mirror of https://github.com/hpcaitech/ColossalAI
remove real data path
parent
12fe8b5858
commit
b0e15d563e
|
@ -17,8 +17,8 @@ set_n_least_used_CUDA_VISIBLE_DEVICES 4
|
|||
|
||||
PROJECT_NAME="dpo"
|
||||
PARENT_CONFIG_FILE="./benchmark_config" # Path to a folder to save training config logs
|
||||
PRETRAINED_MODEL_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local tokenizer path
|
||||
PRETRAINED_MODEL_PATH="" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="" # huggingface or local tokenizer path
|
||||
BENCHMARK_DATA_DIR="./temp/dpo" # Path to benchmark data
|
||||
DATASET_SIZE=320
|
||||
|
||||
|
|
|
@ -17,8 +17,8 @@ set_n_least_used_CUDA_VISIBLE_DEVICES 4
|
|||
|
||||
PROJECT_NAME="kto"
|
||||
PARENT_CONFIG_FILE="./benchmark_config" # Path to a folder to save training config logs
|
||||
PRETRAINED_MODEL_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local tokenizer path
|
||||
PRETRAINED_MODEL_PATH="" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="" # huggingface or local tokenizer path
|
||||
BENCHMARK_DATA_DIR="./temp/kto" # Path to benchmark data
|
||||
DATASET_SIZE=80
|
||||
|
||||
|
|
|
@ -17,8 +17,8 @@ set_n_least_used_CUDA_VISIBLE_DEVICES 2
|
|||
|
||||
PROJECT_NAME="orpo"
|
||||
PARENT_CONFIG_FILE="./benchmark_config" # Path to a folder to save training config logs
|
||||
PRETRAINED_MODEL_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local tokenizer path
|
||||
PRETRAINED_MODEL_PATH="" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="" # huggingface or local tokenizer path
|
||||
BENCHMARK_DATA_DIR="./temp/orpo" # Path to benchmark data
|
||||
DATASET_SIZE=160
|
||||
|
||||
|
|
|
@ -17,8 +17,8 @@ set_n_least_used_CUDA_VISIBLE_DEVICES 4
|
|||
|
||||
PROJECT_NAME="sft"
|
||||
PARENT_CONFIG_FILE="./benchmark_config" # Path to a folder to save training config logs
|
||||
PRETRAINED_MODEL_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local tokenizer path
|
||||
PRETRAINED_MODEL_PATH="" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="" # huggingface or local tokenizer path
|
||||
BENCHMARK_DATA_DIR="./temp/sft" # Path to benchmark data
|
||||
DATASET_SIZE=640
|
||||
|
||||
|
|
|
@ -17,8 +17,8 @@ set_n_least_used_CUDA_VISIBLE_DEVICES 4
|
|||
|
||||
PROJECT_NAME="simpo"
|
||||
PARENT_CONFIG_FILE="./benchmark_config" # Path to a folder to save training config logs
|
||||
PRETRAINED_MODEL_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="/root/commonData/Llama-2-7b-hf" # huggingface or local tokenizer path
|
||||
PRETRAINED_MODEL_PATH="" # huggingface or local model path
|
||||
PRETRAINED_TOKENIZER_PATH="" # huggingface or local tokenizer path
|
||||
BENCHMARK_DATA_DIR="./temp/simpo" # Path to benchmark data
|
||||
DATASET_SIZE=640
|
||||
|
||||
|
|
|
@ -5,9 +5,9 @@ rm -rf $SAVE_DIR/jsonl
|
|||
rm -rf $SAVE_DIR/arrow
|
||||
|
||||
python prepare_dataset.py --type sft \
|
||||
--data_input_dirs /home/nvme-share/home/yeanbang/data/dataset/hh_rlhf/sft \
|
||||
--conversation_template_config /home/nvme-share/home/yeanbang/ColossalAI/applications/ColossalChat/config/conversation_template/llama2.json \
|
||||
--tokenizer_dir "/home/nvme-share/share/models/Sheared-LLaMA-1.3B" \
|
||||
--data_input_dirs /PATH/TO/SFT/DATASET \
|
||||
--conversation_template_config /PATH/TO/CHAT/TEMPLATE/CONFIG.json \
|
||||
--tokenizer_dir "" \
|
||||
--data_cache_dir $SAVE_DIR/cache \
|
||||
--data_jsonl_output_dir $SAVE_DIR/jsonl \
|
||||
--data_arrow_output_dir $SAVE_DIR/arrow \
|
||||
|
|
Loading…
Reference in New Issue