ZeroOps: Ho fihlela Meralo ea Boipheliso e Ikemetseng ka li-AI Agents, StackGen, le LangChain | B4N1

ZeroOps: Ho fihlela Meralo ea Boipheliso e Ikemetseng ka li-AI Agents, StackGen, le LangChain

7 months ago · b4n1 Read time: 5 min

ZeroOps: Ho fihlela Meralo ea Boipheliso e Ikemetseng ka li-AI Agents, StackGen, le LangChain

Selelekela

Sebakeng se lulang se fetoha sa tikoloho ea maru, tlhokahalo ea taolo e sebetsang ea litšebeletso tsa motheo le e mpe ha e so ka e phahama. Lisebelisoa tsa setso tsa othomathike li thusitse haholo ho ntlafatseng lits'ebetso, empa hangata li haelloa ke ho sebetsana le mathata le menahano ea meaho ea sejoale-joale. Mona ke moo ZeroOps e kenang - mokhoa oa phetoho o matlafatsang Agetic AI ho fihlela taolo ea meaho e ikemetseng. Sengoliloeng sena, re tla shebana le lefats'e la ZeroOps, re lekola likarolo tsa lona tsa mantlha, melemo, le lits'ebetso tse sebetsang.

ZeroOps ke eng?

ZeroOps ke paradigm e hlahang e kopanyang matla a Artificial Intelligence (AI) le ho feto-fetoha ha maemo a sebaka sa maru ho theha lisebelisoa tsa ho ipusa. Ka ho sebelisa bokhoni ba liakhente tsa AI, StackGen, le LangChain, ZeroOps e thusa mekhatlo ho iketsetsa tsamaiso ea litšebeletso tsa motheo, ho fokotsa liphoso tsa batho, le ho ntlafatsa katleho ka kakaretso.

Agent AI: Pelo ea ZeroOps

Agetic AI e bua ka mofuta oa AI o bonts'ang boits'oaro bo bohlale, bokhoni ba ho etsa liqeto, le ho ikamahanya le maemo. Boemong ba ZeroOps, liakhente tsa AI tsa AI li etselitsoe ho ithuta ho tsoa lipateroneng tsa lits'ebetso, ho supa mefokolo, le ho nka mehato ea ho lokisa. Baemeli bana ba ka koetlisoa mehloling e fapaneng ea data, ho kenyeletsoa li-log, metrics, le lifaele tsa tlhophiso, ho hlaolela kutloisiso e tebileng ea boitšoaro ba litšebeletso.

StackGen: The Infrastructure-as-Code (IaC) Engine

StackGen ke enjene ea IaC e bulehileng e thusang mekhatlo ho hlalosa le ho laola lits'ebetso tsa motheo joalo ka khoutu. Ka ho sebelisa StackGen, bahlahisi ba ka etsa litempele tsa litšebeletso tsa motheo, ba tsamaisa lits'ebetso, le ho laola lisebelisoa ka mokhoa o tsitsitseng le o ka hlahisoang hape. Khokahano ea StackGen le li-agents tsa AI e lumella hore ho thehoe lisebelisoa tsa ho ipholisa, moo baemeli ba ka lemohang le ho lokisa litaba ka nako ea sebele.

LangChain: The AI-Powered Infrastructure Orchestrator

LangChain ke moralo o thehiloeng ho Python o nolofalletsang batho ba ntlafatsang ho haha ​​le ho tsamaisa li-orchestrator tsa motheo tse tsamaisoang ke AI. Ka ho sebelisa LangChain, mekhatlo e ka theha mahlahana a tloaelo a AI a kopantsoeng le StackGen le lisebelisoa tse ling tsa meralo ho fihlela taolo ea meaho e ikemetseng. Mehaho e tloaelehileng ea LangChain le laebrari e pharaletseng ea likarolo tse hahiloeng esale pele li etsa hore e be khetho e nepahetseng bakeng sa ho haha ​​​​litharollo tse rarahaneng tsa mekhoa ea motheo ea AI.

Tšebeliso e sebetsang: Mohlala o bonolo

Ha re nahaneng ka mohlala o bonolo oa hore na ZeroOps e ka sebelisoa joang tikolohong ea maru. Ha re re re na le ts'ebeliso ea webo e kentsoeng sehlopheng sa Kubernetes, 'me re batla ho iketsetsa ts'ebetso ea ho lekanya kopo ho latela tlhoko ea sephethephethe.


import langchain
ho tsoa ho langchain.agents import LLMChain
ho tsoa ho langchain.llms import LangChainLLM

# Hlalosa moemeli oa AI
moemeli = LLMChain(
    llm=LangChainLLM(
        model_name = "code-davinci-002",
        mocheso=0.7,
        max_tokens=2048,
        stop_sequences=["\n", "\n\n"]
    ),
    chain_name="scale_app"
)

# Hlalosa template ea StackGen
mohlala = {
    "kind": "Deployment",
    "metadata": {
        "name": "web-app"
    },
    "spec": {
        "Replicas": 1
    }
}

# Hlalosa leano la sekhahla
leano = {
    "condition": "traffic > 1000",
    "ketso": {
        "Replicas": 2
    }
}

# Theha 'mino oa motheo o tsamaisoang ke AI
sebini = LangChainLLM(
    model_name = "code-davinci-002",
    mocheso=0.7,
    max_tokens=2048,
    stop_sequences=["\n", "\n\n"]
)

# Hlalosa phepelo ea StackGen
deployment = StackGen(
    template=template,
    leano=leano
)

# Kenya ts'ebeliso
deployment.deploy()

# Eketsa ts'ebeliso ho latela tlhoko ea sephethephethe
orchestrator.scale_app(moemeli, thomello)

Mohlala ona, re hlalositse moemeli oa AI ea sebelisang LangChain e kopanyang le StackGen ho iketsetsa mokhoa oa ho eketsa ts'ebeliso ea marang-rang ho latela tlhoko ea sephethephethe. Moemeli o sebelisa mokhoa oa puo o koetlisitsoeng esale pele ho sekaseka lintlha tsa sephethephethe le ho etsa liqeto mabapi le ho eketsa ts'ebeliso.

Kaholimo ho 5-10 Litaelo/Lipotlana tse Sebelisitsoeng ka ho Fetisisa

Mona ke tse ling tsa litaelo le likonopo tse sebelisoang haholo bakeng sa ZeroOps:

  1. langchain init: E qala morero o mocha oa LangChain.
langchain init my_project
  1. llm chain: E theha ketane e ncha ea LLM.
ho tsoa ho langchain.agents import LLMChain
ho tsoa ho langchain.llms import LangChainLLM

moemeli = LLMChain(
    llm=LangChainLLM(
        model_name = "code-davinci-002",
        mocheso=0.7,
        max_tokens=2048,
        stop_sequences=["\n", "\n\n"]
    ),
    chain_name="my_chain"
)
  1. stackgen deploy: E romela thempleite ea StackGen ho mofani oa maru.
stackgen deploy my_template.yaml
  1. orchestrator scale_app: E lekanya sesebelisoa ho latela tlhoko ea sephethephethe.
ho tsoa ho langchain.agents import LLMChain
ho tsoa ho langchain.llms import LangChainLLM

sebini = LangChainLLM(
    model_name = "code-davinci-002",
    mocheso=0.7,
    max_tokens=2048,
    stop_sequences=["\n", "\n\n"]
)

orchestrator.scale_app(moemeli, thomello)
  1. Terene ea moemeli: E koetlisa moemeli oa AI ka setlankaneng sa data.
moemeli oa terene my_dataset.csv
  1. Leano la stackgen: E hlalosa leano la ho eketsa kopo.
leano = {
    "condition": "traffic > 1000",
    "ketso": {
        "Replicas": 2
    }
}
  1. langchain visualize: E bona ka mahlo a kelello tshebetso ya ho etsa diqeto ya moemedi wa AI.
langchain bona my_agent.json
  1. maemo a orchestrator: E bontsha boemo ba tshebediso.
boemo ba sebini my_app
  1. stackgen logs: E bonts'a li-log tsa ho romelloa ha StackGen.
stackgen logs my_deployment
  1. ente e lekola: E lekola tšebetso ea moemeli oa AI.
moemeli hlahloba my_agent.json

Pono ya pono

Mona ke pono ea meralo ea ZeroOps e sebelisang Chart.js:

{
  "type": "bar",
  "data": {
    "labels": ["Agent AI", "StackGen", "LangChain"],
    "lintlha tsa boitsebiso": [
      {
        "label": "Likaroloana",
        "data": [10, 8, 12],
        "backgroundColour": [
          "rgba(255, 99, 132, 0.2)",
          "rgba(54, 162, 235, 0.2)",
          "rgba(255, 206, 86, 0.2)"
        ],
        "Mmala oa moeli": [
          "rgba(255, 99, 132, 1)",
          "rgba(54, 162, 235, 1)",
          "rgba(255, 206, 86, 1)"
        ],
        "borderWidth": 1
      }
    ]
  },
  "likgetho": {
    "makala": {
      "yAxes": [
        {
          "li-ticks": {
            "beginAtZero": 'nete
          }
        }
      ]
    }
  }
}

Pono ena e bonts'a likarolo tse tharo tsa mantlha tsa ZeroOps: Agent AI, StackGen, le LangChain. Chate ea bara e bonts'a bohlokoa bo lekanyelitsoeng ba karolo e 'ngoe le e 'ngoe mohahong oa ZeroOps.

Qetello

ZeroOps ke mokhoa oa phetoho oa taolo ea meralo ea ts'ebetso e matlafatsang Agetic AI, StackGen, le LangChain ho fihlela taolo ea meaho e ikemetseng. Ka ho sebelisa matla a AI le IaC, ZeroOps e thusa mekhatlo ho iketsetsa tsamaiso ea litšebeletso tsa motheo, ho fokotsa liphoso tsa batho, le ho ntlafatsa katleho ka kakaretso. Ka lits'ebetso tsa eona tse sebetsang, litaelo tse holimo, le pono ea pono, ZeroOps e mothating oa ho fetola tsela eo re tsamaisang lits'oants'o tsa motheo maemong a maru.

Have Questions?

Let's Talk

Contact me on LinkedIn