Baptiste Rozière
@b_roziere
Followers
2K
Following
384
Statuses
116
Research scientist at Mistral AI. Working on ML for programming languages. Lead author of CodeLlama, now working on the Codestral models!
Paris, France
Joined September 2016
Today, we release CodeLlama, a collection of base and instruct-finetuned models with 7B, 13B and 34B parameters. For coding tasks, CodeLlama 7B is competitive with Llama 2 70B and CodeLlama 34B is state-of-the-art among open models. Paper and weights:
8
87
363
RT @MistralAI: Introducing the all new Le Chat: your ultimate AI sidekick for life and work! Now live on web and mobile!
0
2K
0
RT @MistralAI: Le Chat is fast (1,100 tok/s for flash queries on an updated Mistral Large). Download it at or http…
0
372
0
RT @MistralAI: Introducing Small 3, our most efficient and versatile model yet! Pre-trained and instructed version, Apache 2.0, 24B, 81% MM…
0
495
0
RT @MistralAI: magnet:?xt=urn:btih:11f2d1ca613ccf5a5c60104db9f3babdfa2e6003&dn=Mistral-Small-3-Instruct&tr=udp%3A%2F%
0
470
0
If it takes too long to start generating your completion, it will be irrelevant by the time it arrives. No model that's fast enough for code completion comes close to Codestral 2501's ELO on copilot arena.
0
0
4
TestGenEval evaluates models for generating full test suites or adding test cases to an existing test suite. On top of the pass rate, it measures the quality of the test suite using the coverage and mutation rate. Check out the paper and Kush's online leaderboard!
Thrilled to announce our new work TestGenEval, a benchmark that measures unit test generation and test completion capabilities. This work was done in collaboration with the FAIR CodeGen team. Preprint: Leaderboard:
0
0
12
RT @syhw: Just gave a talk on "Grounding LLMs in Code Execution" at the NeurIPS Hacker-Cup AI Competition, here are the slides https://t.co…
0
36
0
RT @cognitivecompai: Today, I experimented with @continuedev + @ollama for day-to-day troubleshooting and coding tasks. I tested Codestral…
0
16
0
RT @GuillaumeLample: Nice review (in French) of le Chat: where Pixtral Large easily answers questions about complex…
0
16
0
RT @GuillaumeLample: Today we are releasing a Batch API and a Moderation API ! The batch API allows to process high-volume requests to all…
0
6
0
RT @LucileSaulnier: 🌟 AI enthusiasts! Join @MistralAI and shape the future of generative AI! 🌟 We're hiring AI Scientists, Research Engine…
0
32
0
RT @dchaplot: We just released the weights of Pixtral 12B base model on HuggingFace: Pixtral 12B Base: Also link…
0
97
0
RT @sophiamyang: 🎉 We @MistralAI just dropped two new edge models - Ministral 3B and Ministral 8B!
0
104
0
You can now find the slides of my LLM4Code talk here. Thanks to @Lin0Tan and all the organizers of the llm4code workshop for organizing this talk.
If you missed the fantastic keynotes on #CodeLlama and #Starcoder2 from their initiators and core contributors @b_roziere and @LoubnaBenAllal1, you can still check out the slides Hope to see your submissions to the 2nd @LLM4Code workshop. #LLM4code #llm
0
0
17
RT @Lin0Tan: If you missed the fantastic keynotes on #CodeLlama and #Starcoder2 from their initiators and core contributors @b_roziere and…
0
4
0