A new Text-to-SQL generative AI model, SambaCoder-nsql-Llama-2-70B, has been jointly developed by Numbers Station and SambaNova Systems, setting a new benchmark in SQL query generation. The model has achieved a notable accuracy of 78.1% on the Spider test set, outperforming the previous leader, GPT-4. This advancement has been highlighted across various platforms, with both companies expressing their excitement over the collaboration and the model's performance. SambaNova Systems has also taken a step further by open-sourcing the model and releasing its artifacts, aiming to contribute to the AI community's growth. The model's development involved pre-training on the SQL splits of the Big Code Project's stack dataset, followed by fine-tuning, which has been instrumental in surpassing GPT-4's execution accuracy on the Spider test.
This week, @SambaNovaAI launched the SambaCoder-nsql-llama2-70B model, a new Text-to-SQL generative #AI solution. Check out more below about how the #LLM can unlock developer productivity. https://t.co/JBcmgo8ouB
MODEL RELEASE: @NumbersStnAI and SambaNova Systems are proud to release SambaCoder-nsql-Llama-2-70B, a Text-to-SQL model that surpasses GPT-4 on the spider benchmark! We pre-trained @AIatMeta ‘s Llama-2-70B on the SQL splits of @BigCodeProject ‘s stack dataset and further fine… https://t.co/xptFPdUYfi
🧱How Dataherald Makes Natural Language to SQL Easy Text-to-SQL is a HUGE use case, and @DataheraldCo is the open-source leader in the space This is a great look behind the curtains to see what makes it tick. Blog: https://t.co/njcpu77zYo https://t.co/euKWSSsDxz
Announcing SambaCoder-nsql-Llama-2-70B, a collaboration w/ @SambaNova! 🚀 🏅This model sets a new benchmark in SQL query generation, achieving 78.1% accuracy on the Spider test set, outperforming GPT-4 👀Learn more in our blog! https://t.co/pFloHLnXjW #AI #SQL #enterprisedata
TECH BLOG: We are excited to open source and release the artifacts of our collaboration with Numbers Station on a Text-to-SQL generative AI solution. The SambaCoder-nsql-llama2-70B model surpasses GPT-4 in execution accuracy on the spider test. https://t.co/spm0t4wQyM