speed

openr/infer.sh

@@ -15,13 +15,13 @@
 # (TODO)
 # Please modify your requirements
 
-#SBATCH -p r8nv-gpu-hw-80g                    # Submit to 'nv-gpu' Partitiion
+#SBATCH -p r8nv-gpu-hw                   # Submit to 'nv-gpu' Partitiion
 #SBATCH -t 1-06:00:00                # Run for a maximum time of 0 days, 12 hours, 00 mins, 00 secs
 #SBATCH --nodes=1                    # Request N nodes
 #SBATCH --gres=gpu:8                 # Request M GPU per node
 #SBATCH --gres-flags=enforce-binding # CPU-GPU Affinity
 #SBATCH --qos=gpu-normal           # Request QOS Type
-#SBATCH --constraint="A100_80G"
+#SBATCH --constraint="L40S"
 
 ###
 ### The system will alloc 8 or 16 cores per gpu by default.
@@ -81,7 +81,7 @@ cd /nfs_global/S/zhangxiaoyun/prm/openr
 export PYTHONPATH=$(pwd)
 
 bash reason/llm_service/create_service_qwen2.5_math_vllm_gold_prm_speed.sh --acc 1.0 --policy_model_name Qwen2.5-Math-1.5B-Instruct
-sleep 30s
+sleep 100s
 bash scripts/eval/beam_search.sh --acc 1.0 --policy_model_name Qwen2.5-Math-1.5B-Instruct
 
 #- End