KwaiAICoder
@KwaiAICoder
Kwaipilot team, illuminating the world with AI and building dreams with code. The official account of the Kwaipilot team (Kuaishou's Large Model for coding).
The ascent of Gen Z founders showcases the great potential in redefining AI Coding. 🌟 In the talent intensive dev tool market, from serial entrepreneurs to corporate veterans, these exceptional leaders are shaping the future of development. 🚀 In Feature II, we dive into the…
🚀 Excited to introduce KAT-V1 (Kwaipilot-AutoThink) – a breakthrough 40B large language model from the Kwaipilot team! KAT-V1 dynamically switches between reasoning and non-reasoning modes to address the “overthinking” problem in complex reasoning tasks. Key Highlights: 📌 40B…
🌟 Introducing KwaiCoder-23BA4-v1 🚀 23B-wide MoE architecture code completion model 💰Trained at 1/30 of the cost compared to traditional methods 🏆 Achieves SOTA performance on multiple code-related evaluation datasets 🔗 Check out the details: huggingface.co/Kwaipilot/Kwai…
Model Performance for Kwai-Coder-DS-V2-Lite-Base!（4/4） 「MATH and GSM8K test sets」 🧮 Surpassed Deepseek-v2-Lite-Base with improvements of 3.79% and 21.46% respectively 📙 Outperformed CodeLlama-34B with substantial improvements of 90.95% and 40.03% respectively 🔥 Surpassed…
Performance for Kwai-Coder-DS-V2-Lite-Base!（3/4） 「BigCodeBench-Complete-Full Set」 🚀 Achieved a 6% improvement over DeepSeek-Coder-33B, reaching SOTA levels 「BigCodeBench-Complete-Hard Set」 🔥 Significantly outperformed the CodeLlama-70B and the Qwen2.5-Coder-7B 🔗…
Performance for Kwai-Coder-DS-V2-Lite-Base!（2/4） 「MBPP and MBPP+ Test Sets」 🌟 Outperformed Deepseek-v2-Lite-Base 📈 Achieved an average improvement of nearly 5 percentage points over Qwen2.5-Coder-7B 🔗 Model: huggingface.co/Kwaipilot/Kwai…
Performance for Kwai-Coder-DS-V2-Lite-Base!（1/4） 「HumanEval and HumanEval+ Test Sets」 🎯 Achieved Pass@1 scores of 75.0% and 68.9% respectively 📈 Represented an improvement of 83.37% and 102.05% over Deepseek-v2-Lite-Base 🏅 Surpassed OpenCoder-8B, achieving SOTA levels 🔗…
🚀 Excited to introduce Kwai-Coder-DS-V2-Lite-Base! Kuaishou has released the open-source model Kwai-Coder-DS-V2-Lite-Base, which is built on Deepseek-v2-Lite-Base. This model features a total of 16B parameters, with 2.4B activated parameters, and was developed at a remarkably…
Surpassing OpenAI with Just 5M Tokens? Kuaishou releases OASIS Code Embedding Model! Recently, the Kwaipilot team officially announced the open-sourcing of the OASIS (Optimized Augmentation Strategy for Improved Code Search) Code Embedding Model. Kwaipilot collaborates with the…