ChatQnA - Adding files to deploy an application in the K8S environment using Helm

Signed-off-by: Chingis Yundunov <c.yundunov@datamonsters.com>

Author: chyundunovDatamonsters

ChatQnA/kubernetes/helm/faqgen-rocm-tgi-values.yaml

@@ -21,7 +21,7 @@ tgi:
   MAX_TOTAL_TOKENS: "4096"
   PYTORCH_TUNABLEOP_ENABLED: "0"
   USE_FLASH_ATTENTION: "true"
-  FLASH_ATTENTION_RECOMPUTE: "true"
+  FLASH_ATTENTION_RECOMPUTE: "false"
   HIP_VISIBLE_DEVICES: "0,1"
   MAX_BATCH_SIZE: "2"
   extraCmdArgs: [ "--num-shard","2" ]

ChatQnA/kubernetes/helm/rocm-tgi-values.yaml

@@ -15,7 +15,7 @@ tgi:
   MAX_TOTAL_TOKENS: "4096"
   PYTORCH_TUNABLEOP_ENABLED: "0"
   USE_FLASH_ATTENTION: "true"
-  FLASH_ATTENTION_RECOMPUTE: "true"
+  FLASH_ATTENTION_RECOMPUTE: "false"
   HIP_VISIBLE_DEVICES: "0,1"
   MAX_BATCH_SIZE: "2"
   extraCmdArgs: [ "--num-shard","2" ]