smothiki

kserve % make deploy-dev
# Move LLMISVC CRD to llmisvc folder	                   
mv config/crd/full/serving.kserve.io_llminferenceservices.yaml config/crd/full/llmisvc/serving.kserve.io_llminferenceservices.yaml
mv config/crd/full/serving.kserve.io_llminferenceserviceconfigs.yaml config/crd/full/llmisvc/serving.kserve.io_llminferenceserviceconfigs.yaml
# Move LocalModel CRD to localmodel folder
mv config/crd/full/serving.kserve.io_localmodelcaches.yaml config/crd/full/localmodel/serving.kserve.io_localmodelcaches.yaml
mv config/crd/full/serving.kserve.io_localmodelnodegroups.yaml config/crd/full/localmodel/serving.kserve.io_localmodelnodegroups.yaml
mv config/crd/full/serving.kserve.io_localmodelnodes.yaml config/crd/full/localmodel/serving.kserve.io_localmodelnodes.yaml
# Copy the cluster role to the helm chart
cp config/rbac/auth_proxy_role.yaml charts/kserve-resources/templates/clusterrole.yaml

smothiki

service/kserve-webhook-server-service serverside-applied
certificate.cert-manager.io/serving-cert serverside-applied
Apply failed with 2 conflicts: conflicts with "helm":
- .metadata.labels.app.kubernetes.io/managed-by
- .metadata.labels.app.kubernetes.io/name
Please review the fields above--they currently have other managers. Here
are the ways you can resolve this warning:
* If you intend to manage all of these fields, please re-run the apply
  command with the `--force-conflicts` flag.
* If you do not intend to manage all of the fields, please edit your

smothiki

#!/usr/bin/env python3
"""
Script to download model metadata files from S3 and create structured JSON output.
Uses AWS CLI to download modelmetadata.json and metadata.yaml files from a single S3 path.

Usage examples:
  # Basic usage for NGC
  python download_model_metadata.py \
    --repo-type NGC \
    --s3-path "s3://my-bucket/models/model1/" \

smothiki

#!/usr/bin/env python3

import argparse
import json
import os
import subprocess
import sys
import logging

import yaml

smothiki

1. Pull and Run the Squid Container Run the following command in your terminal. This uses the popular ubuntu/squid image.

docker run -d --name squid-local -e 'TZ=UTC' -p 3128:3128 ubuntu/squid:latest

-d: Runs in detached mode (background).

-p 3128:3128: Maps port 3128 on your machine to port 3128 in the container (Squid's default port).

smothiki

  - name: Llama 3.1 Instruct
    displayName: Llama 3.1 Instruct
    modelHubID: llama-3.1-instruct
    category: Text Generation
    type: HF
    description: The Llama 3.1 8B-Instruct, 8B instruct and 8B base NIM simplifies the deployment of the Llama 3.1 70B-Instruct, 8B instruct and 8B base tuned models which is optimized for language understanding, reasoning, and text generation use cases, and outperforms many of the available open source chat models on common industry benchmarks.
    requireLicense: True
    licenseAgreements:
    - label: Use Policy
      url: https://llama.meta.com/llama3/use-policy/

smothiki

models:
  - name: Llama-2-Chat
    displayName: Llama-2-Chat (Before 24.02)
    description: Llama-2-Chat model published by Meta
    versions:
      - latestVersionId: LLAMA-2-70B-CHAT
        sha: ff6d1c3ba37b31d4af421951c2300f2256fb3691
        latestVersionSizeInBytes: 107224858679
        createdDate: 2023-10-24T18:21:04.610Z
        updatedDate: 2024-04-23T22:57:49.590Z

smothiki

import logging
import logging as log

from pathlib import Path
from pyspark.sql import DataFrame
from pyspark.sql.types import StructType
from pyspark.sql.functions import udf
from pyspark.sql.types import StructField
from pyspark.sql.functions import lit

smothiki

train_dir = '/home/cdsw/datasettrees'
import numpy as np
import glob
import os
import matplotlib.pyplot as plt
from PIL import Image
train_list = glob.glob(train_dir+'/**/*.jpg',recursive=True)

print(train_list)

smothiki

!pip3 install torch torchvision pandas numpy onnx onnxruntime

import argparse

import torch
import torch.nn.functional as F
from torch import nn, optim
from torch.optim.lr_scheduler import StepLR
from torchvision import datasets, transforms