Quantong Qiu
|
d904d663a4
|
Add dsv4_pro6000_vllm_tp_dp_matrix experiment for RTX 6000D
- vLLM adaptive concurrency benchmark adapted from H200 version
- Key changes:
- Model path: /data/hf_models/DeepSeek-V4-Flash
- Venv path: /root/.miniconda3/envs/sglang
- Dataset path: /data/yy/sskj/datasets/
- DSL renamed to OSL (Output Sequence Length)
- Includes: config.env, adaptive_config.env, run_adaptive_concurrency.sh,
start_vllm_dp.sh, start_vllm_docker.sh
|
2026-07-13 02:59:33 +00:00 |
|