AI - Meta AI OPT-66B

Leading developer of large-scale language models based on Transformers architecture, renowned for generating human-like text through Opt-66b.

Logo of Meta AI OPT-66B
Last Audited At

About Meta AI OPT-66B

Meta AI OPT-66B is a leading developer of large-scale language models based on the Transformers architecture. Their flagship model, Opt-66b, is a causal language model that utilizes AutoModelForCausalLM and AutoTokenizer for processing and generation. This model is renowned for its ability to generate human-like text based on given prompts. The preprocessing of texts involves the use of Byte Pair Encoding (BPE) with a vocabulary size of 50272, tokenizing sequences into 2048 consecutive tokens. Opt-66b was trained on approximately 175 billion model parameters across 992 80GB A100 GPUs over roughly 33 days of continuous training. The model is accessible through the Transformers library and has been downloaded over 15,000 times in the last month. OPT is open-source and available under a non-restrictive license, with resources including a model card, extensive documentation, and an active community of users for collaboration and learning.

Screenshot of Meta AI OPT-66B Website

More companies

Cambricon

Leading tech firm specializes in AI chips & accelerators, providing high-performance, efficient industry solutions via extensive product line and open software platform.

Read more

PagerDuty

Ensuring operational integrity through industry-leading incident management and operations solutions for various industries.

Read more

Google Colab

Empowering researchers and engineers with accessible advanced computational tools for AI and ML through Google Colab's user-friendly cloud platform.

Read more

Tell us about your project

Our Hubs

London, United Kingdom

A global AI hotspot, thrives on innovation, diverse talent, and a dynamic tech ecosystem, offering unparalleled opportunities for AI engineers.

Munich, Germany

A vibrant AI hub, merges cutting-edge technology with rich cultural experiences, creating an inspiring environment for AI engineers.