Datasets:
Download model_inference.sh from uilab/BLEnD: direct link, hf CLI and curl.
- Browser
- Download file 3.88 kB
-
https://huggingface.co/datasets/uilab/BLEnD/resolve/8aa8498ecbac0c880b1e7d426defb84c78d219f5/model_inference.sh
- Command line
-
hf download hf://datasets/uilab/BLEnD@8aa8498ecbac0c880b1e7d426defb84c78d219f5/model_inference.sh
-
curl -L -o model_inference.sh https://huggingface.co/datasets/uilab/BLEnD/resolve/8aa8498ecbac0c880b1e7d426defb84c78d219f5/model_inference.sh
3.88 kB
| export CUDA_VISIBLE_DEVICES="" | |
| export HF_TOKEN="" | |
| export COHERE_API_KEY="" | |
| export OPENAI_API_KEY="" | |
| export OPENAI_ORG_ID="" | |
| export AZURE_OPENAI_API_KEY="" | |
| export AZURE_OPENAI_API_VER="" | |
| export AZURE_OPENAI_API_ENDPT="" | |
| export CLAUDE_API_KEY="" | |
| export GOOGLE_API_KEY="" | |
| export GOOGLE_APPLICATION_CREDENTIALS="" | |
| export GOOGLE_PROJECT_NAME="" | |
| # Define model keys | |
| MODEL_KEYS=( | |
| "gpt-4-1106-preview" | |
| "gpt-3.5-turbo-1106" | |
| "aya-101" | |
| "gemini-pro" | |
| "claude-3-opus-20240229" | |
| "claude-3-sonnet-20240229" | |
| "claude-3-haiku-20240307" | |
| "Qwen1.5-72B-Chat" | |
| "Qwen1.5-14B-Chat" | |
| "Qwen1.5-32B-Chat" | |
| "text-bison-002" | |
| "c4ai-command-r-v01" | |
| "c4ai-command-r-plus" | |
| "aya-23" | |
| "SeaLLM-7B-v2.5" | |
| "Merak-7B-v4" | |
| "jais-13b-chat" | |
| ) | |
| # Define countries and languages as an associative array | |
| declare -A COUNTRY_LANG | |
| COUNTRY_LANG["UK"]="English" | |
| COUNTRY_LANG["US"]="English" | |
| COUNTRY_LANG["South_Korea"]="Korean" | |
| COUNTRY_LANG["Algeria"]="Arabic" | |
| COUNTRY_LANG["China"]="Chinese" | |
| COUNTRY_LANG["Indonesia"]="Indonesian" | |
| COUNTRY_LANG["Spain"]="Spanish" | |
| COUNTRY_LANG["Iran"]="Persian" | |
| COUNTRY_LANG["Mexico"]="Spanish" | |
| COUNTRY_LANG["Assam"]="Assamese" | |
| COUNTRY_LANG["Greece"]="Greek" | |
| COUNTRY_LANG["Ethiopia"]="Amharic" | |
| COUNTRY_LANG["Northern_Nigeria"]="Hausa" | |
| COUNTRY_LANG["Azerbaijan"]="Azerbaijani" | |
| COUNTRY_LANG["North_Korea"]="Korean" | |
| COUNTRY_LANG["West_Java"]="Sundanese" | |
| # Prompt numbers | |
| PROMPT_NUMBERS=("inst-4" "pers-3") | |
| # Iterate over models, countries, languages, and prompts | |
| for model_key in "${MODEL_KEYS[@]}"; do | |
| for country in "${!COUNTRY_LANG[@]}"; do | |
| language="${COUNTRY_LANG[$country]}" | |
| for prompt_no in "${PROMPT_NUMBERS[@]}"; do | |
| python model_inference.py --model "$model_key" \ | |
| --language "$language" \ | |
| --country "$country" \ | |
| --question_dir "./data/questions" \ | |
| --question_file "${country}_questions.csv" \ | |
| --question_col Translation \ | |
| --prompt_dir "./data/prompts" \ | |
| --prompt_file "${country}_prompts.csv" \ | |
| --prompt_no "$prompt_no" \ | |
| --id_col ID \ | |
| --output_dir "./model_inference_results" \ | |
| --output_file "${model_key}-${country}_${language}_${prompt_no}_result.csv" \ | |
| --model_cache_dir ".cache" \ | |
| --gpt_azure "True" \ | |
| --temperature 0 \ | |
| --top_p 1 | |
| if [ "$language" != "English" ]; then | |
| python model_inference.py --model "$model_key" \ | |
| --language "$language" \ | |
| --country "$country" \ | |
| --question_dir "./data/questions" \ | |
| --question_file "${country}_questions.csv" \ | |
| --question_col Question \ | |
| --prompt_dir "./data/prompts" \ | |
| --prompt_file "${country}_prompts.csv" \ | |
| --prompt_no "$prompt_no" \ | |
| --id_col ID \ | |
| --output_dir "./model_inference_results" \ | |
| --output_file "${model_key}-${country}_English_${prompt_no}_result.csv" \ | |
| --model_cache_dir ".cache" \ | |
| --gpt_azure "True" \ | |
| --temperature 0 \ | |
| --top_p 1 | |
| fi | |
| done | |
| done | |
| done | |