Add files using upload-large-folder tool

english_only/tok_english_only/tok_jobs_check/tokenizer_14.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-only-14
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-only-14
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-only-14-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/datasets/english_only//tok_english_only/tok_files_check/tokenizer_14.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

english_only/tok_english_only/tok_jobs_check/tokenizer_15.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-only-15
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-only-15
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-only-15-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/datasets/english_only//tok_english_only/tok_files_check/tokenizer_15.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

english_only/tok_english_only/tok_jobs_check/tokenizer_22.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-only-22
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-only-22
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-only-22-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/datasets/english_only//tok_english_only/tok_files_check/tokenizer_22.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

english_only/tok_english_only/tok_jobs_check/tokenizer_23.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-only-23
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-only-23
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-only-23-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/datasets/english_only//tok_english_only/tok_files_check/tokenizer_23.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

english_only/tok_english_only/tok_jobs_check/tokenizer_34.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-only-34
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-only-34
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-only-34-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/datasets/english_only//tok_english_only/tok_files_check/tokenizer_34.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

english_only/tok_english_only/tok_jobs_check/tokenizer_39.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-only-39
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-only-39
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-only-39-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/datasets/english_only//tok_english_only/tok_files_check/tokenizer_39.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock

english_only/tok_english_only/tok_jobs_check/tokenizer_5.yaml

@@ -0,0 +1,104 @@
+apiVersion: kubeflow.org/v2beta1
+kind: MPIJob
+metadata:
+  name: peacock-tokenizer-english-only-5
+  namespace: peacock
+  labels:
+    app: peacock-tokenizer-english-only-5
+spec:
+  slotsPerWorker: 8
+  runPolicy:
+    backoffLimit: 1
+    cleanPodPolicy: Running
+  mpiReplicaSpecs:
+    Launcher:
+      replicas: 1
+      template:
+        spec:
+          hostIPC: true
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-tokenizer-english-only-5-launcher
+              imagePullPolicy: IfNotPresent
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock
+              command: ["/bin/bash", "-c"]
+              args:
+                - >-
+                  
+                  export SYNAPSE_VERSION="1.15.1";
+                  export WORKER_DIR="/mnt/weka/peacock/experiments/llama";
+
+                  export MEGATRON_SETUP_CMD="$WORKER_DIR/launch/setup.sh"
+                  export TOKENIZER_CMD=/mnt/weka/peacock/idc/datasets/english_only//tok_english_only/tok_files_check/tokenizer_5.sh
+                  HOSTSFILE=$OMPI_MCA_orte_default_hostfile;
+                  echo "HOSTSFILE=$HOSTSFILE";
+                  MASTER_ADDR="$(head -n 1 $HOSTSFILE | sed -n s/[[:space:]]slots.*//p)";
+                  NUM_NODES=$(wc -l < $HOSTSFILE);
+                  CARDS_PER_NODE=8;
+                  N_CARDS=$((NUM_NODES*CARDS_PER_NODE));
+                  echo "MPI_ROOT=$MPI_ROOT";
+                  echo "N_CARDS=$N_CARDS";
+                  echo "MASTER_ADDR=$MASTER_ADDR";
+                  sleep 20;
+
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $MEGATRON_SETUP_CMD;
+
+                  mpirun -np $N_CARDS  -npernode 8 \
+                    --tag-output \
+                    --allow-run-as-root \
+                    --prefix $MPI_ROOT \
+                    -x WORKER_DIR=$WORKER_DIR \
+                    -x SYNAPSE_VERSION=$SYNAPSE_VERSION $TOKENIZER_CMD;
+                    
+                   
+    Worker:
+      replicas: 1
+      template:
+        spec:
+          volumes:
+            - name: work-dir
+              persistentVolumeClaim:
+                claimName: peacock-fs-pvc
+          tolerations:
+            - key: "habana.ai/gaudi"
+              operator: "Exists"
+              effect: "NoSchedule"
+            - key: "k8s/namespace"
+              operator: "Equal"
+              value: "peacock"
+              effect: "NoSchedule"
+          hostIPC: true
+          containers:
+            - image: vault.habana.ai/gaudi-docker/1.15.1/ubuntu22.04/habanalabs/pytorch-installer-2.2.0:latest
+              name: peacock-llama-worker
+              imagePullPolicy: IfNotPresent
+              securityContext:
+                capabilities:
+                  add:
+                    - SYSLOG
+              resources:
+                limits:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+                requests:
+                  habana.ai/gaudi: 8
+                  hugepages-2Mi: 300Gi
+                  memory: 700Gi
+                  cpu: 150
+              volumeMounts:
+                - name: work-dir
+                  mountPath: /mnt/weka/peacock