.9e-05	0.7644	73.1872	52.2019	72.5199	72.5183
2.0	8570	0.8076	4.9071	4.2e-05	0.7268	73.9182	53.1365	73.2551	73.2570
3.0	12855	0.7485	4.4814	3.5e-05	0.7160	74.4752	53.8076	73.7712	73.7792
4.0	17140	0.7082	5.3145	2.8e-05	0.7023	74.7628	54.3316	74.0811	74.0790
5.0	21425	0.6788	4.4266	2.1e-05	0.7013	74.9437	54.5630	74.2637	74.2668
6.0	25710	0.6561	5.2897	1.4e-05	0.6998	75.0834	54.7163	74.3907	74.3977
7.0	29995	0.6396	3.5197	7.0e-06	0.7005	75.2151	54.8735	74.5142	74.5262
8.0	34280	0.6278	4.4625	0.0	0.7016	75.1899	54.8423	74.4695	74.4801

Model Analysis

Average scores by model

The checkpoint-29995 (T5-GenQ-T-v1) model outperforms query-gen-msmarco-t5-base-v1 across all ROUGE metrics.

The largest performance gap is in ROUGE2, where checkpoint-29995 achieves 52.27, whereas query-gen-msmarco-t5-base-v1 scores 19.52.

ROUGE1, ROUGEL, and ROUGELSUM scores are very similar in both trends, with checkpoint-29995 consistently scoring above 72, while query-gen-msmarco-t5-base-v1 stays below 41.

Density comparison

T5-GenQ-T-v1 - Higher concentration of high ROUGE scores, especially near 100%, indicating strong text overlap with references.

query-gen-msmarco-t5-base-v1 – more spread-out distribution, with multiple peaks at 10-40%, suggesting greater variability but lower precision.

ROUGE-1 & ROUGE-L: T5-GenQ-T-v1 peaks at 100%, while query-gen-msmarco-t5-base-v1 has lower, broader peaks.

ROUGE-2: query-gen-msmarco-t5-base-v1 has a high density at 0%, indicating many low-overlap outputs.

Histogram comparison

T5-GenQ-T-v1 – higher concentration of high ROUGE scores, especially near 100%, indicating strong text overlap with references.

query-gen-msmarco-t5-base-v1 – more spread-out distribution, with peaks in the 10-40% range, suggesting greater variability but lower precision.

ROUGE-1 & ROUGE-L: T5-GenQ-T-v1 shows a rising trend towards higher scores, while query-gen-msmarco-t5-base-v1 has multiple peaks at lower scores.

ROUGE-2: query-gen-msmarco-t5-base-v1 has a high concentration of low-score outputs, whereas T5-GenQ-T-v1 achieves more high-scoring outputs.

Scores by generated query length

This visualization analyzes average ROUGE scores and score differences across different query sizes.

High ROUGE Scores for Most Sizes (3-9 words).

ROUGE-1, ROUGE-2, ROUGE-L, and ROUGE-LSUM scores remain consistently high across most word sizes.

Sharp Spike at Size 2:

A large positive score difference at 2 words, suggesting strong alignment for very short phrases.

Stable Score Differences (Sizes 3-9):

After the initial spike at size 2, score differences stay close to zero, indicating consistent performance across phrase lengths.

Semantic similarity distribution

This histogram visualizes the distribution of cosine similarity scores, which measure the semantic similarity between paired texts.

The majority of similarity scores cluster near 1.0, indicating that most text pairs are highly similar.

A gradual increase in frequency is observed as similarity scores rise, with a sharp peak at 1.0.

Lower similarity scores (0.0–0.4) are rare, suggesting fewer instances of dissimilar text pairs.

Semantic similarity score against ROUGE scores

This scatter plot matrix compares semantic similarity (cosine similarity) with ROUGE scores, showing their correlation.

Higher similarity → Higher ROUGE scores, indicating strong n-gram overlap in semantically similar texts. ROUGE-1 & ROUGE-L show the strongest correlation, while ROUGE-2 has more variability. Low-similarity outliers exist, where texts share words but differ semantically.

More Information

Please visit the GitHub Repository

Authors

Mentor: Milutin Studen
Engineers: Petar Surla, Andjela Radojevic

Model Card Contact

For questions, please open an issue on the GitHub Repository

smartcat
/

T5-GenQ-T-v1