蓬莱3 7m emmbedding ja

architecture:

  • lfm2moe

tarined by:

  • rtx3060 * 20m training code in ./training_codes

tokenizer from ku-nlp/gpt2-small-japanese-char
(added unused qwen3 special tokens from base model )

dataset:

  • hotchpotch/sentence_transformer_japanese use llmjp-kaken 30k rows 1 epoch

evals:

{
  "dataset_revision": "5bac629e25799df4c9c80a6a5db983d6cba9e77d",
  "task_name": "JSTS",
  "mteb_version": "2.18.5",
  "scores": {
    "validation": [
      {
        "pearson": 0.486124,
        "spearman": 0.481482,
        "cosine_pearson": 0.486124,
        "cosine_spearman": 0.481483,
        "manhattan_pearson": 0.502434,
        "manhattan_spearman": 0.481838,
        "euclidean_pearson": 0.502667,
        "euclidean_spearman": 0.481521,
        "main_score": 0.481483,
        "hf_subset": "default",
        "languages": [
          "jpn-Jpan"
        ],
        "mteb_version": "2.18.5"
      }
    ]
  },
  "evaluation_time": 6.252074241638184,
  "kg_co2_emissions": null,
  "date": null,
  "evaluation_phases": null
}
{
  "dataset_revision": "5713cd1b76a22b79c3595827c717c72b6cab8852",
  "task_name": "JaQuADRetrieval",
  "mteb_version": "2.18.5",
  "scores": {
    "validation": [
      {
        "ndcg_at_1": 0.05322,
        "ndcg_at_3": 0.05539,
        "ndcg_at_5": 0.06654,
        "ndcg_at_10": 0.08348,
        "ndcg_at_20": 0.10187,
        "ndcg_at_100": 0.14915,
        "ndcg_at_1000": 0.21649,
        "map_at_1": 0.02661,
        "map_at_3": 0.04114,
        "map_at_5": 0.0465,
        "map_at_10": 0.05274,
        "map_at_20": 0.05754,
        "map_at_100": 0.06367,
        "map_at_1000": 0.06633,
        "recall_at_1": 0.02661,
        "recall_at_3": 0.05957,
        "recall_at_5": 0.08154,
        "recall_at_10": 0.12451,
        "recall_at_20": 0.18335,
        "recall_at_100": 0.39771,
        "recall_at_1000": 0.85205,
        "accuracy": 0.02661,
        "precision_at_1": 0.05322,
        "precision_at_3": 0.03971,
        "precision_at_5": 0.03262,
        "precision_at_10": 0.0249,
        "precision_at_20": 0.01833,
        "precision_at_100": 0.00795,
        "precision_at_1000": 0.0017,
        "mrr_at_1": 0.053223,
        "mrr_at_3": 0.078695,
        "mrr_at_5": 0.088241,
        "mrr_at_10": 0.098402,
        "mrr_at_20": 0.10552,
        "mrr_at_100": 0.112814,
        "mrr_at_1000": 0.11427,
        "nauc_ndcg_at_1_max": 0.060183,
        "nauc_ndcg_at_1_std": 0.186826,
        "nauc_ndcg_at_1_diff1": 0.200748,
        "nauc_ndcg_at_3_max": 0.034569,
        "nauc_ndcg_at_3_std": 0.193343,
        "nauc_ndcg_at_3_diff1": 0.183758,
        "nauc_ndcg_at_5_max": 0.019013,
        "nauc_ndcg_at_5_std": 0.211276,
        "nauc_ndcg_at_5_diff1": 0.166229,
        "nauc_ndcg_at_10_max": 0.015117,
        "nauc_ndcg_at_10_std": 0.221909,
        "nauc_ndcg_at_10_diff1": 0.159813,
        "nauc_ndcg_at_20_max": 0.006519,
        "nauc_ndcg_at_20_std": 0.242097,
        "nauc_ndcg_at_20_diff1": 0.147229,
        "nauc_ndcg_at_100_max": 0.007738,
        "nauc_ndcg_at_100_std": 0.280225,
        "nauc_ndcg_at_100_diff1": 0.135159,
        "nauc_ndcg_at_1000_max": 0.031105,
        "nauc_ndcg_at_1000_std": 0.272176,
        "nauc_ndcg_at_1000_diff1": 0.125747,
        "nauc_map_at_1_max": 0.060183,
        "nauc_map_at_1_std": 0.186826,
        "nauc_map_at_1_diff1": 0.200748,
        "nauc_map_at_3_max": 0.036661,
        "nauc_map_at_3_std": 0.19899,
        "nauc_map_at_3_diff1": 0.186028,
        "nauc_map_at_5_max": 0.027618,
        "nauc_map_at_5_std": 0.21289,
        "nauc_map_at_5_diff1": 0.175309,
        "nauc_map_at_10_max": 0.024948,
        "nauc_map_at_10_std": 0.217784,
        "nauc_map_at_10_diff1": 0.173213,
        "nauc_map_at_20_max": 0.020302,
        "nauc_map_at_20_std": 0.228483,
        "nauc_map_at_20_diff1": 0.166941,
        "nauc_map_at_100_max": 0.019263,
        "nauc_map_at_100_std": 0.240281,
        "nauc_map_at_100_diff1": 0.163332,
        "nauc_map_at_1000_max": 0.020397,
        "nauc_map_at_1000_std": 0.240768,
        "nauc_map_at_1000_diff1": 0.162718,
        "nauc_recall_at_1_max": 0.060183,
        "nauc_recall_at_1_std": 0.186826,
        "nauc_recall_at_1_diff1": 0.200748,
        "nauc_recall_at_3_max": 0.025022,
        "nauc_recall_at_3_std": 0.191194,
        "nauc_recall_at_3_diff1": 0.189284,
        "nauc_recall_at_5_max": 2.7e-05,
        "nauc_recall_at_5_std": 0.222344,
        "nauc_recall_at_5_diff1": 0.158639,
        "nauc_recall_at_10_max": -0.003379,
        "nauc_recall_at_10_std": 0.240262,
        "nauc_recall_at_10_diff1": 0.147606,
        "nauc_recall_at_20_max": -0.019003,
        "nauc_recall_at_20_std": 0.279154,
        "nauc_recall_at_20_diff1": 0.123961,
        "nauc_recall_at_100_max": -0.009469,
        "nauc_recall_at_100_std": 0.326349,
        "nauc_recall_at_100_diff1": 0.084294,
        "nauc_recall_at_1000_max": 0.125963,
        "nauc_recall_at_1000_std": 0.28499,
        "nauc_recall_at_1000_diff1": -0.035354,
        "nauc_precision_at_1_max": 0.060183,
        "nauc_precision_at_1_std": 0.186826,
        "nauc_precision_at_1_diff1": 0.200748,
        "nauc_precision_at_3_max": 0.025022,
        "nauc_precision_at_3_std": 0.191194,
        "nauc_precision_at_3_diff1": 0.189284,
        "nauc_precision_at_5_max": 2.7e-05,
        "nauc_precision_at_5_std": 0.222344,
        "nauc_precision_at_5_diff1": 0.158639,
        "nauc_precision_at_10_max": -0.003379,
        "nauc_precision_at_10_std": 0.240262,
        "nauc_precision_at_10_diff1": 0.147606,
        "nauc_precision_at_20_max": -0.019003,
        "nauc_precision_at_20_std": 0.279154,
        "nauc_precision_at_20_diff1": 0.123961,
        "nauc_precision_at_100_max": -0.009469,
        "nauc_precision_at_100_std": 0.326349,
        "nauc_precision_at_100_diff1": 0.084294,
        "nauc_precision_at_1000_max": 0.125963,
        "nauc_precision_at_1000_std": 0.28499,
        "nauc_precision_at_1000_diff1": -0.035354,
        "nauc_mrr_at_1_max": 0.060183,
        "nauc_mrr_at_1_std": 0.186826,
        "nauc_mrr_at_1_diff1": 0.200748,
        "nauc_mrr_at_3_max": 0.042032,
        "nauc_mrr_at_3_std": 0.189095,
        "nauc_mrr_at_3_diff1": 0.17712,
        "nauc_mrr_at_5_max": 0.030667,
        "nauc_mrr_at_5_std": 0.197766,
        "nauc_mrr_at_5_diff1": 0.16525,
        "nauc_mrr_at_10_max": 0.028667,
        "nauc_mrr_at_10_std": 0.203482,
        "nauc_mrr_at_10_diff1": 0.160405,
        "nauc_mrr_at_20_max": 0.026953,
        "nauc_mrr_at_20_std": 0.206592,
        "nauc_mrr_at_20_diff1": 0.155442,
        "nauc_mrr_at_100_max": 0.027542,
        "nauc_mrr_at_100_std": 0.209632,
        "nauc_mrr_at_100_diff1": 0.153106,
        "nauc_mrr_at_1000_max": 0.027904,
        "nauc_mrr_at_1000_std": 0.209228,
        "nauc_mrr_at_1000_diff1": 0.153605,
        "hit_rate_at_1": 0.05322,
        "hit_rate_at_3": 0.11426,
        "hit_rate_at_5": 0.15576,
        "hit_rate_at_10": 0.23291,
        "hit_rate_at_20": 0.33252,
        "hit_rate_at_100": 0.64014,
        "hit_rate_at_1000": 0.96826,
        "main_score": 0.08348,
        "hf_subset": "default",
        "languages": [
          "jpn-Jpan"
        ],
        "mteb_version": "2.18.5"
      }
    ]
  },
  "evaluation_time": 14.237053155899048,
  "kg_co2_emissions": null,
  "date": null,
  "evaluation_phases": null
}
Downloads last month
15
Safetensors
Model size
6.55M params
Tensor type
F32
·
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for WariHima/hourai3-7m-embedding-ja

Finetuned
(1)
this model
Quantizations
1 model

Collection including WariHima/hourai3-7m-embedding-ja