repo_id:
  - 'meta-llama/Llama-2-7b-chat-hf'
  - 'THUDM/chatglm2-6b'
  - 'THUDM/chatglm3-6b'
  - 'baichuan-inc/Baichuan2-7B-Chat'
  - 'Qwen/Qwen-7B-Chat'
local_model_hub: '/mnt/disk1/models'
warm_up: 1
num_trials: 3
num_beams: 1 # default to greedy search
low_bit: 'fp8' # default to use 'sym_int4' (i.e. symmetric int4)
batch_size: 1 # default to 1
in_out_pairs:
  - '32-32'
  - '512-256'
  - '1024-128'
  - '2048-256'
test_api:
  - "transformer_int4_gpu"  # on Intel GPU
cpu_embedding: False # whether put embedding to CPU (only avaiable now for gpu win related test_api)
exclude:
  - 'meta-llama/Llama-2-7b-chat-hf:2048:4'
  - 'meta-llama/Llama-2-7b-chat-hf:512:8'
  - 'meta-llama/Llama-2-7b-chat-hf:1024:8'
  - 'meta-llama/Llama-2-7b-chat-hf:2048:8'
  - 'THUDM/chatglm2-6b:2048:8'
  - 'THUDM/chatglm3-6b:2048:8'
  - 'baichuan-inc/Baichuan2-7B-Chat:2048:2'
  - 'baichuan-inc/Baichuan2-7B-Chat:1024:4'
  - 'baichuan-inc/Baichuan2-7B-Chat:2048:4'
  - 'baichuan-inc/Baichuan2-7B-Chat:512:8'
  - 'baichuan-inc/Baichuan2-7B-Chat:1024:8'
  - 'baichuan-inc/Baichuan2-7B-Chat:2048:8'
  - 'Qwen/Qwen-7B-Chat:2048:1'
  - 'Qwen/Qwen-7B-Chat:1024:2'
  - 'Qwen/Qwen-7B-Chat:2048:2'
  - 'Qwen/Qwen-7B-Chat:512:4'
  - 'Qwen/Qwen-7B-Chat:1024:4'
  - 'Qwen/Qwen-7B-Chat:2048:4'
  - 'Qwen/Qwen-7B-Chat:512:8'
  - 'Qwen/Qwen-7B-Chat:1024:8'
  - 'Qwen/Qwen-7B-Chat:2048:8'