vllm/vllm_v0.10.0/.buildkite/scripts/tpu/config_v6e_1.env

15 lines
290 B
Bash

# Environment config
TEST_NAME=llama8b
CONTAINER_NAME=vllm-tpu
# vllm config
MODEL=meta-llama/Llama-3.1-8B-Instruct
MAX_NUM_SEQS=256
MAX_NUM_BATCHED_TOKENS=1024
TENSOR_PARALLEL_SIZE=1
MAX_MODEL_LEN=2048
DOWNLOAD_DIR=/mnt/disks/persist
EXPECTED_THROUGHPUT=8.0
INPUT_LEN=1800
OUTPUT_LEN=128