Sentence Similarity
sentence-transformers
Safetensors
English
grt
geometric-rotation-transformer
feature-extraction
circular-pooling
rotational
qwen3
Instructions to use MatterWaveAI/grt-30m-embedding with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use MatterWaveAI/grt-30m-embedding with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("MatterWaveAI/grt-30m-embedding") sentences = [ "That is a happy person", "That is a happy dog", "That is a very happy person", "Today is a sunny day" ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [4, 4] - Notebooks
- Google Colab
- Kaggle
Update 17-way classical geometric optimizer ablation (add HuygensSymplecticAdam): eval_results.json
8956eb8 verified Download eval_results.json from MatterWaveAI/grt-30m-embedding: direct link, hf CLI and curl.
- Browser
- Download file 12.3 kB
-
https://huggingface.co/MatterWaveAI/grt-30m-embedding/resolve/main/eval_results.json
- Command line
-
hf download hf://MatterWaveAI/grt-30m-embedding/eval_results.json
-
curl -L -o eval_results.json https://huggingface.co/MatterWaveAI/grt-30m-embedding/resolve/main/eval_results.json
12.3 kB
| { | |
| "optimizer": "PairAdam", | |
| "stsb_dev_spearman_before": 0.4551, | |
| "stsb_dev_spearman_after": 0.7179, | |
| "stsb_test_spearman_before": 0.2849, | |
| "stsb_test_spearman_after": 0.6242, | |
| "stsb_test_pearson_before": 0.2981, | |
| "stsb_test_pearson_after": 0.633, | |
| "hessian_pair_adam_ablation": { | |
| "stsb_dev_spearman_after": 0.7278, | |
| "stsb_dev_pearson_after": 0.6917, | |
| "stsb_test_spearman_after": 0.6288, | |
| "stsb_test_pearson_after": 0.6316 | |
| }, | |
| "adamw_ablation": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_test_spearman_after": 0.6317, | |
| "stsb_test_pearson_after": 0.6372 | |
| }, | |
| "parameters": 29305497, | |
| "trainable_parameters": 9891865, | |
| "optimizer_ablations_8way": { | |
| "polar_natural_adam": { | |
| "stsb_dev_spearman_after": 0.7262, | |
| "stsb_dev_pearson_after": 0.702, | |
| "stsb_test_spearman_after": 0.6471, | |
| "stsb_test_pearson_after": 0.6494 | |
| }, | |
| "cyclide_adam": { | |
| "s1_dev_spearman": 0.6255, | |
| "stsb_dev_spearman_after": 0.7368, | |
| "stsb_dev_pearson_after": 0.7179, | |
| "stsb_test_spearman_after": 0.6287, | |
| "stsb_test_pearson_after": 0.6476, | |
| "rocm_elapsed_sec": 380.42 | |
| }, | |
| "hessian_pair_adam_1x_coupling": { | |
| "stsb_dev_spearman_after": 0.7242, | |
| "stsb_dev_pearson_after": 0.6876, | |
| "stsb_test_spearman_after": 0.6337, | |
| "stsb_test_pearson_after": 0.6369 | |
| }, | |
| "rotor_adam": { | |
| "stsb_dev_spearman_after": 0.7322, | |
| "stsb_dev_pearson_after": 0.7099, | |
| "stsb_test_spearman_after": 0.6329, | |
| "stsb_test_pearson_after": 0.6309 | |
| }, | |
| "toroidal_lion": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_dev_pearson_after": 0.697, | |
| "stsb_test_spearman_after": 0.6319, | |
| "stsb_test_pearson_after": 0.6341 | |
| }, | |
| "adamw": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_test_spearman_after": 0.6317, | |
| "stsb_test_pearson_after": 0.6372 | |
| }, | |
| "hessian_pair_adam_3x_coupling": { | |
| "stsb_dev_spearman_after": 0.7278, | |
| "stsb_dev_pearson_after": 0.6917, | |
| "stsb_test_spearman_after": 0.6288, | |
| "stsb_test_pearson_after": 0.6316 | |
| }, | |
| "pair_adam_3x_coupling": { | |
| "stsb_dev_spearman_after": 0.7179, | |
| "stsb_test_spearman_after": 0.6242, | |
| "stsb_test_pearson_after": 0.633 | |
| }, | |
| "toroidal_muon": { | |
| "stsb_dev_spearman_after": 0.7361, | |
| "stsb_dev_pearson_after": 0.7105, | |
| "stsb_test_spearman_after": 0.6197, | |
| "stsb_test_pearson_after": 0.6304 | |
| } | |
| }, | |
| "optimizer_ablations_9way": { | |
| "polar_natural_adam": { | |
| "stsb_dev_spearman_after": 0.7262, | |
| "stsb_dev_pearson_after": 0.702, | |
| "stsb_test_spearman_after": 0.6471, | |
| "stsb_test_pearson_after": 0.6494 | |
| }, | |
| "cyclide_adam": { | |
| "s1_dev_spearman": 0.6255, | |
| "stsb_dev_spearman_after": 0.7368, | |
| "stsb_dev_pearson_after": 0.7179, | |
| "stsb_test_spearman_after": 0.6287, | |
| "stsb_test_pearson_after": 0.6476, | |
| "rocm_elapsed_sec": 380.42 | |
| }, | |
| "hessian_pair_adam_1x_coupling": { | |
| "stsb_dev_spearman_after": 0.7242, | |
| "stsb_dev_pearson_after": 0.6876, | |
| "stsb_test_spearman_after": 0.6337, | |
| "stsb_test_pearson_after": 0.6369 | |
| }, | |
| "rotor_adam": { | |
| "stsb_dev_spearman_after": 0.7322, | |
| "stsb_dev_pearson_after": 0.7099, | |
| "stsb_test_spearman_after": 0.6329, | |
| "stsb_test_pearson_after": 0.6309 | |
| }, | |
| "toroidal_lion": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_dev_pearson_after": 0.697, | |
| "stsb_test_spearman_after": 0.6319, | |
| "stsb_test_pearson_after": 0.6341 | |
| }, | |
| "adamw": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_test_spearman_after": 0.6317, | |
| "stsb_test_pearson_after": 0.6372 | |
| }, | |
| "hessian_pair_adam_3x_coupling": { | |
| "stsb_dev_spearman_after": 0.7278, | |
| "stsb_dev_pearson_after": 0.6917, | |
| "stsb_test_spearman_after": 0.6288, | |
| "stsb_test_pearson_after": 0.6316 | |
| }, | |
| "pair_adam_3x_coupling": { | |
| "stsb_dev_spearman_after": 0.7179, | |
| "stsb_test_spearman_after": 0.6242, | |
| "stsb_test_pearson_after": 0.633 | |
| }, | |
| "toroidal_muon": { | |
| "stsb_dev_spearman_after": 0.7361, | |
| "stsb_dev_pearson_after": 0.7105, | |
| "stsb_test_spearman_after": 0.6197, | |
| "stsb_test_pearson_after": 0.6304 | |
| } | |
| }, | |
| "optimizer_ablations_16way": { | |
| "symplectic_adam": { | |
| "s1_dev_spearman": 0.6198, | |
| "stsb_dev_spearman_after": 0.736, | |
| "stsb_dev_pearson_after": 0.7087, | |
| "stsb_test_spearman_after": 0.6516, | |
| "stsb_test_pearson_after": 0.6564, | |
| "rocm_elapsed_sec": 243.28 | |
| }, | |
| "huygens_symplectic_adam": { | |
| "s1_dev_spearman": 0.6282, | |
| "stsb_dev_spearman_after": 0.7368, | |
| "stsb_dev_pearson_after": 0.709, | |
| "stsb_test_spearman_after": 0.6509, | |
| "stsb_test_pearson_after": 0.6548, | |
| "rocm_elapsed_sec": 236.41 | |
| }, | |
| "polar_natural_adam": { | |
| "stsb_dev_spearman_after": 0.7262, | |
| "stsb_dev_pearson_after": 0.702, | |
| "stsb_test_spearman_after": 0.6471, | |
| "stsb_test_pearson_after": 0.6494 | |
| }, | |
| "gyric_adam": { | |
| "s1_dev_spearman": 0.6221, | |
| "stsb_dev_spearman_after": 0.7304, | |
| "stsb_dev_pearson_after": 0.7023, | |
| "stsb_test_spearman_after": 0.6455, | |
| "stsb_test_pearson_after": 0.6503, | |
| "rocm_elapsed_sec": 419.85 | |
| }, | |
| "huygens_adam": { | |
| "s1_dev_spearman": 0.6383, | |
| "stsb_dev_spearman_after": 0.7295, | |
| "stsb_dev_pearson_after": 0.7035, | |
| "stsb_test_spearman_after": 0.642, | |
| "stsb_test_pearson_after": 0.6469, | |
| "rocm_elapsed_sec": 331.8 | |
| }, | |
| "hessian_pair_adam_1x_coupling": { | |
| "stsb_dev_spearman_after": 0.7242, | |
| "stsb_dev_pearson_after": 0.6876, | |
| "stsb_test_spearman_after": 0.6337, | |
| "stsb_test_pearson_after": 0.6369 | |
| }, | |
| "beltrami_adam": { | |
| "s1_dev_spearman": 0.6337, | |
| "stsb_dev_spearman_after": 0.7363, | |
| "stsb_dev_pearson_after": 0.7068, | |
| "stsb_test_spearman_after": 0.6333, | |
| "stsb_test_pearson_after": 0.6328, | |
| "rocm_elapsed_sec": 321.44 | |
| }, | |
| "hannay_adam": { | |
| "s1_dev_spearman": 0.6115, | |
| "stsb_dev_spearman_after": 0.7257, | |
| "stsb_dev_pearson_after": 0.699, | |
| "stsb_test_spearman_after": 0.6333, | |
| "stsb_test_pearson_after": 0.633, | |
| "rocm_elapsed_sec": 239.35 | |
| }, | |
| "rotor_adam": { | |
| "stsb_dev_spearman_after": 0.7322, | |
| "stsb_dev_pearson_after": 0.7099, | |
| "stsb_test_spearman_after": 0.6329, | |
| "stsb_test_pearson_after": 0.6309 | |
| }, | |
| "toroidal_lion": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_dev_pearson_after": 0.697, | |
| "stsb_test_spearman_after": 0.6319, | |
| "stsb_test_pearson_after": 0.6341 | |
| }, | |
| "adamw": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_test_spearman_after": 0.6317, | |
| "stsb_test_pearson_after": 0.6372 | |
| }, | |
| "poncelet_adam": { | |
| "s1_dev_spearman": 0.6262, | |
| "stsb_dev_spearman_after": 0.7275, | |
| "stsb_dev_pearson_after": 0.6987, | |
| "stsb_test_spearman_after": 0.6315, | |
| "stsb_test_pearson_after": 0.6374, | |
| "rocm_elapsed_sec": 267.34 | |
| }, | |
| "hessian_pair_adam_3x_coupling": { | |
| "stsb_dev_spearman_after": 0.7278, | |
| "stsb_dev_pearson_after": 0.6917, | |
| "stsb_test_spearman_after": 0.6288, | |
| "stsb_test_pearson_after": 0.6316 | |
| }, | |
| "cyclide_adam": { | |
| "s1_dev_spearman": 0.6255, | |
| "stsb_dev_spearman_after": 0.7368, | |
| "stsb_dev_pearson_after": 0.7179, | |
| "stsb_test_spearman_after": 0.6287, | |
| "stsb_test_pearson_after": 0.6476, | |
| "rocm_elapsed_sec": 380.42 | |
| }, | |
| "rindler_adam": { | |
| "s1_dev_spearman": 0.6208, | |
| "stsb_dev_spearman_after": 0.7318, | |
| "stsb_dev_pearson_after": 0.7069, | |
| "stsb_test_spearman_after": 0.6264, | |
| "stsb_test_pearson_after": 0.6353, | |
| "rocm_elapsed_sec": 271.68 | |
| }, | |
| "pair_adam_3x_coupling": { | |
| "stsb_dev_spearman_after": 0.7179, | |
| "stsb_test_spearman_after": 0.6242, | |
| "stsb_test_pearson_after": 0.633 | |
| }, | |
| "toroidal_muon": { | |
| "stsb_dev_spearman_after": 0.7361, | |
| "stsb_dev_pearson_after": 0.7105, | |
| "stsb_test_spearman_after": 0.6197, | |
| "stsb_test_pearson_after": 0.6304 | |
| } | |
| }, | |
| "optimizer_ablations_17way": { | |
| "symplectic_adam": { | |
| "s1_dev_spearman": 0.6198, | |
| "stsb_dev_spearman_after": 0.736, | |
| "stsb_dev_pearson_after": 0.7087, | |
| "stsb_test_spearman_after": 0.6516, | |
| "stsb_test_pearson_after": 0.6564, | |
| "rocm_elapsed_sec": 243.28 | |
| }, | |
| "huygens_symplectic_adam": { | |
| "s1_dev_spearman": 0.6282, | |
| "stsb_dev_spearman_after": 0.7368, | |
| "stsb_dev_pearson_after": 0.709, | |
| "stsb_test_spearman_after": 0.6509, | |
| "stsb_test_pearson_after": 0.6548, | |
| "rocm_elapsed_sec": 236.41 | |
| }, | |
| "polar_natural_adam": { | |
| "stsb_dev_spearman_after": 0.7262, | |
| "stsb_dev_pearson_after": 0.702, | |
| "stsb_test_spearman_after": 0.6471, | |
| "stsb_test_pearson_after": 0.6494 | |
| }, | |
| "gyric_adam": { | |
| "s1_dev_spearman": 0.6221, | |
| "stsb_dev_spearman_after": 0.7304, | |
| "stsb_dev_pearson_after": 0.7023, | |
| "stsb_test_spearman_after": 0.6455, | |
| "stsb_test_pearson_after": 0.6503, | |
| "rocm_elapsed_sec": 419.85 | |
| }, | |
| "huygens_adam": { | |
| "s1_dev_spearman": 0.6383, | |
| "stsb_dev_spearman_after": 0.7295, | |
| "stsb_dev_pearson_after": 0.7035, | |
| "stsb_test_spearman_after": 0.642, | |
| "stsb_test_pearson_after": 0.6469, | |
| "rocm_elapsed_sec": 331.8 | |
| }, | |
| "hessian_pair_adam_1x_coupling": { | |
| "stsb_dev_spearman_after": 0.7242, | |
| "stsb_dev_pearson_after": 0.6876, | |
| "stsb_test_spearman_after": 0.6337, | |
| "stsb_test_pearson_after": 0.6369 | |
| }, | |
| "beltrami_adam": { | |
| "s1_dev_spearman": 0.6337, | |
| "stsb_dev_spearman_after": 0.7363, | |
| "stsb_dev_pearson_after": 0.7068, | |
| "stsb_test_spearman_after": 0.6333, | |
| "stsb_test_pearson_after": 0.6328, | |
| "rocm_elapsed_sec": 321.44 | |
| }, | |
| "hannay_adam": { | |
| "s1_dev_spearman": 0.6115, | |
| "stsb_dev_spearman_after": 0.7257, | |
| "stsb_dev_pearson_after": 0.699, | |
| "stsb_test_spearman_after": 0.6333, | |
| "stsb_test_pearson_after": 0.633, | |
| "rocm_elapsed_sec": 239.35 | |
| }, | |
| "rotor_adam": { | |
| "stsb_dev_spearman_after": 0.7322, | |
| "stsb_dev_pearson_after": 0.7099, | |
| "stsb_test_spearman_after": 0.6329, | |
| "stsb_test_pearson_after": 0.6309 | |
| }, | |
| "toroidal_lion": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_dev_pearson_after": 0.697, | |
| "stsb_test_spearman_after": 0.6319, | |
| "stsb_test_pearson_after": 0.6341 | |
| }, | |
| "adamw": { | |
| "stsb_dev_spearman_after": 0.7289, | |
| "stsb_test_spearman_after": 0.6317, | |
| "stsb_test_pearson_after": 0.6372 | |
| }, | |
| "poncelet_adam": { | |
| "s1_dev_spearman": 0.6262, | |
| "stsb_dev_spearman_after": 0.7275, | |
| "stsb_dev_pearson_after": 0.6987, | |
| "stsb_test_spearman_after": 0.6315, | |
| "stsb_test_pearson_after": 0.6374, | |
| "rocm_elapsed_sec": 267.34 | |
| }, | |
| "hessian_pair_adam_3x_coupling": { | |
| "stsb_dev_spearman_after": 0.7278, | |
| "stsb_dev_pearson_after": 0.6917, | |
| "stsb_test_spearman_after": 0.6288, | |
| "stsb_test_pearson_after": 0.6316 | |
| }, | |
| "cyclide_adam": { | |
| "s1_dev_spearman": 0.6255, | |
| "stsb_dev_spearman_after": 0.7368, | |
| "stsb_dev_pearson_after": 0.7179, | |
| "stsb_test_spearman_after": 0.6287, | |
| "stsb_test_pearson_after": 0.6476, | |
| "rocm_elapsed_sec": 380.42 | |
| }, | |
| "rindler_adam": { | |
| "s1_dev_spearman": 0.6208, | |
| "stsb_dev_spearman_after": 0.7318, | |
| "stsb_dev_pearson_after": 0.7069, | |
| "stsb_test_spearman_after": 0.6264, | |
| "stsb_test_pearson_after": 0.6353, | |
| "rocm_elapsed_sec": 271.68 | |
| }, | |
| "pair_adam_3x_coupling": { | |
| "stsb_dev_spearman_after": 0.7179, | |
| "stsb_test_spearman_after": 0.6242, | |
| "stsb_test_pearson_after": 0.633 | |
| }, | |
| "toroidal_muon": { | |
| "stsb_dev_spearman_after": 0.7361, | |
| "stsb_dev_pearson_after": 0.7105, | |
| "stsb_test_spearman_after": 0.6197, | |
| "stsb_test_pearson_after": 0.6304 | |
| } | |
| } | |
| } |