ggml-mllm / config.toml
matthoffner's picture
Update config.toml
bc1988d
raw
history blame
128 Bytes
model="wizardLM-7B-GGML/wizardLM-7B.ggml.q5_1.bin"
use_mmap="True"
use_mlock="True"
n_threads="4"
n_ctx="1024"
n_gpu_layers="20"