Topics
Latest
AI
Amazon
Image Credits:Rafael Henrique/SOPA Images/LightRocket / Getty Images
Apps
Biotech & Health
Climate
Cloud Computing
Commerce
Crypto
Enterprise
EVs
Fintech
Fundraising
Gadgets
stake
Government & Policy
Hardware
Layoffs
Media & Entertainment
Meta
Microsoft
seclusion
Robotics
protection
Social
Space
Startups
TikTok
Transportation
speculation
More from TechCrunch
Events
Startup Battlefield
StrictlyVC
Podcasts
television
Partner Content
TechCrunch Brand Studio
Crunchboard
Contact Us
French AI inauguration Mistral is present new AI model customization selection , including pay plans , to lease developers and enterprises o.k. - line its reproductive role model for picky use case .
The first is self - service . Mistral has free a software development kit ( SDK),Mistral - Finetune , for fine - tuning its models on workstation , server and small-scale datacenter nodes .
In the readme for the SDK ’s GitHub repository , Mistral notice that the SDK is optimise for multi - GPU setup but can descale down to a undivided Nvidia A100 or H100 GPU for fine - tune up small models like Mistral 7B. Fine - tuning on a dataset such as UltraChat , a compendium of 1.4 million duologue with OpenAI ’s ChatGPT , takes around half an hour using Mistral - Finetune across eight H100s , Mistral says .
For developers and companies that prefer a more manage solution , there ’s Mistral ’s newly launched fine - tuning services available through the company ’s API . Compatible with two of Mistral ’s models for now , Mistral Small and the aforementioned Mistral 7B , Mistral enunciate that the delicately - tuning services will gain support for more of its models in the coming weeks .
Lastly , Mistral is debuting usage training services , presently only available to take client , to fine - tune any Mistral model for an governance ’s apps using their data . “ This approach start the universe of highly specialized and optimized models for their specific domain , ” the company excuse in a berth on its officialblog .
Mistral , which my colleague Ingrid Lunden recently reported isseeking to elicit around $ 600 million at a $ 6 billion valuationfrom investors including DST , General Catalyst and Lightspeed Venture Partners , is no doubt looking to develop revenue as it face considerable — and growing — contender in the generative AI quad .
Since Mistral unveiled its first generative model in September 2023 , it ’s released several more , including acode - generating model , and roll out pay up genus Apis . But it has n’t disclosed how many users it has , nor what its revenues are look like .
