[DeepSeek] Update Docs for Setting Measurement Results Path

[yiliu30]

Depends on intel/neural-compressor#2210

Configure the Measurement Statistics Results

The environment variable INC_MEASUREMENT_DUMP_PATH_PREFIX specifies the root directory where measurement statistics were saved.
The final path is constructed by joining this root directory with the dump_stats_path defined in the quantization JSON file specified by the QUANT_CONFIG environment variable.

Example:

INC_MEASUREMENT_DUMP_PATH_PREFIX=/mnt/disk3/vllm-fork
dump_stats_path (from config): "scripts/nc_workspace_measure_kvache/inc_measure_output"
Resulting full path: "/mnt/disk3/vllm-fork/scripts/nc_workspace_measure_kvache/inc_measure_output_xx"

Caution

Before running the benchmark, update the following variables in the single_16k_len_inc.sh script:

• model_path
• QUANT_CONFIG
• INC_MEASUREMENT_DUMP_PATH_PREFIX

cc @thuang6

[Copilot]

Pull Request Overview

This PR updates the documentation regarding the configuration of measurement statistics results for the quantization workflow. It adds a new section to explain how the full output path is constructed and clarifies the benchmark preparation steps by listing the environment variables that need updating.

• Added a "Configure the Measurement Statistics Results" section with an example.
• Updated the benchmark caution message to include additional environment variables.

[yiliu30]

@taotod @czhu15 @yangulei Please help review, thx

[czhu15]

LGTM