{ "results": { "afrimgsm_direct_xho": { "alias": "afrimgsm_direct_xho", "exact_match,remove_whitespace": 0.0, "exact_match_stderr,remove_whitespace": 0.0, "exact_match,flexible-extract": 0.008, "exact_match_stderr,flexible-extract": 0.005645483676690173 }, "afrimgsm_direct_zul": { "alias": "afrimgsm_direct_zul", "exact_match,remove_whitespace": 0.0, "exact_match_stderr,remove_whitespace": 0.0, "exact_match,flexible-extract": 0.02, "exact_match_stderr,flexible-extract": 0.008872139507342683 }, "afrimmlu_direct_xho": { "alias": "afrimmlu_direct_xho", "acc,none": 0.228, "acc_stderr,none": 0.018781306529363204, "f1,none": 0.22499762903231807, "f1_stderr,none": "N/A" }, "afrimmlu_direct_zul": { "alias": "afrimmlu_direct_zul", "acc,none": 0.25, "acc_stderr,none": 0.019384310743640384, "f1,none": 0.2522065219763795, "f1_stderr,none": "N/A" }, "afrixnli_en_direct_xho": { "alias": "afrixnli_en_direct_xho", "acc,none": 0.335, "acc_stderr,none": 0.019285007627653686, "f1,none": 0.27778903523584375, "f1_stderr,none": "N/A" }, "afrixnli_en_direct_zul": { "alias": "afrixnli_en_direct_zul", "acc,none": 0.33166666666666667, "acc_stderr,none": 0.019236854622688142, "f1,none": 0.26541415513303723, "f1_stderr,none": "N/A" } }, "group_subtasks": { "afrimgsm_direct_xho": [], "afrimgsm_direct_zul": [], "afrimmlu_direct_xho": [], "afrimmlu_direct_zul": [], "afrixnli_en_direct_xho": [], "afrixnli_en_direct_zul": [] }, "configs": { "afrimgsm_direct_xho": { "task": "afrimgsm_direct_xho", "tag": [ "afrimgsm", "afrimgsm_direct" ], "group": [ "afrimgsm", "afrimgsm_direct" ], "dataset_path": "masakhane/afrimgsm", "dataset_name": "xho", "test_split": "test", "doc_to_text": "{% if answer is not none %}{{question+\"\\nAnswer:\"}}{% else %}{{\"Question: \"+question+\"\\nAnswer:\"}}{% endif %}", "doc_to_target": "{% if answer is not none %}{{answer[21:]}}{% else %}{{answer_number|string}}{% endif %}", "description": "", "target_delimiter": "", "fewshot_delimiter": "\n\n", "num_fewshot": 0, "metric_list": [ { "metric": "exact_match", "aggregation": "mean", "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true } ], "output_type": "generate_until", "generation_kwargs": { "do_sample": false, "until": [ "Question:", "", "<|im_end|>" ] }, "repeats": 1, "filter_list": [ { "name": "remove_whitespace", "filter": [ { "function": "remove_whitespace" }, { "function": "take_first" } ] }, { "filter": [ { "function": "regex", "group_select": -1, "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)" }, { "function": "take_first" } ], "name": "flexible-extract" } ], "should_decontaminate": false, "metadata": { "version": 2.0 } }, "afrimgsm_direct_zul": { "task": "afrimgsm_direct_zul", "tag": [ "afrimgsm", "afrimgsm_direct" ], "group": [ "afrimgsm", "afrimgsm_direct" ], "dataset_path": "masakhane/afrimgsm", "dataset_name": "zul", "test_split": "test", "doc_to_text": "{% if answer is not none %}{{question+\"\\nAnswer:\"}}{% else %}{{\"Question: \"+question+\"\\nAnswer:\"}}{% endif %}", "doc_to_target": "{% if answer is not none %}{{answer[21:]}}{% else %}{{answer_number|string}}{% endif %}", "description": "", "target_delimiter": "", "fewshot_delimiter": "\n\n", "num_fewshot": 0, "metric_list": [ { "metric": "exact_match", "aggregation": "mean", "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true } ], "output_type": "generate_until", "generation_kwargs": { "do_sample": false, "until": [ "Question:", "", "<|im_end|>" ] }, "repeats": 1, "filter_list": [ { "name": "remove_whitespace", "filter": [ { "function": "remove_whitespace" }, { "function": "take_first" } ] }, { "filter": [ { "function": "regex", "group_select": -1, "regex_pattern": "(-?[$0-9.,]{2,})|(-?[0-9]+)" }, { "function": "take_first" } ], "name": "flexible-extract" } ], "should_decontaminate": false, "metadata": { "version": 2.0 } }, "afrimmlu_direct_xho": { "task": "afrimmlu_direct_xho", "tag": [ "afrimmlu", "afrimmlu_direct" ], "group": [ "afrimmlu", "afrimmlu_direct" ], "dataset_path": "masakhane/afrimmlu", "dataset_name": "xho", "validation_split": "validation", "test_split": "test", "fewshot_split": "validation", "doc_to_text": "def doc_to_text(doc):\n output = \"\"\"You are a highly knowledgeable and intelligent artificial intelligence\n model answers multiple-choice questions about {subject}\n\n Question: {question}\n\n Choices:\n A: {choice1}\n B: {choice2}\n C: {choice3}\n D: {choice4}\n\n Answer: \"\"\"\n\n choices = eval(doc[\"choices\"])\n text = output.format(\n subject=doc[\"subject\"],\n question=doc[\"question\"],\n choice1=choices[0],\n choice2=choices[1],\n choice3=choices[2],\n choice4=choices[3],\n )\n return text\n", "doc_to_target": "{{['A', 'B', 'C', 'D'].index(answer)}}", "doc_to_choice": "def doc_to_choice(doc):\n choices = eval(doc[\"choices\"])\n return choices\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "num_fewshot": 0, "metric_list": [ { "metric": "f1", "aggregation": "def weighted_f1_score(items):\n from sklearn.metrics import f1_score\n\n unzipped_list = list(zip(*items))\n golds = unzipped_list[0]\n preds = unzipped_list[1]\n fscore = f1_score(golds, preds, average=\"weighted\")\n return fscore\n", "average": "weighted", "hf_evaluate": true, "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true, "regexes_to_ignore": [ ",", "\\$" ] }, { "metric": "acc", "aggregation": "mean", "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true, "regexes_to_ignore": [ ",", "\\$" ] } ], "output_type": "multiple_choice", "repeats": 1, "should_decontaminate": true, "doc_to_decontamination_query": "Question: {{question}}\nAnswer:", "metadata": { "version": 1.0 } }, "afrimmlu_direct_zul": { "task": "afrimmlu_direct_zul", "tag": [ "afrimmlu", "afrimmlu_direct" ], "group": [ "afrimmlu", "afrimmlu_direct" ], "dataset_path": "masakhane/afrimmlu", "dataset_name": "zul", "validation_split": "validation", "test_split": "test", "fewshot_split": "validation", "doc_to_text": "def doc_to_text(doc):\n output = \"\"\"You are a highly knowledgeable and intelligent artificial intelligence\n model answers multiple-choice questions about {subject}\n\n Question: {question}\n\n Choices:\n A: {choice1}\n B: {choice2}\n C: {choice3}\n D: {choice4}\n\n Answer: \"\"\"\n\n choices = eval(doc[\"choices\"])\n text = output.format(\n subject=doc[\"subject\"],\n question=doc[\"question\"],\n choice1=choices[0],\n choice2=choices[1],\n choice3=choices[2],\n choice4=choices[3],\n )\n return text\n", "doc_to_target": "{{['A', 'B', 'C', 'D'].index(answer)}}", "doc_to_choice": "def doc_to_choice(doc):\n choices = eval(doc[\"choices\"])\n return choices\n", "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "num_fewshot": 0, "metric_list": [ { "metric": "f1", "aggregation": "def weighted_f1_score(items):\n from sklearn.metrics import f1_score\n\n unzipped_list = list(zip(*items))\n golds = unzipped_list[0]\n preds = unzipped_list[1]\n fscore = f1_score(golds, preds, average=\"weighted\")\n return fscore\n", "average": "weighted", "hf_evaluate": true, "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true, "regexes_to_ignore": [ ",", "\\$" ] }, { "metric": "acc", "aggregation": "mean", "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true, "regexes_to_ignore": [ ",", "\\$" ] } ], "output_type": "multiple_choice", "repeats": 1, "should_decontaminate": true, "doc_to_decontamination_query": "Question: {{question}}\nAnswer:", "metadata": { "version": 1.0 } }, "afrixnli_en_direct_xho": { "task": "afrixnli_en_direct_xho", "tag": [ "afrixnli", "afrixnli_en_direct" ], "group": [ "afrixnli", "afrixnli_en_direct" ], "dataset_path": "masakhane/afrixnli", "dataset_name": "xho", "validation_split": "validation", "test_split": "test", "fewshot_split": "validation", "doc_to_text": "{{premise}}\nQuestion: {{hypothesis}} True, False, or Neither?\nAnswer:", "doc_to_target": "def doc_to_target(doc):\n replacements = {0: \"True\", 1: \"Neither\", 2: \"False\"}\n return replacements[doc[\"label\"]]\n", "doc_to_choice": [ "True", "Neither", "False" ], "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "num_fewshot": 0, "metric_list": [ { "metric": "f1", "aggregation": "def weighted_f1_score(items):\n from sklearn.metrics import f1_score\n\n unzipped_list = list(zip(*items))\n golds = unzipped_list[0]\n preds = unzipped_list[1]\n fscore = f1_score(golds, preds, average=\"weighted\")\n return fscore\n", "average": "weighted", "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true }, { "metric": "acc", "aggregation": "mean", "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true } ], "output_type": "multiple_choice", "repeats": 1, "should_decontaminate": true, "doc_to_decontamination_query": "premise", "metadata": { "version": 1.0 } }, "afrixnli_en_direct_zul": { "task": "afrixnli_en_direct_zul", "tag": [ "afrixnli", "afrixnli_en_direct" ], "group": [ "afrixnli", "afrixnli_en_direct" ], "dataset_path": "masakhane/afrixnli", "dataset_name": "zul", "validation_split": "validation", "test_split": "test", "fewshot_split": "validation", "doc_to_text": "{{premise}}\nQuestion: {{hypothesis}} True, False, or Neither?\nAnswer:", "doc_to_target": "def doc_to_target(doc):\n replacements = {0: \"True\", 1: \"Neither\", 2: \"False\"}\n return replacements[doc[\"label\"]]\n", "doc_to_choice": [ "True", "Neither", "False" ], "description": "", "target_delimiter": " ", "fewshot_delimiter": "\n\n", "num_fewshot": 0, "metric_list": [ { "metric": "f1", "aggregation": "def weighted_f1_score(items):\n from sklearn.metrics import f1_score\n\n unzipped_list = list(zip(*items))\n golds = unzipped_list[0]\n preds = unzipped_list[1]\n fscore = f1_score(golds, preds, average=\"weighted\")\n return fscore\n", "average": "weighted", "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true }, { "metric": "acc", "aggregation": "mean", "higher_is_better": true, "ignore_case": true, "ignore_punctuation": true } ], "output_type": "multiple_choice", "repeats": 1, "should_decontaminate": true, "doc_to_decontamination_query": "premise", "metadata": { "version": 1.0 } } }, "versions": { "afrimgsm_direct_xho": 2.0, "afrimgsm_direct_zul": 2.0, "afrimmlu_direct_xho": 1.0, "afrimmlu_direct_zul": 1.0, "afrixnli_en_direct_xho": 1.0, "afrixnli_en_direct_zul": 1.0 }, "n-shot": { "afrimgsm_direct_xho": 0, "afrimgsm_direct_zul": 0, "afrimmlu_direct_xho": 0, "afrimmlu_direct_zul": 0, "afrixnli_en_direct_xho": 0, "afrixnli_en_direct_zul": 0 }, "higher_is_better": { "afrimgsm_direct_xho": { "exact_match": true }, "afrimgsm_direct_zul": { "exact_match": true }, "afrimmlu_direct_xho": { "f1": true, "acc": true }, "afrimmlu_direct_zul": { "f1": true, "acc": true }, "afrixnli_en_direct_xho": { "f1": true, "acc": true }, "afrixnli_en_direct_zul": { "f1": true, "acc": true } }, "n-samples": { "afrixnli_en_direct_zul": { "original": 600, "effective": 600 }, "afrixnli_en_direct_xho": { "original": 600, "effective": 600 }, "afrimmlu_direct_zul": { "original": 500, "effective": 500 }, "afrimmlu_direct_xho": { "original": 500, "effective": 500 }, "afrimgsm_direct_zul": { "original": 250, "effective": 250 }, "afrimgsm_direct_xho": { "original": 250, "effective": 250 } }, "config": { "model": "hf", "model_args": "pretrained=CohereForAI/aya-23-8B", "model_num_parameters": 8028033024, "model_dtype": "torch.float16", "model_revision": "main", "model_sha": "5f6dc63f4b4071cff399af26e576f4540554236d", "batch_size": "auto:4", "batch_sizes": [ 8, 32, 64, 64 ], "device": null, "use_cache": null, "limit": null, "bootstrap_iters": 100000, "gen_kwargs": null, "random_seed": 0, "numpy_seed": 1234, "torch_seed": 1234, "fewshot_seed": 1234 }, "git_hash": "15ffb0d", "date": 1727750420.4119575, "pretty_env_info": "PyTorch version: 2.4.1+cu121\nIs debug build: False\nCUDA used to build PyTorch: 12.1\nROCM used to build PyTorch: N/A\n\nOS: Ubuntu 22.04.3 LTS (x86_64)\nGCC version: (Ubuntu 11.4.0-1ubuntu1~22.04) 11.4.0\nClang version: Could not collect\nCMake version: version 3.22.1\nLibc version: glibc-2.35\n\nPython version: 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0] (64-bit runtime)\nPython platform: Linux-6.2.0-37-generic-x86_64-with-glibc2.35\nIs CUDA available: True\nCUDA runtime version: 12.2.140\nCUDA_MODULE_LOADING set to: LAZY\nGPU models and configuration: \nGPU 0: NVIDIA RTX A6000\nGPU 1: NVIDIA RTX A6000\n\nNvidia driver version: 535.129.03\ncuDNN version: Could not collect\nHIP runtime version: N/A\nMIOpen runtime version: N/A\nIs XNNPACK available: True\n\nCPU:\nArchitecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nAddress sizes: 40 bits physical, 48 bits virtual\nByte Order: Little Endian\nCPU(s): 28\nOn-line CPU(s) list: 0-27\nVendor ID: AuthenticAMD\nModel name: AMD EPYC-Rome Processor\nCPU family: 23\nModel: 49\nThread(s) per core: 1\nCore(s) per socket: 1\nSocket(s): 28\nStepping: 0\nBogoMIPS: 4999.23\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm rep_good nopl cpuid extd_apicid pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy svm cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw topoext perfctr_core ssbd ibrs ibpb stibp vmmcall fsgsbase tsc_adjust bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 clzero xsaveerptr wbnoinvd arat npt nrip_save umip rdpid arch_capabilities\nVirtualization: AMD-V\nL1d cache: 896 KiB (28 instances)\nL1i cache: 896 KiB (28 instances)\nL2 cache: 14 MiB (28 instances)\nL3 cache: 448 MiB (28 instances)\nNUMA node(s): 1\nNUMA node0 CPU(s): 0-27\nVulnerability Gather data sampling: Not affected\nVulnerability Itlb multihit: Not affected\nVulnerability L1tf: Not affected\nVulnerability Mds: Not affected\nVulnerability Meltdown: Not affected\nVulnerability Mmio stale data: Not affected\nVulnerability Retbleed: Mitigation; untrained return thunk; SMT disabled\nVulnerability Spec rstack overflow: Mitigation; SMT disabled\nVulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Retpolines, IBPB conditional, STIBP disabled, RSB filling, PBRSB-eIBRS Not affected\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\n\nVersions of relevant libraries:\n[pip3] flake8==4.0.1\n[pip3] numpy==1.25.2\n[pip3] torch==2.4.1+cu121\n[pip3] torchaudio==2.4.1+cu121\n[pip3] torchvision==0.19.1+cu121\n[pip3] triton==3.0.0\n[conda] Could not collect", "transformers_version": "4.45.1", "upper_git_hash": null, "tokenizer_pad_token": [ "", "0" ], "tokenizer_eos_token": [ "<|END_OF_TURN_TOKEN|>", "255001" ], "tokenizer_bos_token": [ "", "5" ], "eot_token_id": 255001, "max_length": 8192, "task_hashes": { "afrixnli_en_direct_zul": "011b872bfe35d1ead7694b59c7023bc079845f39fb417791e0c2e19e49c8ce6e", "afrixnli_en_direct_xho": "812b77def909fef6b7ec5373d4bfa09d6a6f5b2971b0bcad3e81a1f94d743411", "afrimmlu_direct_zul": "460ed49479021e40a2b7b112085638761d2b46580532bb66a18403f43575d9d5", "afrimmlu_direct_xho": "7cb5c1bd5911e13faf3f2e7c2740974738d8396d115a4fe06ab4af64e8dee56b", "afrimgsm_direct_zul": "afc89857751cbc97ed864d974b6032c80c182128e51964077051627b45798654", "afrimgsm_direct_xho": "56a4760bd96dbcd55fb7f296c706a2846e0533cb832b638f98f56d8f96d4d3ad" }, "model_source": "hf", "model_name": "CohereForAI/aya-23-8B", "model_name_sanitized": "CohereForAI__aya-23-8B", "system_instruction": null, "system_instruction_sha": null, "fewshot_as_multiturn": false, "chat_template": null, "chat_template_sha": null, "start_time": 6430.178763367, "end_time": 7735.584634235, "total_evaluation_time_seconds": "1305.4058708679995" }