model
stringclasses 3
values | rouge1
float64 0.04
0.13
| rouge2
float64 0.01
0.06
| rougeL
float64 0.03
0.09
| normalized_levenshtein
float64 0.12
0.13
| levenshtein_distance
float64 5.44k
5.9k
| num_examples
int64 42
42
| evaluation_date
stringdate 2025-12-09 10:56:10
2025-12-09 11:14:26
|
|---|---|---|---|---|---|---|---|
unsloth/Qwen3-0.6B-GGUF
| 0.055981
| 0.009523
| 0.044265
| 0.124479
| 5,890.880952
| 42
|
2025-12-09T10:56:10.311010
|
espsluar/crawlerlm-qwen3-0.6b-test
| 0.041047
| 0.011858
| 0.030611
| 0.125309
| 5,902.238095
| 42
|
2025-12-09T11:10:40.583629
|
espsluar/qwen-crawlerlm-sft
| 0.125857
| 0.055374
| 0.090374
| 0.134227
| 5,444.857143
| 42
|
2025-12-09T11:14:26.851934
|
YAML Metadata
Warning:
empty or missing yaml metadata in repo card
(https://huggingface.co/docs/hub/datasets-cards)
CrawlerLM Evaluation Results
This dataset contains evaluation metrics for models tested on the CrawlerLM HTML-to-JSON extraction task.
Current Results
| Model | ROUGE-1 | ROUGE-2 | ROUGE-L | Norm. Lev. | Lev. Dist |
|---|---|---|---|---|---|
| unsloth/Qwen3-0.6B-GGUF | 0.0560 | 0.0095 | 0.0443 | 0.1245 | 5890.88 |
| espsluar/crawlerlm-qwen3-0.6b-test | 0.0410 | 0.0119 | 0.0306 | 0.1253 | 5902.24 |
| espsluar/qwen-crawlerlm-sft | 0.1259 | 0.0554 | 0.0904 | 0.1342 | 5444.86 |
Dataset Structure
model: HuggingFace model ID (namespace/model-name)rouge1: ROUGE-1 F1 score (higher is better)rouge2: ROUGE-2 F1 score (higher is better)rougeL: ROUGE-L F1 score (higher is better)normalized_levenshtein: Normalized Levenshtein similarity (0-1, higher is better)levenshtein_distance: Average Levenshtein edit distance (lower is better)num_examples: Number of test examples evaluatedevaluation_date: ISO timestamp of evaluation
Usage
from datasets import load_dataset
ds = load_dataset("espsluar/crawlerlm-eval-results")
print(ds)
Test Dataset
All models are evaluated on: espsluar/crawlerlm-html-to-json (test split)
- Downloads last month
- 18