Dria
@driaforall
Followers
7K
Following
2K
Statuses
407
Dria democratizes access to high-quality synthetic data to improve AI through a multi-agent network
Joined January 2024
Introducing Dria-Agent-α, the first agentic LLM trained for Pythonic function calling: - 3B parameters, yet matches GPT-4o in BFCL - Parallel multi-function calls at once - Reasoning beyond JSON's limitations - Trained with synthetic data generated by 2000 edge devices
23
109
737
RT @cognitivecompai: Following up - I announce the release of the Dolphin-R1 dataset with Apache 2.0 license! Half Gemini Flash Thinking a…
0
76
0
RT @driaforall: We’re open-sourcing the Pythonic Function Calling Dataset for Dria-Agent-α—alongside the synthetic data generation pipeline…
0
8
0
2-) Build custom function-calling agents for your workflows using the Synthetic Data Generation Pipeline:
1
0
7
Avaliable models: DEEPSEEK_R1_1_5B = "deepseek-r1:1.5b" DEEPSEEK_R1_7B = "deepseek-r1:7b" DEEPSEEK_R1_8B = "deepseek-r1:8b" DEEPSEEK_R1_14B = "deepseek-r1:14b" DEEPSEEK_R1_32B = "deepseek-r1:32b" DEEPSEEK_R1_70B = "deepseek-r1:70b" Generate now:
1
0
30
RT @rsarrow: Ever sit down and think "Gee, I wish I could monitor decentralized data generation in real time?" Well now you can, thanks to…
0
5
0
Decentralized Synthetic Data Generation across the globe shortens the gap between Closed Labs and Open Source. Distill task-specific models from great open-source reasoner models like @deepseek_ai R1 and feed reasoning traces to your model at the lowest cost.
5
8
61
RT @mervenoyann: Everything that happened this week in open AI, a recap 🤠 👀 Multimodal - MiniCPM-o 2.6 is a new sota any-to-any model by @…
0
43
0
RT @mervenoyann: Dria has been cooking models and benchmarks for agentic capabilities (the hot topic of this year) 🤖💗 make sure to check t…
0
5
0
DPAB-α is a comprehensive benchmark designed to evaluate LLMs function calling capabilities through both Pythonic and JSON-based approaches. This benchmark contains 100 synthetically generated and validated problems across different difficulty levels. Benchmark: Blog:
0
1
12
RT @digitizedweekly: 1/ Introducing: Dria-Agent-ɑ An LLM focused on Pythonic function calls, addressing the constraints of JSON-based tool…
0
6
0
RT @gm8xx8: Dria-Agent-α uses Pythonic Function Calling, enabling LLMs to interact with tools through Python code instead of JSON schemas.…
0
8
0