Merge pull request #71 from duggalsu/fix_vid_vec_graviton_dockerfile

Fix Graviton commands

src/api/Dockerfile.vid_vec_rep_resnet

@@ -17,23 +17,24 @@ RUN pip install --no-cache-dir --user -r /app/core/operators/vid_vec_rep_resnet_
 
 ### AWS Graviton Optimization ###
 # Graviton3(E) (e.g. c7g, c7gn and Hpc7g instances) supports BF16 format for ML acceleration. This can be enabled in oneDNN by setting the below environment variable
-grep -q bf16 /proc/cpuinfo && export DNNL_DEFAULT_FPMATH_MODE=BF16
+ENV DNNL_DEFAULT_FPMATH_MODE=BF16
 
 # Enable primitive caching to avoid the redundant primitive allocation
 # latency overhead. Please note this caching feature increases the
 # memory footprint. Tune this cache capacity to a lower value to
 # reduce the additional memory requirement.
-export LRU_CACHE_CAPACITY=1024
+ENV LRU_CACHE_CAPACITY=1024
 
 # Enable Transparent huge page allocations from PyTorch C10 allocator
-export THP_MEM_ALLOC_ENABLE=1
+ENV THP_MEM_ALLOC_ENABLE=1
 
 # Make sure the openmp threads are distributed across all the processes for multi process applications to avoid over subscription for the vcpus. For example if there is a single application process, then num_processes should be set to '1' so that all the vcpus are assigned to it with one-to-one mapping to omp threads
-num_vcpus=$(getconf _NPROCESSORS_ONLN)
-num_processes=1
-export OMP_NUM_THREADS=$((1 > ($num_vcpus/$num_processes) ? 1 : ($num_vcpus/$num_processes)))
-export OMP_PROC_BIND=false
-export OMP_PLACES=cores
+# RUN num_vcpus=8
+# RUN num_processes=1
+# RUN temp = $((1 > ($num_vcpus/$num_processes) ? 1 : ($num_vcpus/$num_processes)))
+ENV OMP_NUM_THREADS=1
+ENV OMP_PROC_BIND=false
+ENV OMP_PLACES=cores
 ###
 
 COPY ./core/operators/vid_vec_rep_resnet.py /app/core/operators/vid_vec_rep_resnet.py