Google's Gemma 7B model has been fine-tuned to work with 15 Indian languages, enabling developers to build advanced AI models. Navarasa 2.0, a chat model with optimized parameters, has been deployed and praised for its capabilities. Mistral 7B and Gemma 2B models are being deployed serverless using Hugging Face and Beam Cloud.
Last week, we released Navarasa 2.0 - A Gemma 7B and 2B instruction-tuned model in 15 Indian Languages (+ English) But the real test of a model is its deployment and usage in the real-world! If you are wondering how to deploy Navarasa 2.0, we have put together a deployment… https://t.co/eJFi5WlMUG
🚀 Deploying Navarasa-2.0 - a Gemma finetuned model in 15 Indian Languages 🌟 On March 18th, 2024, we proudly unveiled Navarasa-2.0 - a cutting-edge Gemma 7B/2B instruction-tuned model proficient in 15 Indian languages (+ English). We are thrilled to observe the growing… https://t.co/waXa6wnrfi
🍀 In our latest blog, we implemented a simple inference engine using @huggingface to deploy Mistral 7B serverless using Modal Labs. In this follow-up post, we'll achieve the same goal with @GoogleAI 's Gemma 2B model. But this time leveraging @beam_cloud as the serverless…
Someone deployed our model Navarasa 2.0 (15 langs) as a clean chat UI we can use on the internet! Playing it as a chat model with optimized parameters, streaming inference, and clean UI actually blew our minds in what the model is capable of! Lot more work to be done on… https://t.co/UaAGc33M55
Introducing Navarasa 2.0! Our own @ravithejads has been doing amazing work: fine-tuning a new version of Google Gemma 7B to work with 15 Indian languages! Localizing gen AI models to regional languages is valuable (and complex!) work, allowing developers to build advanced RAG… https://t.co/qsntwtJBXL