Datasets:

bdatm-project
/

evaluation-results-task1

model_name stringlengths 25 56	fen_exact_match float64 0 0.03	character_error_rate float64 0.04 3.38	square_by_square_accuracy float64 0 1	levenshtein_distance float64 2.23 177	adapter_sha stringlengths 40 64 ⌀	board_exact_match float64 0 0.99	board_character_error_rate float64 0 4.31	side_to_move_match float64 0 0.66	castling_match float64 0 1	en_passant_match float64 0 1	clocks_match float64 0 0.05
Vanilla Qwen2.5-VL (Zero-Shot)	0	3.259093	0	169.225	null	0	3.867622	0	0.0025	0	0
Qwen + LoRA (Zigzag - TF)	0	0.314656	0.500781	16.45	df1de1e657acce1785c7604a3206a4ec22b1a28f	0	0.340622	0.5925	0.945	0.99	0.0375
Qwen + LoRA (Spiral - TF)	0	0.621696	0.379414	33.0075	df1de1e657acce1785c7604a3206a4ec22b1a28f	0	0.722563	0.545	0.8925	0.9925	0.0125
Qwen + LoRA (File_wise - TF)	0	0.653481	0.570391	34.98	df1de1e657acce1785c7604a3206a4ec22b1a28f	0	0.762722	0.505	0.8775	0.9975	0.01
Qwen + LoRA (Zigzag - SFT)	0.0125	0.044541	0.974492	2.2975	2e7358f780fb18aca54b0b96e626811ef9b34335	0.95	0.001425	0.6075	0.995	0.9975	0.0325
Qwen + LoRA (Spiral - SFT)	0.0125	0.046081	0.99668	2.385	cb9ddf3ac519846bd91e962711ff4667bc7a1573	0.9625	0.003055	0.6175	0.9775	0.995	0.0225
Qwen + LoRA (File_wise - SFT)	0.0225	0.04463	0.999648	2.305	27bf9ae48a8332ef6fe9d1ad70e6a5954fa38359	0.9775	0.000713	0.615	0.985	0.9975	0.0425
Qwen + LoRA (Learned Reordering)	0.02	0.055516	0.981055	2.9025	04e3799b675534ad419bb77505afeb670301fbe2	0.755	0.015249	0.6225	0.985	0.995	0.0475
Vanilla Qwen3.5-0.8B (Zero-Shot) [tok256]	0	3.381385	0	177.2625	null	0	4.308628	0.0025	0.0025	0.0025	0
Vanilla Qwen3.5-0.8B (3-Shot) [tok256]	0	0.697983	0.085508	36.6075	null	0	0.761413	0.5025	0.2975	0.9925	0.01
Qwen + LoRA (Raster - SFT) [tok256]	0.0225	0.042988	0.984766	2.225	6da2be8cc6fe54aae0332f0985021e7750ebc3cd7b9fa0bdd10fee0310008ff4	0.97	0.000838	0.66	0.98	0.9975	0.04
Qwen + LoRA Raster, read zigzag (TF) [tok256]	0	0.512802	0	27.4575	6da2be8cc6fe54aae0332f0985021e7750ebc3cd7b9fa0bdd10fee0310008ff4	0	0.553607	0.44	0.7875	0.86	0.0175
Qwen + LoRA Raster, read spiral_center_out (TF) [tok256]	0	0.78379	0	42.195	6da2be8cc6fe54aae0332f0985021e7750ebc3cd7b9fa0bdd10fee0310008ff4	0	0.889686	0.38	0.7125	0.7875	0
Qwen + LoRA Raster, read file_wise (TF) [tok256]	0	0.793535	0.006289	42.6175	6da2be8cc6fe54aae0332f0985021e7750ebc3cd7b9fa0bdd10fee0310008ff4	0	0.905924	0.355	0.6575	0.735	0.005
Qwen + LoRA Raster, read hilbert (TF) [tok256]	0	0.718401	0.007695	38.7075	6da2be8cc6fe54aae0332f0985021e7750ebc3cd7b9fa0bdd10fee0310008ff4	0	0.792901	0.4375	0.7175	0.8025	0.02
Qwen + LoRA (Zigzag - SFT) [tok256]	0.02	0.044088	0.997031	2.2875	b5567af702189fa03bab77522a8cc9c19023c63d812d801cd212deb59d5f4858	0.975	0.00083	0.61	0.98	0.9975	0.035
Qwen + LoRA (Spiral_center_out - SFT) [tok256]	0.01	0.044427	0.999727	2.3025	538810b337ea03d2be6258cee094ddca57c3ffefbd406e4a55366a6caabb1288	0.985	0.000563	0.605	0.985	0.9975	0.0275
Qwen + LoRA (File_wise - SFT) [tok256]	0.02	0.043646	0.99207	2.26	f944b66181d2a7eb393c7dd526d5863aef7bcf71c1882be0624cf1fe4b99b994	0.9675	0.002206	0.62	0.985	0.9975	0.0325
Qwen + LoRA (Hilbert - SFT) [tok256]	0.025	0.043923	0.97707	2.28	5c6319fc48e7db0a4b466d7f1c8e7e9f626fc6bee65037b99ef81149951b684b	0.9525	0.001315	0.635	0.9775	0.9975	0.0375
Qwen + LoRA (Learned Reordering) [tok256]	0.0175	0.044122	0.996602	2.295	997e28ac3871e51ecfe2028ad01fd77037d3a9890c29c4b508c718dedc55cf34	0.9575	0.00155	0.6475	0.9725	0.9975	0.0325
Qwen + LoRA (Learned Reordering, frozen policy) [tok256]	0.0225	0.043736	0.994531	2.2625	61c6f70943b3a9a2d8a3c5137cc47658955aa33e0f9b983a97b93ec17baaf38e	0.9675	0.000881	0.6525	0.9875	0.9975	0.04
Qwen + LoRA Fine-Tuning (from HF)	0.0175	0.043973	0.994922	2.2775	e3b1d9d57c37a439bbddd86498db6b5304612c74cf074c2ea6874f2691f832ae	0.99	0.000617	0.62	0.9925	0.9975	0.0375

Model-comparison table for this task: one row per evaluated model, written by push_results_table in src/eval/results.py. Per-sample predictions are in per_sample/.

Downloads last month: 259

Number of rows:

Total file size:

2.16 MB