Kedis ku xidhan hagaha · Dhexdhexaad Heerka
Yi Models by 01.AI Kedis
Dib u eeg asalka qoyska nooca Yi, diiradda laba-luqadlaha, naqshadaynta-kaliya, kala duwanaanshaha macnaha guud, iyo waxbarashada ku jihaysan tababarka dambe.
Su'aal 1 ee 8
Yaa aasaasay 01.AI, shirkadda ka dambeysa moodooyinka Yi?
Horudhac su'aal kasta8 su'aalo
Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.
- Yaa aasaasay 01.AI, shirkadda ka dambeysa moodooyinka Yi?
- Kai-Fu Lee
- Robin Li
- Jack Ma
- Andrew Ng
- Maxay ahaayeen labada nooc ee saldhigga Yi ee ugu horreeyay ee la sii daayay?
- Yi-1 iyo Yi-2
- Yi-6B iyo Yi-34B
- Yi-100 iyo Yi-200
- Yi-Mini iyo Yi-Max
- Awood noocee ah ayaa moodooyinka Yi si gaar ah loogu yaqaanaa?
- Soo saarida sawirada kaliya
- Ku shaqeeya iyada oo aan wax GPU ah
- Waxqabad xooggan oo labada luqadood ku hadla ee Shiine-Ingiriisi
- Turjumaada kaliya ee codka lagu hadlo
- Waa maxay nasabka dhismaha aasaasiga ah ee moodooyinka Yi raacaan?
- Naqshad LSTM ah oo soo noqnoqota
- Qaab fidin
- Qaabka codeeyaha saafiga ah sida BERT
- Beddeliyaha-qaabka Llama-kaliya
- Qiyaastii intee in le'eg ayay tahay daaqadda macnaha guud ee kala duwanaanshaha macnaha-dheer ee Yi?
- 200k calaamadood
- 4K calaamado
- 32K calaamadood
- 1 milyan oo calaamad
- What did 01.AI’s Yi technical report identify as an important driver of model performance?
- Katagitaanka AI gebi ahaanba
- Data quality resulting from data-engineering efforts
- Kaliya dhis chips hardware
- Ku biirista OpenAI
- What does supervised fine-tuning plus reinforcement learning from human feedback aim to improve in Yi chat models?
- Compressing retinal images
- Changing the model into an encoder-only network
- Following instructions in conversational use
- Removing all text-generation ability
- What data strategy did 01.AI emphasize for the Yi models?
- Training exclusively on synthetic audio
- Avoiding all data filtering
- Using only programming-language source code
- Careful data quality and curation rather than relying only on sheer scale