Kuzvichengetera-mu-mudziyo Meta's Llama models paNVIDIA Tesla P40 (24 GB VRAM, Pascal architecture). Kuenderana ne8B models pachokwadi chakazara uye mazhinji mamodheru kana achiitwa. GPU servers hazvigone kupihwa; taura nesu kuti uzive kana zvaitwa.
$ curl -fsSL https://ollama.com/install.sh | sh && ollama run llama3
# Mufananidzo wekuisa command
Llama ndeimwe yemhuri yeMeta yezvinyorwa zvemavara zvinowanikwa zvakasununguka. Kushandisa Llama paGPU yako inochengeta maprompts nedata zvakavanzika.
Kufamba 8B mamodheru pa 24 GB ye VRAM. 70B mamodheru haasi kunyatsoenderana pamwechete 24 GB kadhi pazviyero zvakajairika quantumization.
Maitiro ako uye data anogara paseva yako, pasina API yekukumbira kwechitatu bato.
Usashandisa GGUF nelama.cpp kana Ollama, kana nePyTorch.
Kushandira mamodheru ne Ollama kana llama.cpp server yeOpenAI-inowirirana endpoint.
LLaMA inogona kushandisa NVIDIA GPU kuti iite basa rayo. VPS.org GPU servers hazvina kuwanikwa kuti utore; taura nesu kuti uzive kana vachiri. LLaMA inowanzo GPU-yakakwira basa. Ita shuwa kuti software yako ine CUDA rutsigiro uye kuti yako driver / runtime versions inosangana nezvinodiwa zvebasa re LLaMA.
GPU servers hadzina kuwanikwa kuti utore. Nhasi unogona kuisa LLaMA pane yakajairika CPU package, uko inoita basa rakakura kupfuura pane GPU: SSH in uye shandisa curl -fsSL https://ollama.com/install.sh | sh && ollama run llama3.
VRAM zvinodiwa LLaMA kuenderana nemhando uye batch size; tarisa chirongwa ruzivo rwezvinoreva.
GPU mutengo achaburitswa kana GPU servers anogona kupihwa. Tibate kana iwe uchida kuziva kana ichiitika.
Yega. Every VPS.org server inouya nefull root access, saka unogona kutamba zvimwe zvinhu pamwe LLaMA mukati server's RAM uye kuchengetera.
Every VPS.org server inouya nepfupi root SSH kuwanikwa, saka unogona kuisa madhiraivha, mabhuku uye LLaMA asi iwe unofanirwa.
Hapana server inouya neCUDA runtime kana NVIDIA mutyairi akaiswa, uye GPU servers hazvina kuwanikwa kuti aite kusvika parizvino. Unogona kuisa zvinyorwa LLaMA zvinoda iwe pachako.
VPS.org servers ndezvemapurojekiti akaomarara, kwete ephemeral instances: LLaMA models, configs uye data inogara paSSD storage yeserver kusvika wabvisa.
Tenga data rekushandisa pane SSD ye server yekuwedzera kuwanikwa kwenguva pfupi panguva ye LLaMA runs, uye kopa zvakapera zviwanikwa (zviyero, zvizvarwa, embeddings) kune yekuchengetera kunze kwe VPS.org; snapshots uye backups vanogara pamwechete kuchengetera cluster.
Unogona kusimudzira chero VPS.org server kune yakakura chirongwa kubva kudzora panel; resize rinotangazve uye yako LLaMA kuiswa inotora pamusoro. GPU servers hazvina kuwanikwa kusvika parizvino.
Yeah. Backups ndivo akabhadhara add-on (zuva nezuva + 30% yemari chirongwa, akachengetwa 7 mazuva; zuva nezuva + 20%, akachengetwa 28 mazuva) uye snapshots vakasununguka, izvo zvinokosha pamberi LLaMA mazuva. Both stay on the same storage cluster as the server.
Contact support within 30 days of your first purchase to request a refund. GPU servers are not sold yet, so no GPU order is involved.
GPU kugona kuri kuwanikwa chete. Tiudzai nezvebasa rako uye tichakuudza kuti ripi rinowanikwa uye ripi risiri.