Whakataki
Ko te mara o DevOps kua puta nga huringa nui i roto i nga tau tata nei, me te whakaurunga o Artificial Intelligence (AI) me te Ako Miihini (ML) kua piki haere te rongonui. Ko tetahi o nga waahanga nui kei te hangai e AI he paanga nui kei roto i te ao o te automation. Ma te whakamahi i nga kaihoko AI, ka taea e nga whakahaere te whakamaarama i a raatau rerenga mahi whanaketanga rorohiko, te whakaheke i nga hapa a-ringa, me te whakapai ake i te pai.
I roto i tenei tuhinga, ka rukuhia e matou te ao o te miihini DevOps i peia e AI, e arotahi ana ki nga angamahi LangChain me DSPy. Ka tirotirohia me pehea te whakamahi i enei taputapu ki te hanga i nga punaha whakatau take me te whakaaunoa i nga rerenga mahi whanaketanga rorohiko. I tua atu, ka whakaratohia e matou he tauira whaitake mo te whakauru i nga Tauira Reo Nui (LLM) ki roto i to paipa DevOps.
He aha te LangChain me te DSPy?
I mua i te ruku ki roto i nga korero mo te miihini DevOps i peia e AI, me ata titiro ki nga anga LangChain me DSPy.
LangChain
Ko te LangChain he anga-puna tuwhera e taea ai e nga kaihanga te hanga i nga tono korero AI ma te whakamahi i nga Tauira Reo Nui (LLM). Ka whakaratohia he API ngawari me te mohio mo te taunekeneke me nga LLM, ka taea e nga kaiwhakawhanake te hanga atanga korero ritenga mo o raatau tono.
Kei te tautoko a LangChain i te whānuitanga o nga LLM, tae atu ki nga tauira rongonui penei i a LLaMA, BERT, me RoBERTa. Ka whakarato hoki i te whānuitanga o nga taputapu mo te whakatikatika me te whakarite i nga LLM kia rite ki nga keehi whakamahi motuhake.
DSPy
Ko DSPy he whare pukapuka Python e whakarato ana i te huarahi ngawari me te pai ki te whakauru i nga tauira AI ki roto i to paipa DevOps. Kei te tautoko i te whānuitanga o nga anga AI, tae atu ki a LangChain, me te whakarato i te whānuitanga o nga taputapu mo te whakaaunoa i nga rerenga mahi a AI.
I hoahoatia a DSPy ki te mahi pai me nga taputapu DevOps rongonui penei i a Jenkins, GitLab, me Docker, he mea tino pai mo nga whakahaere e whai ana ki te whakauru AI ki o raatau hanganga DevOps.
Whakaritea nga Kaitohu AI mo te DevOps Automation
Inaianei kei a maatau he tirohanga poto mo LangChain me DSPy, me tirotiro me pehea te whakamahi i enei anga ki te whakarite i nga kaihoko AI mo te miihini DevOps.
Whakataunga Putanga Whakaoho
Ko tetahi o nga painga nui o te miihini DevOps i peia e AI ko te kaha ki te whakatau i nga take i mua i te raru nui. Ka taea te whakamahi i te LangChain me te DSPy ki te hanga i nga atanga korero ritenga e mahi tahi ana me nga LLM ki te tautuhi i nga take pea me te tuku taunakitanga mo te whakatau.
Anei he tauira ngawari mo te mahi:
kawemai langchain
mai i langchain.chat_models kawemai LLaMA
# Waihangahia he tauira korerorero LangChain
korerorero_tauira = LLaMA()
# Tautuhia he mahi hei taunekeneke me te tauira korerorero
def interact_with_chat_model(tohutohu):
whakautu = chat_model.run(prompt)
whakahoki whakahoki
# Tautuhia he mahi hei tautuhi i nga take pea
def identify_issues(waehere):
prompt = f"Tautuhia nga take pea kei te waehere e whai ake nei: {code}"
whakautu = interact_with_chat_model(prompt)
whakahoki whakahoki
# Tautuhia he mahi hei whakarato i nga taunakitanga mo te whakatau
def provide_recommendations(take):
prompt = f"Homai kupu tohutohu mo te whakatau i nga take e whai ake nei: {take}"
whakautu = interact_with_chat_model(prompt)
whakahoki whakahoki
# Tauira whakamahi
waehere = """
def tāpiri(x, y):
hoki x + y
"""
take = tautohu_take(waehere)
tūtohu = whakarato_tohutohu(take)
tā(tohutohu)
I roto i tenei tauira, ka tautuhia e matou he mahi ngawari interact_with_chat_model e mahi tahi ana me te tauira korerorero a LangChain ki te tuku whakautu ki te tono tere. Ka tautuhia etahi atu mahi e rua tautuhi_take me te provide_recommendations e whakamahi ana i te tauira korerorero ki te tautuhi i nga take pea kei roto i etahi tauira tauira me te tuku taunakitanga hei whakatau.
Whakaaunoa i nga Reremahi Whanaketanga Pūmanawa
Ka taea hoki te whakamahi i te LangChain me te DSPy ki te whakaaunoa i nga rerenga mahi whanaketanga rorohiko ma te whakauru i nga tauira AI ki roto i to paipa DevOps.
Anei tetahi tauira o te mahi penei:
kawemai dspy
mai i dspy kawemai Rerengamahi
# Tautuhia he rerengamahi DSPy
reremahi = Rerengamahi()
# Tautuhia he mahi hei whakahaere i tetahi taputapu tātari waehere
def run_code_analysis(waehere):
# Whakahaerehia he taputapu tātari waehere hei tautuhi i nga take pea
take = []
#...
take whakahoki
# Tautuhia he mahi hei whakarato i nga taunakitanga mo te whakatau
def provide_recommendations(take):
# Whakamahia te tauira korerorero LangChain ki te tuku tohutohu
prompt = f"Homai kupu tohutohu mo te whakatau i nga take e whai ake nei: {take}"
whakautu = langchain.chat_model.run(prompt)
whakahoki whakahoki
# Tautuhia he mahi hei tuku i te waehere
def deploy_code(waehere):
# Tukuna te waehere ki tetahi taiao whakaputa
#...
# Tautuhia te rerenga mahi
rerengamahi.add_task(run_code_analysis)
rerengamahi.add_task(provide_recommendations)
rerengamahi.add_task(deploy_code)
# Whakahaerehia te rerenga mahi
rerengamahi.run()
I roto i tenei tauira, ka tautuhia e matou he rerengamahi DSPy ngawari e toru nga mahi: te whakahaere i tetahi taputapu tātari waehere, whakarato tohutohu mo te whakatau, me te tuku i te waehere. Ka tautuhia te rerengamahi ma te taapiri i ia mahi me te whakahaere.
Top 5-10 Whakamahinga Whakahau/Snippets
Anei etahi o nga whakahau e whakamahia nuitia ana mo LangChain me DSPy:
LangChain
langchain.chat_models.LLaMA(): Waihangahia he tauira korerorero LangChain ma te whakamahi i te tauira LLaMA.langchain.chat_model.run(prompt): Whakahaerehia te tauira korerorero me te tuku wawe.langchain.chat_model.interact_with(prompt): Whakawhitiwhiti me te tauira korerorero ma te whakamahi i te tere.
DSPy
dspy.Workflow(): Waihangahia he rerengamahi DSPy.workflow.add_task(task): Tāpirihia he mahi ki te rerengamahi.workflow.run(): Whakahaerehia te rerengamahi.
Whakaaturanga
Anei he tirohanga ngawari o te rerengamahi ma te whakamahi i te Chart.js:
{
"type": "pae",
"raraunga": {
"labels": ["Tatari Waehere", "Tautohu", "Whakamahinga"],
"raunga raraunga": [
{
"label": "Mahinga Mahi",
"raraunga": [10, 20, 30],
"Tae papamuri": [
"rgba(255, 99, 132, 0.2)",
"rgba(54, 162, 235, 0.2)",
"rgba(255, 206, 86, 0.2)"
],
"Taitapa Tae": [
"rgba(255, 99, 132, 1)",
"rgba(54, 162, 235, 1)",
"rgba(255, 206, 86, 1)"
],
"BorderWidth": 1
}
]
},
"kōwhiringa": {
"taitara": {
"whakaatu": pono,
"text": "Mahinga Mahi"
},
"tauine": {
"yAxes": [
{
"tika": {
"beginAtZero": pono
}
}
]
}
}
}
Ko tenei tirohanga e whakaatu ana i te tūtohi pae ngawari e tohu ana i te rerengamahi, e toru nga pae e tohu ana mo te tātari waehere, nga taunakitanga, me nga mahi tuku.
Whakamutunga
I roto i tenei tuhinga, i tirotirohia e matou te ao o te AI-driven DevOps automation ma te whakamahi LangChain me DSPy. I kite matou me pehea te whakamahi i enei anga ki te hanga i nga punaha whakatau take me te whakaaunoa i nga rerenga mahi whanaketanga rorohiko. I tukuna ano e matou etahi tauira whaihua mo te whakauru i nga LLM ki to paipa DevOps me te tiro i te rerengamahi ma te whakamahi i te Chart.js.
Ma te whakamahi i te mana o AI me te aunoatanga, ka taea e nga whakahaere te whakamaarama i o raatau rerenga mahi whanaketanga rorohiko, te whakaiti i nga hapa a-ringa, me te whakapai ake i te pai katoa. Ahakoa he miihini DevOps koe, he kaiwhakawhanake rorohiko, he kaiarahi pakihi ranei, he taputapu kaha a LangChain me DSPy hei awhina i a koe ki te whakatutuki i o whainga.