初始化项目，由ModelHub XC社区提供模型
Model: lomahony/eleuther-pythia410m-hh-sft Source: Original Platform
This commit is contained in:
ModelHub XC
35
.gitattributes
vendored
Normal file
35
.gitattributes
vendored
Normal file
@@ -0,0 +1,35 @@
|
||||
*.7z filter=lfs diff=lfs merge=lfs -text
|
||||
*.arrow filter=lfs diff=lfs merge=lfs -text
|
||||
*.bin filter=lfs diff=lfs merge=lfs -text
|
||||
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
||||
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
||||
*.ftz filter=lfs diff=lfs merge=lfs -text
|
||||
*.gz filter=lfs diff=lfs merge=lfs -text
|
||||
*.h5 filter=lfs diff=lfs merge=lfs -text
|
||||
*.joblib filter=lfs diff=lfs merge=lfs -text
|
||||
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
||||
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
||||
*.model filter=lfs diff=lfs merge=lfs -text
|
||||
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
||||
*.npy filter=lfs diff=lfs merge=lfs -text
|
||||
*.npz filter=lfs diff=lfs merge=lfs -text
|
||||
*.onnx filter=lfs diff=lfs merge=lfs -text
|
||||
*.ot filter=lfs diff=lfs merge=lfs -text
|
||||
*.parquet filter=lfs diff=lfs merge=lfs -text
|
||||
*.pb filter=lfs diff=lfs merge=lfs -text
|
||||
*.pickle filter=lfs diff=lfs merge=lfs -text
|
||||
*.pkl filter=lfs diff=lfs merge=lfs -text
|
||||
*.pt filter=lfs diff=lfs merge=lfs -text
|
||||
*.pth filter=lfs diff=lfs merge=lfs -text
|
||||
*.rar filter=lfs diff=lfs merge=lfs -text
|
||||
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
||||
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
||||
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
||||
*.tar filter=lfs diff=lfs merge=lfs -text
|
||||
*.tflite filter=lfs diff=lfs merge=lfs -text
|
||||
*.tgz filter=lfs diff=lfs merge=lfs -text
|
||||
*.wasm filter=lfs diff=lfs merge=lfs -text
|
||||
*.xz filter=lfs diff=lfs merge=lfs -text
|
||||
*.zip filter=lfs diff=lfs merge=lfs -text
|
||||
*.zst filter=lfs diff=lfs merge=lfs -text
|
||||
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
||||
20
README.md
Normal file
20
README.md
Normal file
@@ -0,0 +1,20 @@
|
||||
---
|
||||
language:
|
||||
- en
|
||||
tags:
|
||||
- pytorch
|
||||
- causal-lm
|
||||
- pythia
|
||||
license: apache-2.0
|
||||
datasets:
|
||||
- Anthropic/hh-rlhf
|
||||
---
|
||||
|
||||
[Pythia-410m](https://huggingface.co/EleutherAI/pythia-410m) supervised finetuned with [Anthropic-hh-rlhf dataset](https://huggingface.co/datasets/Anthropic/hh-rlhf) for 1 epoch.
|
||||
|
||||
[wandb log](https://wandb.ai/pythia_dpo/Pythia_LOM/runs/0tlrjao6?workspace=user-lauraomahony)
|
||||
|
||||
Benchmark evaluations included in repo done using [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness/tree/big-refactor).
|
||||
|
||||
See [Pythia-410m](https://huggingface.co/EleutherAI/pythia-160m) for model details [(paper)](https://arxiv.org/abs/2101.00027).
|
||||
|
||||
@@ -0,0 +1,430 @@
|
||||
bootstrapping for stddev: perplexity
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.21416382252559726,
|
||||
"acc_stderr,none": 0.011988383205966515,
|
||||
"acc_norm,none": 0.2431740614334471,
|
||||
"acc_norm_stderr,none": 0.012536554144587084
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.5189393939393939,
|
||||
"acc_stderr,none": 0.01025242049689449,
|
||||
"acc_norm,none": 0.45707070707070707,
|
||||
"acc_norm_stderr,none": 0.010221897564256049
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.6058103975535168,
|
||||
"acc_stderr,none": 0.008546995661233639
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.33718382792272455,
|
||||
"acc_stderr,none": 0.004717820714968757,
|
||||
"acc_norm,none": 0.4060944035052778,
|
||||
"acc_norm_stderr,none": 0.004900988997414242
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 10.780459714601333,
|
||||
"perplexity_stderr,none": 0.32049412467424027,
|
||||
"acc,none": 0.5163982146322531,
|
||||
"acc_stderr,none": 0.006962230326368326
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.182,
|
||||
"acc_stderr,none": 0.01727277329773045,
|
||||
"acc_norm,none": 0.294,
|
||||
"acc_norm_stderr,none": 0.020395095484936624
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6670293797606094,
|
||||
"acc_stderr,none": 0.010995648822619082,
|
||||
"acc_norm,none": 0.6719260065288357,
|
||||
"acc_norm_stderr,none": 0.010954487135124227
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.815,
|
||||
"acc_stderr,none": 0.012285191326386667,
|
||||
"acc_norm,none": 0.725,
|
||||
"acc_norm_stderr,none": 0.014127086556490528
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 34.50450469911897,
|
||||
"byte_perplexity,none": 1.7927778872125213,
|
||||
"bits_per_byte,none": 0.842196759334895
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5335438042620363,
|
||||
"acc_stderr,none": 0.014020826677598103
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7fae1b130040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7fae1b102ef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7fae1b103370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7fae1b1035b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=EleutherAI/pythia-410m",
|
||||
"num_fewshot": 0,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
hf (pretrained=EleutherAI/pythia-410m), limit: None, num_fewshot: 0, batch_size: 16
|
||||
| Task |Version|Filter| Metric | Value | |Stderr|
|
||||
|--------------|-------|------|---------------|------:|---|-----:|
|
||||
|arc_challenge |Yaml |none |acc | 0.2142|± |0.0120|
|
||||
| | |none |acc_norm | 0.2432|± |0.0125|
|
||||
|arc_easy |Yaml |none |acc | 0.5189|± |0.0103|
|
||||
| | |none |acc_norm | 0.4571|± |0.0102|
|
||||
|boolq |Yaml |none |acc | 0.6058|± |0.0085|
|
||||
|hellaswag |Yaml |none |acc | 0.3372|± |0.0047|
|
||||
| | |none |acc_norm | 0.4061|± |0.0049|
|
||||
|lambada_openai|Yaml |none |perplexity |10.7805|± |0.3205|
|
||||
| | |none |acc | 0.5164|± |0.0070|
|
||||
|openbookqa |Yaml |none |acc | 0.1820|± |0.0173|
|
||||
| | |none |acc_norm | 0.2940|± |0.0204|
|
||||
|piqa |Yaml |none |acc | 0.6670|± |0.0110|
|
||||
| | |none |acc_norm | 0.6719|± |0.0110|
|
||||
|sciq |Yaml |none |acc | 0.8150|± |0.0123|
|
||||
| | |none |acc_norm | 0.7250|± |0.0141|
|
||||
|wikitext |Yaml |none |word_perplexity|34.5045| | |
|
||||
| | |none |byte_perplexity| 1.7928| | |
|
||||
| | |none |bits_per_byte | 0.8422| | |
|
||||
|winogrande |Yaml |none |acc | 0.5335|± |0.0140|
|
||||
|
||||
406
base-410m-eval-files/EleutherAI-pythia-410m-0shot/results.json
Normal file
406
base-410m-eval-files/EleutherAI-pythia-410m-0shot/results.json
Normal file
@@ -0,0 +1,406 @@
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.21416382252559726,
|
||||
"acc_stderr,none": 0.011988383205966515,
|
||||
"acc_norm,none": 0.2431740614334471,
|
||||
"acc_norm_stderr,none": 0.012536554144587084
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.5189393939393939,
|
||||
"acc_stderr,none": 0.01025242049689449,
|
||||
"acc_norm,none": 0.45707070707070707,
|
||||
"acc_norm_stderr,none": 0.010221897564256049
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.6058103975535168,
|
||||
"acc_stderr,none": 0.008546995661233639
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.33718382792272455,
|
||||
"acc_stderr,none": 0.004717820714968757,
|
||||
"acc_norm,none": 0.4060944035052778,
|
||||
"acc_norm_stderr,none": 0.004900988997414242
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 10.780459714601333,
|
||||
"perplexity_stderr,none": 0.32049412467424027,
|
||||
"acc,none": 0.5163982146322531,
|
||||
"acc_stderr,none": 0.006962230326368326
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.182,
|
||||
"acc_stderr,none": 0.01727277329773045,
|
||||
"acc_norm,none": 0.294,
|
||||
"acc_norm_stderr,none": 0.020395095484936624
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6670293797606094,
|
||||
"acc_stderr,none": 0.010995648822619082,
|
||||
"acc_norm,none": 0.6719260065288357,
|
||||
"acc_norm_stderr,none": 0.010954487135124227
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.815,
|
||||
"acc_stderr,none": 0.012285191326386667,
|
||||
"acc_norm,none": 0.725,
|
||||
"acc_norm_stderr,none": 0.014127086556490528
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 34.50450469911897,
|
||||
"byte_perplexity,none": 1.7927778872125213,
|
||||
"bits_per_byte,none": 0.842196759334895
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5335438042620363,
|
||||
"acc_stderr,none": 0.014020826677598103
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7fae1b130040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7fae1b102ef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7fae1b103370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7fae1b1035b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=EleutherAI/pythia-410m",
|
||||
"num_fewshot": 0,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
@@ -0,0 +1,440 @@
|
||||
Downloading and preparing dataset super_glue/boolq to /home/laura/.cache/huggingface/datasets/super_glue/boolq/1.0.3/bb9675f958ebfee0d5d6dc5476fafe38c79123727a7258d515c450873dbdbbed...
|
||||
Dataset super_glue downloaded and prepared to /home/laura/.cache/huggingface/datasets/super_glue/boolq/1.0.3/bb9675f958ebfee0d5d6dc5476fafe38c79123727a7258d515c450873dbdbbed. Subsequent calls will reuse this data.
|
||||
Downloading and preparing dataset openbookqa/main to /home/laura/.cache/huggingface/datasets/openbookqa/main/1.0.1/f338ccacfbc86fb8c2de3aa1c06d2ce686933de3bca284dba97d32592c52b33f...
|
||||
Dataset openbookqa downloaded and prepared to /home/laura/.cache/huggingface/datasets/openbookqa/main/1.0.1/f338ccacfbc86fb8c2de3aa1c06d2ce686933de3bca284dba97d32592c52b33f. Subsequent calls will reuse this data.
|
||||
Downloading and preparing dataset piqa/plain_text to /home/laura/.cache/huggingface/datasets/piqa/plain_text/1.1.0/6c611c1a9bf220943c4174e117d3b660859665baf1d43156230116185312d011...
|
||||
Dataset piqa downloaded and prepared to /home/laura/.cache/huggingface/datasets/piqa/plain_text/1.1.0/6c611c1a9bf220943c4174e117d3b660859665baf1d43156230116185312d011. Subsequent calls will reuse this data.
|
||||
Downloading and preparing dataset sciq/default to /home/laura/.cache/huggingface/datasets/sciq/default/0.1.0/50e5c6e3795b55463819d399ec417bfd4c3c621105e00295ddb5f3633d708493...
|
||||
Dataset sciq downloaded and prepared to /home/laura/.cache/huggingface/datasets/sciq/default/0.1.0/50e5c6e3795b55463819d399ec417bfd4c3c621105e00295ddb5f3633d708493. Subsequent calls will reuse this data.
|
||||
Downloading and preparing dataset winogrande/winogrande_xl to /home/laura/.cache/huggingface/datasets/winogrande/winogrande_xl/1.1.0/a826c3d3506aefe0e9e9390dcb53271070536586bab95849876b2c1743df56e2...
|
||||
Dataset winogrande downloaded and prepared to /home/laura/.cache/huggingface/datasets/winogrande/winogrande_xl/1.1.0/a826c3d3506aefe0e9e9390dcb53271070536586bab95849876b2c1743df56e2. Subsequent calls will reuse this data.
|
||||
bootstrapping for stddev: perplexity
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.21843003412969283,
|
||||
"acc_stderr,none": 0.012074291605700959,
|
||||
"acc_norm,none": 0.2645051194539249,
|
||||
"acc_norm_stderr,none": 0.012889272949313368
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.54503367003367,
|
||||
"acc_stderr,none": 0.010218084454602589,
|
||||
"acc_norm,none": 0.5370370370370371,
|
||||
"acc_norm_stderr,none": 0.010231597249131058
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.4871559633027523,
|
||||
"acc_stderr,none": 0.008742169169427067
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.33827922724556864,
|
||||
"acc_stderr,none": 0.004721571443354456,
|
||||
"acc_norm,none": 0.40818562039434375,
|
||||
"acc_norm_stderr,none": 0.004904933500255884
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 14.485555582236119,
|
||||
"perplexity_stderr,none": 0.4358013409476018,
|
||||
"acc,none": 0.4422666407917718,
|
||||
"acc_stderr,none": 0.006919384666875831
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.188,
|
||||
"acc_stderr,none": 0.01749067888034625,
|
||||
"acc_norm,none": 0.28,
|
||||
"acc_norm_stderr,none": 0.020099950647503237
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6806311207834603,
|
||||
"acc_stderr,none": 0.010877964076613737,
|
||||
"acc_norm,none": 0.6692056583242655,
|
||||
"acc_norm_stderr,none": 0.010977520584714429
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.892,
|
||||
"acc_stderr,none": 0.009820001651345682,
|
||||
"acc_norm,none": 0.887,
|
||||
"acc_norm_stderr,none": 0.01001655286669685
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 34.50450469911897,
|
||||
"byte_perplexity,none": 1.7927778872125213,
|
||||
"bits_per_byte,none": 0.842196759334895
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5335438042620363,
|
||||
"acc_stderr,none": 0.014020826677598103
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7efb86530040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7efb86502ef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7efb86503370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7efb865035b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=EleutherAI/pythia-410m",
|
||||
"num_fewshot": 5,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
hf (pretrained=EleutherAI/pythia-410m), limit: None, num_fewshot: 5, batch_size: 16
|
||||
| Task |Version|Filter| Metric | Value | |Stderr|
|
||||
|--------------|-------|------|---------------|------:|---|-----:|
|
||||
|arc_challenge |Yaml |none |acc | 0.2184|± |0.0121|
|
||||
| | |none |acc_norm | 0.2645|± |0.0129|
|
||||
|arc_easy |Yaml |none |acc | 0.5450|± |0.0102|
|
||||
| | |none |acc_norm | 0.5370|± |0.0102|
|
||||
|boolq |Yaml |none |acc | 0.4872|± |0.0087|
|
||||
|hellaswag |Yaml |none |acc | 0.3383|± |0.0047|
|
||||
| | |none |acc_norm | 0.4082|± |0.0049|
|
||||
|lambada_openai|Yaml |none |perplexity |14.4856|± |0.4358|
|
||||
| | |none |acc | 0.4423|± |0.0069|
|
||||
|openbookqa |Yaml |none |acc | 0.1880|± |0.0175|
|
||||
| | |none |acc_norm | 0.2800|± |0.0201|
|
||||
|piqa |Yaml |none |acc | 0.6806|± |0.0109|
|
||||
| | |none |acc_norm | 0.6692|± |0.0110|
|
||||
|sciq |Yaml |none |acc | 0.8920|± |0.0098|
|
||||
| | |none |acc_norm | 0.8870|± |0.0100|
|
||||
|wikitext |Yaml |none |word_perplexity|34.5045| | |
|
||||
| | |none |byte_perplexity| 1.7928| | |
|
||||
| | |none |bits_per_byte | 0.8422| | |
|
||||
|winogrande |Yaml |none |acc | 0.5335|± |0.0140|
|
||||
|
||||
406
base-410m-eval-files/EleutherAI-pythia-410m-5shot/results.json
Normal file
406
base-410m-eval-files/EleutherAI-pythia-410m-5shot/results.json
Normal file
@@ -0,0 +1,406 @@
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.21843003412969283,
|
||||
"acc_stderr,none": 0.012074291605700959,
|
||||
"acc_norm,none": 0.2645051194539249,
|
||||
"acc_norm_stderr,none": 0.012889272949313368
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.54503367003367,
|
||||
"acc_stderr,none": 0.010218084454602589,
|
||||
"acc_norm,none": 0.5370370370370371,
|
||||
"acc_norm_stderr,none": 0.010231597249131058
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.4871559633027523,
|
||||
"acc_stderr,none": 0.008742169169427067
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.33827922724556864,
|
||||
"acc_stderr,none": 0.004721571443354456,
|
||||
"acc_norm,none": 0.40818562039434375,
|
||||
"acc_norm_stderr,none": 0.004904933500255884
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 14.485555582236119,
|
||||
"perplexity_stderr,none": 0.4358013409476018,
|
||||
"acc,none": 0.4422666407917718,
|
||||
"acc_stderr,none": 0.006919384666875831
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.188,
|
||||
"acc_stderr,none": 0.01749067888034625,
|
||||
"acc_norm,none": 0.28,
|
||||
"acc_norm_stderr,none": 0.020099950647503237
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6806311207834603,
|
||||
"acc_stderr,none": 0.010877964076613737,
|
||||
"acc_norm,none": 0.6692056583242655,
|
||||
"acc_norm_stderr,none": 0.010977520584714429
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.892,
|
||||
"acc_stderr,none": 0.009820001651345682,
|
||||
"acc_norm,none": 0.887,
|
||||
"acc_norm_stderr,none": 0.01001655286669685
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 34.50450469911897,
|
||||
"byte_perplexity,none": 1.7927778872125213,
|
||||
"bits_per_byte,none": 0.842196759334895
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5335438042620363,
|
||||
"acc_stderr,none": 0.014020826677598103
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7efb86530040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7efb86502ef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7efb86503370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7efb865035b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=EleutherAI/pythia-410m",
|
||||
"num_fewshot": 5,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
24
config.json
Normal file
24
config.json
Normal file
@@ -0,0 +1,24 @@
|
||||
{
|
||||
"architectures": [
|
||||
"GPTNeoXForCausalLM"
|
||||
],
|
||||
"bos_token_id": 0,
|
||||
"eos_token_id": 0,
|
||||
"hidden_act": "gelu",
|
||||
"hidden_size": 1024,
|
||||
"initializer_range": 0.02,
|
||||
"intermediate_size": 4096,
|
||||
"layer_norm_eps": 1e-05,
|
||||
"max_position_embeddings": 2048,
|
||||
"model_type": "gpt_neox",
|
||||
"num_attention_heads": 16,
|
||||
"num_hidden_layers": 24,
|
||||
"rotary_emb_base": 10000,
|
||||
"rotary_pct": 0.25,
|
||||
"tie_word_embeddings": false,
|
||||
"torch_dtype": "float16",
|
||||
"transformers_version": "4.24.0",
|
||||
"use_cache": true,
|
||||
"use_parallel_residual": true,
|
||||
"vocab_size": 50304
|
||||
}
|
||||
44
config.yaml
Normal file
44
config.yaml
Normal file
@@ -0,0 +1,44 @@
|
||||
seed: 0
|
||||
exp_name: pythia410m_hh_ga4_dpo
|
||||
batch_size: 64
|
||||
eval_batch_size: 32
|
||||
debug: false
|
||||
fsdp_port: 12355
|
||||
datasets:
|
||||
- hh
|
||||
wandb:
|
||||
enabled: true
|
||||
entity: pythia_dpo
|
||||
project: Pythia_LOM
|
||||
local_dirs:
|
||||
- /scr-ssd
|
||||
- /scr
|
||||
- .cache
|
||||
sample_during_eval: false
|
||||
n_eval_model_samples: 16
|
||||
do_first_eval: true
|
||||
local_run_dir: .cache/laura/pythia410m_hh_ga4_dpo_2023-07-19_13-02-35_532776
|
||||
lr: 1.0e-06
|
||||
gradient_accumulation_steps: 4
|
||||
max_grad_norm: 10.0
|
||||
max_length: 512
|
||||
max_prompt_length: 256
|
||||
n_epochs: 1
|
||||
n_examples: null
|
||||
n_eval_examples: 256
|
||||
trainer: FSDPTrainer
|
||||
optimizer: RMSprop
|
||||
warmup_steps: 150
|
||||
activation_checkpointing: false
|
||||
eval_every: 19968
|
||||
minimum_log_interval_secs: 1.0
|
||||
model:
|
||||
name_or_path: EleutherAI/pythia-410m
|
||||
tokenizer_name_or_path: null
|
||||
archive: null
|
||||
block_name: GPTNeoXLayer
|
||||
policy_dtype: float32
|
||||
fsdp_policy_mp: bfloat16
|
||||
reference_dtype: float16
|
||||
loss:
|
||||
name: sft
|
||||
3
optimizer.pt
Normal file
3
optimizer.pt
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:d61e9f7ae162b6ca0f3e6f4310f307a45d8d5f0718a757228d7f9e8d877108df
|
||||
size 1621435717
|
||||
3
policy.pt
Normal file
3
policy.pt
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:f7d5342d3950db163bc8bacb6d0a7eac25502526a3066e30cfd134a6f9da6d9a
|
||||
size 1722144187
|
||||
3
pytorch_model.bin
Normal file
3
pytorch_model.bin
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:d407713fe4a1dc3df845244470f3293ad33f47514896e30b83f69d9384e11d44
|
||||
size 1722124477
|
||||
3
scheduler.pt
Normal file
3
scheduler.pt
Normal file
@@ -0,0 +1,3 @@
|
||||
version https://git-lfs.github.com/spec/v1
|
||||
oid sha256:ef3bb892ec5f85ebcd911ec0358781fc843702c92bef54eec6bc38b127a23878
|
||||
size 627
|
||||
430
sft-410m-eval-files/dpo-pythia-410m-0shot-shelloutput.txt
Normal file
430
sft-410m-eval-files/dpo-pythia-410m-0shot-shelloutput.txt
Normal file
@@ -0,0 +1,430 @@
|
||||
bootstrapping for stddev: perplexity
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.2431740614334471,
|
||||
"acc_stderr,none": 0.012536554144587082,
|
||||
"acc_norm,none": 0.27303754266211605,
|
||||
"acc_norm_stderr,none": 0.013019332762635751
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.5441919191919192,
|
||||
"acc_stderr,none": 0.010219631763437853,
|
||||
"acc_norm,none": 0.47853535353535354,
|
||||
"acc_norm_stderr,none": 0.010250325159456656
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.618960244648318,
|
||||
"acc_stderr,none": 0.008493937524439329
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.3488348934475204,
|
||||
"acc_stderr,none": 0.004756275875018267,
|
||||
"acc_norm,none": 0.41545508862776337,
|
||||
"acc_norm_stderr,none": 0.004917931778593191
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 9.320713747126444,
|
||||
"perplexity_stderr,none": 0.32204382695401407,
|
||||
"acc,none": 0.5148457209392587,
|
||||
"acc_stderr,none": 0.006962906440875391
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.204,
|
||||
"acc_stderr,none": 0.018039369104138645,
|
||||
"acc_norm,none": 0.3,
|
||||
"acc_norm_stderr,none": 0.02051442622562805
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6789989118607181,
|
||||
"acc_stderr,none": 0.010892641574707904,
|
||||
"acc_norm,none": 0.6811751904243744,
|
||||
"acc_norm_stderr,none": 0.010873037534333416
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.809,
|
||||
"acc_stderr,none": 0.012436787112179501,
|
||||
"acc_norm,none": 0.717,
|
||||
"acc_norm_stderr,none": 0.01425181090648174
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 38.90357956698245,
|
||||
"byte_perplexity,none": 1.8285957273924667,
|
||||
"bits_per_byte,none": 0.8707361540620543
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5272296764009471,
|
||||
"acc_stderr,none": 0.014031631629827694
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7f40312f8040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7f40312caef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7f40312cb370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7f40312cb5b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=lomahony/eleuther-pythia410m-hh-dpo",
|
||||
"num_fewshot": 0,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
hf (pretrained=lomahony/eleuther-pythia410m-hh-dpo), limit: None, num_fewshot: 0, batch_size: 16
|
||||
| Task |Version|Filter| Metric | Value | |Stderr|
|
||||
|--------------|-------|------|---------------|------:|---|-----:|
|
||||
|arc_challenge |Yaml |none |acc | 0.2432|± |0.0125|
|
||||
| | |none |acc_norm | 0.2730|± |0.0130|
|
||||
|arc_easy |Yaml |none |acc | 0.5442|± |0.0102|
|
||||
| | |none |acc_norm | 0.4785|± |0.0103|
|
||||
|boolq |Yaml |none |acc | 0.6190|± |0.0085|
|
||||
|hellaswag |Yaml |none |acc | 0.3488|± |0.0048|
|
||||
| | |none |acc_norm | 0.4155|± |0.0049|
|
||||
|lambada_openai|Yaml |none |perplexity | 9.3207|± |0.3220|
|
||||
| | |none |acc | 0.5148|± |0.0070|
|
||||
|openbookqa |Yaml |none |acc | 0.2040|± |0.0180|
|
||||
| | |none |acc_norm | 0.3000|± |0.0205|
|
||||
|piqa |Yaml |none |acc | 0.6790|± |0.0109|
|
||||
| | |none |acc_norm | 0.6812|± |0.0109|
|
||||
|sciq |Yaml |none |acc | 0.8090|± |0.0124|
|
||||
| | |none |acc_norm | 0.7170|± |0.0143|
|
||||
|wikitext |Yaml |none |word_perplexity|38.9036| | |
|
||||
| | |none |byte_perplexity| 1.8286| | |
|
||||
| | |none |bits_per_byte | 0.8707| | |
|
||||
|winogrande |Yaml |none |acc | 0.5272|± |0.0140|
|
||||
|
||||
430
sft-410m-eval-files/dpo-pythia-410m-5shot-shelloutput.txt
Normal file
430
sft-410m-eval-files/dpo-pythia-410m-5shot-shelloutput.txt
Normal file
@@ -0,0 +1,430 @@
|
||||
bootstrapping for stddev: perplexity
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.25426621160409557,
|
||||
"acc_stderr,none": 0.012724999945157746,
|
||||
"acc_norm,none": 0.2883959044368601,
|
||||
"acc_norm_stderr,none": 0.013238394422428173
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.5732323232323232,
|
||||
"acc_stderr,none": 0.010149141043955643,
|
||||
"acc_norm,none": 0.5542929292929293,
|
||||
"acc_norm_stderr,none": 0.010199118183322989
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.5715596330275229,
|
||||
"acc_stderr,none": 0.008655028561519784
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.3482374029077873,
|
||||
"acc_stderr,none": 0.004754380554929206,
|
||||
"acc_norm,none": 0.41236805417247563,
|
||||
"acc_norm_stderr,none": 0.00491254704013287
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 14.093662143573109,
|
||||
"perplexity_stderr,none": 0.48659081977443114,
|
||||
"acc,none": 0.42907044440131964,
|
||||
"acc_stderr,none": 0.006895529739245138
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.19,
|
||||
"acc_stderr,none": 0.017561800410758988,
|
||||
"acc_norm,none": 0.284,
|
||||
"acc_norm_stderr,none": 0.020186703693570857
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6833514689880305,
|
||||
"acc_stderr,none": 0.010853160531978481,
|
||||
"acc_norm,none": 0.6828073993471164,
|
||||
"acc_norm_stderr,none": 0.01085815545438087
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.864,
|
||||
"acc_stderr,none": 0.010845350230472988,
|
||||
"acc_norm,none": 0.85,
|
||||
"acc_norm_stderr,none": 0.011297239823409314
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 38.90357956698245,
|
||||
"byte_perplexity,none": 1.8285957273924667,
|
||||
"bits_per_byte,none": 0.8707361540620543
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5272296764009471,
|
||||
"acc_stderr,none": 0.014031631629827694
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7fd4f5060040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7fd4f5032ef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7fd4f5033370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7fd4f50335b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=lomahony/eleuther-pythia410m-hh-dpo",
|
||||
"num_fewshot": 5,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
hf (pretrained=lomahony/eleuther-pythia410m-hh-dpo), limit: None, num_fewshot: 5, batch_size: 16
|
||||
| Task |Version|Filter| Metric | Value | |Stderr|
|
||||
|--------------|-------|------|---------------|------:|---|-----:|
|
||||
|arc_challenge |Yaml |none |acc | 0.2543|± |0.0127|
|
||||
| | |none |acc_norm | 0.2884|± |0.0132|
|
||||
|arc_easy |Yaml |none |acc | 0.5732|± |0.0101|
|
||||
| | |none |acc_norm | 0.5543|± |0.0102|
|
||||
|boolq |Yaml |none |acc | 0.5716|± |0.0087|
|
||||
|hellaswag |Yaml |none |acc | 0.3482|± |0.0048|
|
||||
| | |none |acc_norm | 0.4124|± |0.0049|
|
||||
|lambada_openai|Yaml |none |perplexity |14.0937|± |0.4866|
|
||||
| | |none |acc | 0.4291|± |0.0069|
|
||||
|openbookqa |Yaml |none |acc | 0.1900|± |0.0176|
|
||||
| | |none |acc_norm | 0.2840|± |0.0202|
|
||||
|piqa |Yaml |none |acc | 0.6834|± |0.0109|
|
||||
| | |none |acc_norm | 0.6828|± |0.0109|
|
||||
|sciq |Yaml |none |acc | 0.8640|± |0.0108|
|
||||
| | |none |acc_norm | 0.8500|± |0.0113|
|
||||
|wikitext |Yaml |none |word_perplexity|38.9036| | |
|
||||
| | |none |byte_perplexity| 1.8286| | |
|
||||
| | |none |bits_per_byte | 0.8707| | |
|
||||
|winogrande |Yaml |none |acc | 0.5272|± |0.0140|
|
||||
|
||||
430
sft-410m-eval-files/sft-pythia-410m-0shot-shelloutput.txt
Normal file
430
sft-410m-eval-files/sft-pythia-410m-0shot-shelloutput.txt
Normal file
@@ -0,0 +1,430 @@
|
||||
bootstrapping for stddev: perplexity
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.22610921501706485,
|
||||
"acc_stderr,none": 0.012224202097063302,
|
||||
"acc_norm,none": 0.257679180887372,
|
||||
"acc_norm_stderr,none": 0.012780770562768402
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.5256734006734006,
|
||||
"acc_stderr,none": 0.01024624966559123,
|
||||
"acc_norm,none": 0.45496632996632996,
|
||||
"acc_norm_stderr,none": 0.010218084454602594
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.618348623853211,
|
||||
"acc_stderr,none": 0.00849655074117826
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.34156542521410077,
|
||||
"acc_stderr,none": 0.004732654295724439,
|
||||
"acc_norm,none": 0.4032065325632344,
|
||||
"acc_norm_stderr,none": 0.004895390341445623
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 9.865369793814358,
|
||||
"perplexity_stderr,none": 0.31823636930011806,
|
||||
"acc,none": 0.5088298078789055,
|
||||
"acc_stderr,none": 0.006964891360529503
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.192,
|
||||
"acc_stderr,none": 0.017632180454360998,
|
||||
"acc_norm,none": 0.3,
|
||||
"acc_norm_stderr,none": 0.02051442622562805
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6806311207834603,
|
||||
"acc_stderr,none": 0.010877964076613742,
|
||||
"acc_norm,none": 0.6833514689880305,
|
||||
"acc_norm_stderr,none": 0.010853160531978483
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.811,
|
||||
"acc_stderr,none": 0.012386784588117719,
|
||||
"acc_norm,none": 0.715,
|
||||
"acc_norm_stderr,none": 0.014282120955200473
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 35.96785758155289,
|
||||
"byte_perplexity,none": 1.8050958522509775,
|
||||
"bits_per_byte,none": 0.8520754475808856
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5327545382794001,
|
||||
"acc_stderr,none": 0.014022300570434137
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7f9bda880040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7f9bda852ef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7f9bda853370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7f9bda8535b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=lomahony/eleuther-pythia410m-hh-sft",
|
||||
"num_fewshot": 0,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
hf (pretrained=lomahony/eleuther-pythia410m-hh-sft), limit: None, num_fewshot: 0, batch_size: 16
|
||||
| Task |Version|Filter| Metric | Value | |Stderr|
|
||||
|--------------|-------|------|---------------|------:|---|-----:|
|
||||
|arc_challenge |Yaml |none |acc | 0.2261|± |0.0122|
|
||||
| | |none |acc_norm | 0.2577|± |0.0128|
|
||||
|arc_easy |Yaml |none |acc | 0.5257|± |0.0102|
|
||||
| | |none |acc_norm | 0.4550|± |0.0102|
|
||||
|boolq |Yaml |none |acc | 0.6183|± |0.0085|
|
||||
|hellaswag |Yaml |none |acc | 0.3416|± |0.0047|
|
||||
| | |none |acc_norm | 0.4032|± |0.0049|
|
||||
|lambada_openai|Yaml |none |perplexity | 9.8654|± |0.3182|
|
||||
| | |none |acc | 0.5088|± |0.0070|
|
||||
|openbookqa |Yaml |none |acc | 0.1920|± |0.0176|
|
||||
| | |none |acc_norm | 0.3000|± |0.0205|
|
||||
|piqa |Yaml |none |acc | 0.6806|± |0.0109|
|
||||
| | |none |acc_norm | 0.6834|± |0.0109|
|
||||
|sciq |Yaml |none |acc | 0.8110|± |0.0124|
|
||||
| | |none |acc_norm | 0.7150|± |0.0143|
|
||||
|wikitext |Yaml |none |word_perplexity|35.9679| | |
|
||||
| | |none |byte_perplexity| 1.8051| | |
|
||||
| | |none |bits_per_byte | 0.8521| | |
|
||||
|winogrande |Yaml |none |acc | 0.5328|± |0.0140|
|
||||
|
||||
406
sft-410m-eval-files/sft-pythia-410m-0shot/results.json
Normal file
406
sft-410m-eval-files/sft-pythia-410m-0shot/results.json
Normal file
@@ -0,0 +1,406 @@
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.22610921501706485,
|
||||
"acc_stderr,none": 0.012224202097063302,
|
||||
"acc_norm,none": 0.257679180887372,
|
||||
"acc_norm_stderr,none": 0.012780770562768402
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.5256734006734006,
|
||||
"acc_stderr,none": 0.01024624966559123,
|
||||
"acc_norm,none": 0.45496632996632996,
|
||||
"acc_norm_stderr,none": 0.010218084454602594
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.618348623853211,
|
||||
"acc_stderr,none": 0.00849655074117826
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.34156542521410077,
|
||||
"acc_stderr,none": 0.004732654295724439,
|
||||
"acc_norm,none": 0.4032065325632344,
|
||||
"acc_norm_stderr,none": 0.004895390341445623
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 9.865369793814358,
|
||||
"perplexity_stderr,none": 0.31823636930011806,
|
||||
"acc,none": 0.5088298078789055,
|
||||
"acc_stderr,none": 0.006964891360529503
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.192,
|
||||
"acc_stderr,none": 0.017632180454360998,
|
||||
"acc_norm,none": 0.3,
|
||||
"acc_norm_stderr,none": 0.02051442622562805
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6806311207834603,
|
||||
"acc_stderr,none": 0.010877964076613742,
|
||||
"acc_norm,none": 0.6833514689880305,
|
||||
"acc_norm_stderr,none": 0.010853160531978483
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.811,
|
||||
"acc_stderr,none": 0.012386784588117719,
|
||||
"acc_norm,none": 0.715,
|
||||
"acc_norm_stderr,none": 0.014282120955200473
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 35.96785758155289,
|
||||
"byte_perplexity,none": 1.8050958522509775,
|
||||
"bits_per_byte,none": 0.8520754475808856
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5327545382794001,
|
||||
"acc_stderr,none": 0.014022300570434137
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7f9bda880040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7f9bda852ef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7f9bda853370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7f9bda8535b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 0,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=lomahony/eleuther-pythia410m-hh-sft",
|
||||
"num_fewshot": 0,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
514
sft-410m-eval-files/sft-pythia-410m-5shot-shelloutput.txt
Normal file
514
sft-410m-eval-files/sft-pythia-410m-5shot-shelloutput.txt
Normal file
File diff suppressed because one or more lines are too long
406
sft-410m-eval-files/sft-pythia-410m-5shot/results.json
Normal file
406
sft-410m-eval-files/sft-pythia-410m-5shot/results.json
Normal file
@@ -0,0 +1,406 @@
|
||||
{
|
||||
"results": {
|
||||
"arc_challenge": {
|
||||
"acc,none": 0.2363481228668942,
|
||||
"acc_stderr,none": 0.012414960524301842,
|
||||
"acc_norm,none": 0.2696245733788396,
|
||||
"acc_norm_stderr,none": 0.01296804068686915
|
||||
},
|
||||
"arc_easy": {
|
||||
"acc,none": 0.5505050505050505,
|
||||
"acc_stderr,none": 0.010207308833916047,
|
||||
"acc_norm,none": 0.5290404040404041,
|
||||
"acc_norm_stderr,none": 0.01024246382639562
|
||||
},
|
||||
"boolq": {
|
||||
"acc,none": 0.5015290519877675,
|
||||
"acc_stderr,none": 0.008745014084425922
|
||||
},
|
||||
"hellaswag": {
|
||||
"acc,none": 0.33808006373232424,
|
||||
"acc_stderr,none": 0.004720891597174752,
|
||||
"acc_norm,none": 0.40111531567416847,
|
||||
"acc_norm_stderr,none": 0.004891226138578059
|
||||
},
|
||||
"lambada_openai": {
|
||||
"perplexity,none": 14.092313022096198,
|
||||
"perplexity_stderr,none": 0.4548197909283815,
|
||||
"acc,none": 0.44343101106151755,
|
||||
"acc_stderr,none": 0.006921251108304399
|
||||
},
|
||||
"openbookqa": {
|
||||
"acc,none": 0.178,
|
||||
"acc_stderr,none": 0.017123622189062257,
|
||||
"acc_norm,none": 0.288,
|
||||
"acc_norm_stderr,none": 0.020271503835075217
|
||||
},
|
||||
"piqa": {
|
||||
"acc,none": 0.6860718171926007,
|
||||
"acc_stderr,none": 0.010827928134189646,
|
||||
"acc_norm,none": 0.6844396082698585,
|
||||
"acc_norm_stderr,none": 0.010843119201758934
|
||||
},
|
||||
"sciq": {
|
||||
"acc,none": 0.886,
|
||||
"acc_stderr,none": 0.01005510343582333,
|
||||
"acc_norm,none": 0.879,
|
||||
"acc_norm_stderr,none": 0.010318210380946094
|
||||
},
|
||||
"wikitext": {
|
||||
"word_perplexity,none": 35.96785758155289,
|
||||
"byte_perplexity,none": 1.8050958522509775,
|
||||
"bits_per_byte,none": 0.8520754475808856
|
||||
},
|
||||
"winogrande": {
|
||||
"acc,none": 0.5327545382794001,
|
||||
"acc_stderr,none": 0.014022300570434137
|
||||
}
|
||||
},
|
||||
"configs": {
|
||||
"arc_challenge": {
|
||||
"task": "arc_challenge",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Challenge",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"arc_easy": {
|
||||
"task": "arc_easy",
|
||||
"group": [
|
||||
"ai2_arc",
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "ai2_arc",
|
||||
"dataset_name": "ARC-Easy",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "Question: {{question}}\nAnswer:",
|
||||
"doc_to_target": "{{choices.label.index(answerKey)}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "Question: {{question}}\nAnswer:"
|
||||
},
|
||||
"boolq": {
|
||||
"task": "boolq",
|
||||
"group": [
|
||||
"super-glue-lm-eval-v1"
|
||||
],
|
||||
"dataset_path": "super_glue",
|
||||
"dataset_name": "boolq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{{passage}}\nQuestion: {{question}}?\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": [
|
||||
"no",
|
||||
"yes"
|
||||
],
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc"
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "passage"
|
||||
},
|
||||
"hellaswag": {
|
||||
"task": "hellaswag",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "hellaswag",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "{% set text = activity_label ~ ': ' ~ ctx_a ~ ' ' ~ ctx_b.capitalize() %}{{text|trim|replace(' [title]', '. ')|regex_replace('\\[.*?\\]', '')|replace(' ', ' ')}}",
|
||||
"doc_to_target": "{{label}}",
|
||||
"doc_to_choice": "{{endings|map('trim')|map('replace', ' [title]', '. ')|map('regex_replace', '\\[.*?\\]', '')|map('replace', ' ', ' ')|list}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
},
|
||||
"lambada_openai": {
|
||||
"task": "lambada_openai",
|
||||
"group": [
|
||||
"lambada",
|
||||
"loglikelihood",
|
||||
"perplexity"
|
||||
],
|
||||
"dataset_path": "EleutherAI/lambada_openai",
|
||||
"dataset_name": "default",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "{{text.split(' ')[:-1]|join(' ')}}",
|
||||
"doc_to_target": "{{' '+text.split(' ')[-1]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "perplexity",
|
||||
"aggregation": "perplexity",
|
||||
"higher_is_better": false
|
||||
},
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{text}}"
|
||||
},
|
||||
"openbookqa": {
|
||||
"task": "openbookqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "openbookqa",
|
||||
"dataset_name": "main",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "question_stem",
|
||||
"doc_to_target": "{{choices.label.index(answerKey.lstrip())}}",
|
||||
"doc_to_choice": "{{choices.text}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "question_stem"
|
||||
},
|
||||
"piqa": {
|
||||
"task": "piqa",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "piqa",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "Question: {{goal}}\nAnswer:",
|
||||
"doc_to_target": "label",
|
||||
"doc_to_choice": "{{[sol1, sol2]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "goal"
|
||||
},
|
||||
"sciq": {
|
||||
"task": "sciq",
|
||||
"group": [
|
||||
"multiple_choice"
|
||||
],
|
||||
"dataset_path": "sciq",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"doc_to_text": "{{support.lstrip()}}\nQuestion: {{question}}\nAnswer:",
|
||||
"doc_to_target": 3,
|
||||
"doc_to_choice": "{{[distractor1, distractor2, distractor3, correct_answer]}}",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
},
|
||||
{
|
||||
"metric": "acc_norm",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{support}} {{question}}"
|
||||
},
|
||||
"wikitext": {
|
||||
"task": "wikitext",
|
||||
"group": [
|
||||
"perplexity",
|
||||
"loglikelihood_rolling"
|
||||
],
|
||||
"dataset_path": "EleutherAI/wikitext_document_level",
|
||||
"dataset_name": "wikitext-2-raw-v1",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"test_split": "test",
|
||||
"template_aliases": "",
|
||||
"doc_to_text": "",
|
||||
"doc_to_target": "<function wikitext_detokenizer at 0x7f9fa4a64040>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "word_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "byte_perplexity"
|
||||
},
|
||||
{
|
||||
"metric": "bits_per_byte"
|
||||
}
|
||||
],
|
||||
"output_type": "loglikelihood_rolling",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": true,
|
||||
"doc_to_decontamination_query": "{{page}}"
|
||||
},
|
||||
"winogrande": {
|
||||
"task": "winogrande",
|
||||
"dataset_path": "winogrande",
|
||||
"dataset_name": "winogrande_xl",
|
||||
"training_split": "train",
|
||||
"validation_split": "validation",
|
||||
"doc_to_text": "<function doc_to_text at 0x7f9fa4a36ef0>",
|
||||
"doc_to_target": "<function doc_to_target at 0x7f9fa4a37370>",
|
||||
"doc_to_choice": "<function doc_to_choice at 0x7f9fa4a375b0>",
|
||||
"description": "",
|
||||
"target_delimiter": " ",
|
||||
"fewshot_delimiter": "\n\n",
|
||||
"num_fewshot": 5,
|
||||
"metric_list": [
|
||||
{
|
||||
"metric": "acc",
|
||||
"aggregation": "mean",
|
||||
"higher_is_better": true
|
||||
}
|
||||
],
|
||||
"output_type": "multiple_choice",
|
||||
"repeats": 1,
|
||||
"should_decontaminate": false
|
||||
}
|
||||
},
|
||||
"versions": {
|
||||
"arc_challenge": "Yaml",
|
||||
"arc_easy": "Yaml",
|
||||
"boolq": "Yaml",
|
||||
"hellaswag": "Yaml",
|
||||
"lambada_openai": "Yaml",
|
||||
"openbookqa": "Yaml",
|
||||
"piqa": "Yaml",
|
||||
"sciq": "Yaml",
|
||||
"wikitext": "Yaml",
|
||||
"winogrande": "Yaml"
|
||||
},
|
||||
"config": {
|
||||
"model": "hf",
|
||||
"model_args": "pretrained=lomahony/eleuther-pythia410m-hh-sft",
|
||||
"num_fewshot": 5,
|
||||
"batch_size": 16,
|
||||
"batch_sizes": [],
|
||||
"device": "cuda:0",
|
||||
"use_cache": null,
|
||||
"limit": null,
|
||||
"bootstrap_iters": 100000
|
||||
},
|
||||
"git_hash": "4e44f0a"
|
||||
}
|
||||
5
special_tokens_map.json
Normal file
5
special_tokens_map.json
Normal file
@@ -0,0 +1,5 @@
|
||||
{
|
||||
"bos_token": "<|endoftext|>",
|
||||
"eos_token": "<|endoftext|>",
|
||||
"unk_token": "<|endoftext|>"
|
||||
}
|
||||
100528
tokenizer.json
Normal file
100528
tokenizer.json
Normal file
File diff suppressed because it is too large
Load Diff
9
tokenizer_config.json
Normal file
9
tokenizer_config.json
Normal file
@@ -0,0 +1,9 @@
|
||||
{
|
||||
"add_prefix_space": false,
|
||||
"bos_token": "<|endoftext|>",
|
||||
"eos_token": "<|endoftext|>",
|
||||
"name_or_path": "EleutherAI/gpt-neox-20b",
|
||||
"special_tokens_map_file": "/admin/home-hailey/.cache/huggingface/hub/models--EleutherAI--gpt-neox-20b/snapshots/4e49eadb5d14bd22f314ec3f45b69a87b88c7691/special_tokens_map.json",
|
||||
"tokenizer_class": "GPTNeoXTokenizer",
|
||||
"unk_token": "<|endoftext|>"
|
||||
}
|
||||
Reference in New Issue
Block a user