how to run in CPU float32 mode?

[iperov]

I am running "Meta-Llama-3.1-8B-Instruct-f32.gguf" size is 32GB.
RAM size is 64GB
After load llama-cli.exe uses 16GB according process explorer.

also in the log

llama_kv_cache_init:        CPU KV buffer size = 16384.00 MiB
llama_new_context_with_model: KV self size  = 16384.00 MiB, K (f16): 8192.00 MiB, V (f16): 8192.00 MiB

so how to run in float32 mode?

[ggerganov]

Add -ctk f32 -ctv f32

[iperov]

Add -ctk f32 -ctv f32

got error

llama_new_context_with_model: V cache quantization requires flash_attn

but -ctk f32 -ctv f16 works