chatglm-6b-int4 / ptuning /evaluate.sh
NewBreaker's picture
first
23aa310
raw
history blame
760 Bytes
PRE_SEQ_LEN=128
CHECKPOINT=adgen-chatglm-6b-pt-128-2e-2
STEP=1000
CUDA_VISIBLE_DEVICES=0 python main.py \
--do_predict \
--train_file .\\datasets\\AdvertiseGen\\train.json \
--validation_file .\\datasets\\AdvertiseGen\\dev.json \
--test_file .\\datasets\\AdvertiseGen\\dev.json \
--overwrite_cache \
--prompt_column content \
--response_column summary \
--model_name_or_path ..\\models\\chatglm-6b-int4 \
--ptuning_checkpoint .\\output\\$CHECKPOINT\\checkpoint-$STEP \
--output_dir .\\output\\$CHECKPOINT \
--overwrite_output_dir \
--max_source_length 64 \
--max_target_length 64 \
--per_device_eval_batch_size 1 \
--predict_with_generate \
--pre_seq_len $PRE_SEQ_LEN \
--quantization_bit 4