Files
r200_8f_xtrt_llm/examples/gptj/run.sh
2025-08-06 15:49:14 +08:00

9 lines
376 B
Bash

XMLIR_D_XPU_L3_SIZE=0 \
python3 run.py \
--engine_dir=./downloads/gptj-6b/trt_engines/fp16/1-XPU/ \
--hf_model_location=./downloads/gptj-6b \
--max_output_len=2048 \
--performance_test_scale=1x512x512E1x1024x1024E1x2000x64E1x2048x2048E2x512x512E2x1024x1024E2x2000x64E2x2048x2048E4x512x512E\
4x1024x1024E4x2000x64E4x2048x2048E8x512x512E8x1024x1024E8x2000x64 \
--log_level=info