Mutauro AI GUIDE

KV Cache

Iyo KV cache inochengetedza kiyi uye kukosha mavector shanduri yakatoverengera kune yakapfuura tokens, saka haifanirwe kuvadzokorodza kune yega yega izwi idzva raanogadzira.

2 min verengaLast update

Pfupiso

It is the single biggest reason text generation is fast — and the main thing eating your GPU memory during long conversations.

Kudzika Kwakadzika

Transformers inogadzira mavara tokeni imwe panguva, uye imwe neimwe itsva tokeni yekutarisisa inofanirwa kuenzaniswa neyakatangira tokeni. Iyo yekutarisisa meshini inoshandura chiratidzo chega chega kuita mubvunzo, kiyi, uye kukosha vector. Pasina caching, kugadzira nhamba yechiratidzo 1,000 zvinoreva kudzoreredza makiyi uye kukosha kune ese 999 ekutanga tokens padanho rega rega - quadratic, kutambisa basa. Iyo KV cache inochengetedza aya makiyi uye anokoshesa mavheji mushure mekunge atanga kuverengerwa uye kuashandisa zvakare, saka nhanho yega yega nyowani inongoverengera mavheji eiyo imwe nyowani tokeni uye inopinda pamusoro peyakachengetwa cache. Izvi zvinoderedza mutengo we-tokeni kubva pakuyera nehurefu hwekutevedzana kusvika pakungogara. Kutengeserana-yekuyeuka: iyo cache inokura zvakatevedzana nehurefu hwemamiriro, huwandu hwematanho, uye misoro yekutarisisa, kazhinji ichive inotungamira ndangariro mutengi mukureba-mamiriro ekushandira.

Technical Insight

Munguva ye 'prefill' chikamu modhi inobata kukurumidza kwese uye inozadza cache; panguva ye 'decode' inoisa imwe tokeni yeK/V padanho uye inodzokorora. Cache saizi zviyero se 2 (K uye V) × zvidimbu × misoro × musoro_dim × sequence_length × batch, mune yakasarudzwa chaiyo. Kugadzirisa izvi, mamodheru emazuva ano anoshandisa akaungana-mubvunzo kana akawanda-mubvunzo kutarisa kugovera makiyi / hunhu pamisoro yese, uye masevhisi masisitimu akaita sevLLM anoshandisa PagedAttention kugovera cache mumabhuraki asina- contiguous, kucheka kupatsanuka uye marara.

Strategic Impact

Kumhanya uye chiyero

Mutauro workflows inogona kufamba nekukurumidza pasina kupira kuenderana.

Svika uye svika

Inopamhidzira kupinda mumitauro yese nemataera ekutaurirana.

Sarudzo dzakajeka

Zvikwata zvinogona kupedza nguva yakawanda pakutonga uku otomatiki ichibata kudzokorora.

Ramangwana reKV Cache

Sezvo mahwindo emukati anotambanudzira mumazana ezviuru zvezviratidzo, iyo KV cache inova iyo yepakati bhodhoro, saka hutsva hunotyisa: cache quantization kune 8 kana 4 bits, mitemo yekudzinga inodonhedza yakaderera-yakakosha tokens, kuyambuka-chikumbiro prefix kugovera, uye kurodha kuCPU kana diski. Architectural shifts senge multi-head latent attention inomanikidza cache pachayo. Tarisira kuenderera mberi dhizaini yekutarisisa akasiyana uye ndangariro masisitimu ane chinangwa chekushandira akareba mamiriro akachipa uye nepamusoro pekupinda.

Real-World Implementation

Kumhanyisa mhinduro dzechatbot nekushandisa zvakare makiyi akavharirwa / kukosha kubva munhoroondo yehurukuro pane kuidzokororazve.

Prefix caching inogovera cache yenguva refu yekukurumidza kune vakawanda vashandisi, kucheka mutengo uye latency.

vLLM's PagedAttention inobata KV cache mumabhuroko kuti ishumire zvikumbiro zvakawanda panguva imwe chete pane imwe GPU zvakanaka.

Kuverengera iyo KV cache kudzikisa iko kurongeka kuti ikwane kureba mamiriro mune shoma GPU ndangariro.

Njodzi & Guardrails

Chokwadi chehuroyi chinogona kupinda chinyararire mishumo, kuyerera kwetsigiro, kana tsvakiridzo.

Kunzwa nekukasira kunogona kugadzira mhedzisiro isingaenderane pane zvikumbiro zvakafanana.

Sensitive text data inogona kuburitswa kana zvidhiraivho zvisina kusimba.

Implementation Roadmap

1

Tsanangura chimiro chekubuda, toni, uye mhando zviyero usati waburitsa.

2

Mhinduro dzepasi neakavimbika masosi pese pazvine basa.

3

Chengetedza ongororo yekuongorora yemunhu kune yakakwira-stake zvinobuda.

4

Tevera maitiro ekutadza uye dzidzisazve kukurudzira kana mafambiro ebasa nguva nenguva.

Ramba Uchiongorora

Free newsletter

Get the daily AI briefing

Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.

One email each weekday. Unsubscribe in one click. We never sell or share your address.

Test yourself

Take the KV Cache quiz

Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.

Tanga mibvunzo

Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation

Gaidhi rinotevera

KV Cache Optimization

Mibvunzo inowanzo bvunzwa

What is KV Cache?

Iyo KV cache inochengetedza kiyi uye kukosha mavector shanduri yakatoverengera kune yakapfuura tokens, saka haifanirwe kuvadzokorodza kune yega yega izwi idzva raanogadzira. Ndicho chikonzero chimwe chete chikuru chikonzero chizvarwa chemavara chinokurumidza - uye chinhu chikuru chiri kudya ndangariro yako yeGPU panguva yehurukuro refu.

Ko KV cache inochengeterei?

Iyo KV cache inobata kiyi uye kukoshesa mavheji kubva kune ekutanga tokens saka kutarisisa kunogona kuvashandisa zvakare pane kuverengera.

Nderipi dambudziko iro KV cache inonyanya kugadzirisa?

Pasina caching, chiratidzo chitsva chega chega chaizoda kudzokorodza K/V pachiratidzo chega chega, izvo zvinopambadza; iyo cache inoita kuti per-token mutengo ugare uripo.

Chii chakanyanya kuderera cheKV cache?

Iyo cache inokura nekutevedzana kureba, maseru, uye misoro, kazhinji ichiva mutengi weGPU ndangariro mukureba-mamiriro ekushandira.

Ndeipi nzira inoderedza saizi yeKV cache nekugovana makiyi uye kukosha pamisoro yekutarisisa?

Mumapoka-mubvunzo uye kutarisisa kwemibvunzo yakawanda regai misoro yemibvunzo yakawanda igovane mashoma makiyi / kukosha seti, ichideredza cache zvakanyanya.

Ndeapi zvikamu zviviri zveshanduro yetransformer inoenderana neKV cache?

Prefill inozadza cache neiyo K/V yekukurumidza; decode inowedzera imwe tokeni itsva K/V nhanho imwe neimwe uye inodzokorora pamusoro pecache.