BiBo-Mini-Arc-Easy-Challenege.json

{
  "config_general": {
    "lighteval_sha": "?",
    "num_fewshot_seeds": 1,
    "override_batch_size": -1,
    "max_samples": null,
    "job_id": "",
    "start_time": 1342.339773502,
    "end_time": 6498.524528434,
    "total_evaluation_time_secondes": "5156.184754932",
    "model_name": "tinycompany/BiBo-Mini-v0.99",
    "model_sha": "9e66d1135c27c915ecf382453909e5bc1f6a650d",
    "model_dtype": "torch.bfloat16",
    "model_size": "3.31 GB",
    "config": null
  },
  "results": {
    "indiceval|ARC-Challenge:hindi|10": {
      "acc": 0.30887372013651876,
      "acc_stderr": 0.013501770929344003,
      "acc_norm": 0.3361774744027304,
      "acc_norm_stderr": 0.013804855026205752
    },
    "indiceval|ARC-Easy:hindi|5": {
      "acc": 0.5627104377104377,
      "acc_stderr": 0.010178768429321586,
      "acc_norm": 0.5349326599326599,
      "acc_norm_stderr": 0.01023471305272368
    },
    "all": {
      "acc": 0.43579207892347827,
      "acc_stderr": 0.011840269679332795,
      "acc_norm": 0.43555506716769515,
      "acc_norm_stderr": 0.012019784039464717
    }
  },
  "versions": {
    "indiceval|ARC-Challenge:hindi|10": 0,
    "indiceval|ARC-Easy:hindi|5": 0
  },
  "config_tasks": {
    "indiceval|ARC-Challenge:hindi": {
      "name": "ARC-Challenge:hindi",
      "prompt_function": "arc_indic",
      "hf_repo": "Cognitive-Lab/Indic-ARC-Challenge",
      "hf_subset": "hi",
      "metric": [
        "loglikelihood_acc",
        "loglikelihood_acc_norm_nospace"
      ],
      "hf_avail_splits": [
        "train",
        "validation",
        "test"
      ],
      "evaluation_splits": [
        "test"
      ],
      "few_shots_split": null,
      "few_shots_select": "random_sampling_from_train",
      "generation_size": 1,
      "stop_sequence": [
        "\n"
      ],
      "output_regex": null,
      "frozen": false,
      "suite": [
        "indiceval",
        "leaderboard",
        "arc"
      ],
      "original_num_docs": 1172,
      "effective_num_docs": 1172,
      "trust_dataset": true,
      "must_remove_duplicate_docs": null
    },
    "indiceval|ARC-Easy:hindi": {
      "name": "ARC-Easy:hindi",
      "prompt_function": "arc_indic",
      "hf_repo": "Cognitive-Lab/Indic-ARC-Easy",
      "hf_subset": "hi",
      "metric": [
        "loglikelihood_acc",
        "loglikelihood_acc_norm_nospace"
      ],
      "hf_avail_splits": [
        "train",
        "validation",
        "test"
      ],
      "evaluation_splits": [
        "test"
      ],
      "few_shots_split": null,
      "few_shots_select": "random_sampling_from_train",
      "generation_size": 1,
      "stop_sequence": [
        "\n"
      ],
      "output_regex": null,
      "frozen": false,
      "suite": [
        "indiceval",
        "leaderboard",
        "arc"
      ],
      "original_num_docs": 2376,
      "effective_num_docs": 2376,
      "trust_dataset": true,
      "must_remove_duplicate_docs": null
    }
  },
  "summary_tasks": {
    "indiceval|ARC-Challenge:hindi|10": {
      "hashes": {
        "hash_examples": "ede3937107b50671",
        "hash_full_prompts": "939fa143e0be8e76",
        "hash_input_tokens": "260d85f401153bdc",
        "hash_cont_tokens": "b95b113bb2e57385"
      },
      "truncated": 0,
      "non_truncated": 1172,
      "padded": 4680,
      "non_padded": 7,
      "effective_few_shots": 10.0,
      "num_truncated_few_shots": 0
    },
    "indiceval|ARC-Easy:hindi|5": {
      "hashes": {
        "hash_examples": "0186dde6d6cf5f12",
        "hash_full_prompts": "cbaaa6e70e1e350d",
        "hash_input_tokens": "5496daca90725251",
        "hash_cont_tokens": "f7cf5b125bc52602"
      },
      "truncated": 0,
      "non_truncated": 2376,
      "padded": 9441,
      "non_padded": 60,
      "effective_few_shots": 5.0,
      "num_truncated_few_shots": 0
    }
  },
  "summary_general": {
    "hashes": {
      "hash_examples": "799a1387b6c8a4d2",
      "hash_full_prompts": "9938eadd006079ef",
      "hash_input_tokens": "13eac7a7a2c17518",
      "hash_cont_tokens": "14960e7bf91fd26b"
    },
    "truncated": 0,
    "non_truncated": 3548,
    "padded": 14121,
    "non_padded": 67,
    "num_truncated_few_shots": 0
  },
  "email": "[email protected]",
  "language": "hindi"
}