Toyota Research Institute OpenAIは深層学習と自然言語処理を 使って、超高性能なチャットボットを開 発しています 「W&Bは、一人の研究者のインサイトを チームに波及させ、一台のマシンに閉 じ込めず、何千台にスケールさせること ができます。」 WOJCIECH ZAREMBA Co-founder of OpenAI Genentechは深層学習を使って感染 病に有効な新しい抗生物質を 探しています 「W&Bのおかげで、私たちは意味の ある科学研究により多くの時間を費 やすことができます。」 Stephen Ra Machine Learning Lead
use run_name defined above github_version: v2.0.0 #for recording testmode: true # if you don't use api, please set "api" as "false" # if you use api, please select from "openai", "anthoropic", "google", "cohere" api: false model: use_wandb_artifacts: false artifacts_path: "" pretrained_model_name_or_path: 'mistralai/Mistral-7B-Instruct-v0.2' #if you use openai api, put the name of model trust_remote_code: true device_map: "auto" load_in_8bit: false load_in_4bit: false generator: top_p: 1.0 top_k: 0 temperature: 0.1 repetition_penalty: 1.0 tokenizer: use_wandb_artifacts: false artifacts_path: "" pretrained_model_name_or_path: "mistralai/Mistral-7B-Instruct-v0.2" use_fast: true config.yamlの設定 (概要、モデルとトークナイザ)
true, small dataset will be used referenceanswer_artifacts_path: 'wandb-japan/llm-leaderboard/mtbench_ja_referenceanswer:v0' # if testmode is true, small dataset will be used judge_prompt_artifacts_path: 'wandb-japan/llm-leaderboard/mtbench_ja_prompt:v1' bench_name: 'japanese_mt_bench' model_id: null # cannot use '<', '>', ':', '"', '/', '\\', '|', '?', '*', '.' question_begin: null question_end: null max_new_token: 1024 num_choices: 1 num_gpus_per_model: 1 num_gpus_total: 1 max_gpu_memory: null dtype: bfloat16 # None or float32 or float16 or bfloat16 # for gen_judgment judge_model: 'gpt-4' mode: 'single' baseline_model: null parallel: 1 first_n: null # for conv template # added custom_conv_template: true # the following variables will be used when custom_conv_template is set as true conv_name: "custom" conv_system_message: "" conv_roles: "('[INST]', '[/INST]')" conv_sep: "</s> " conv_stop_token_ids: "[2]" conv_stop_str: "</s> " conv_role_message_separator: " " conv_role_only_separator: " " config.yamlの設定 (Japanese MT-Bench)
use run_name defined above github_version: v2.0.0 #for recording testmode: true # if you don't use api, please set "api" as "false" # if you use api, please select from "openai", "anthoropic", "google", "cohere" api: false model: use_wandb_artifacts: false artifacts_path: "" pretrained_model_name_or_path: 'mistralai/Mistral-7B-Instruct-v0.2' #if you use openai api, put the name of model trust_remote_code: true device_map: "auto" load_in_8bit: false load_in_4bit: false generator: top_p: 1.0 top_k: 0 temperature: 0.1 repetition_penalty: 1.0 tokenizer: use_wandb_artifacts: false artifacts_path: "" pretrained_model_name_or_path: "mistralai/Mistral-7B-Instruct-v0.2" use_fast: true config.yamlのwandb.entityとprojectを対応して変更する 実際にはレポートには任意のプ ロジェクトから結果を挿入できる が、対応させた方が管理するの に良いだろう