Tiiuae
Rank:Average Model Cost: $0.0000
Number of Runs: 1,386,559
Models by this creator
falcon-7b
falcon-7b
Falcon-7B is a large language model with 7 billion parameters. It is a causal decoder-only model trained on 1,500 billion tokens of RefinedWeb data that has been enhanced with curated corpora. Falcon-7B outperforms other open-source models and features an optimized architecture for inference. It is available under the Apache 2.0 license and is suitable for research purposes and further specialization. However, it is important to note that Falcon-7B is trained on English and French data only and may carry biases and stereotypes found online. It is recommended to fine-tune the model for specific tasks and use appropriate precautions for production use. The model was trained using a distributed training infrastructure and is available for use with PyTorch 2.0 and the transformers library.
$-/run
409.4K
Huggingface
falcon-7b-instruct
falcon-7b-instruct
Falcon-7B-Instruct is a 7B parameter causal decoder-only model that has been fine-tuned on a mixture of chat and instruct datasets. It is based on Falcon-7B, a high-performing model trained on a large amount of data. The model is optimized for inference and features FlashAttention and multiquery techniques for improved performance. Falcon-7B-Instruct is available under the Apache 2.0 license and is recommended for direct use in chat and instruct applications. It has certain limitations, such as being primarily trained on English data and carrying biases commonly found online. Users are advised to take appropriate precautions for production use. More information about the model's architecture, training data, and technical specifications are provided. Falcon-7B-Instruct was trained on AWS SageMaker using distributed training methods. Citation details and contact information are also provided.
$-/run
401.1K
Huggingface
falcon-40b-instruct
falcon-40b-instruct
Falcon-40B-Instruct is a causal decoder-only model based on Falcon-40B and finetuned on a mixture of Baize. It is optimized for chat/instruct tasks and features an architecture optimized for inference. The model is trained on English and French data and is available under the Apache 2.0 license. It is recommended for direct use but may carry biases and limitations. Users are advised to take appropriate precautions for any production use. The model was trained on AWS SageMaker using a custom distributed training codebase and requires at least 85-100GB of memory for inference. Paper and evaluation details are coming soon.
$-/run
288.5K
Huggingface
falcon-40b
falcon-40b
Falcon-40B is a large language model developed by the Technology Innovation Institute (TII). It has 40 billion parameters and was trained on one trillion tokens of data. Falcon-40B is a causal decoder-only model optimized for inference and features an architecture with FlashAttention and multiquery. It is available under the Apache 2.0 license and is suitable for research on large language models and for further specialization and fine-tuning for specific use cases. However, it is important to note that Falcon-40B is primarily trained on English, German, Spanish, and French, with limited capabilities in other languages. It may carry biases and stereotypes commonly found online. TII is calling for proposals from users worldwide to submit their ideas for Falcon-40B's deployment. To use Falcon-40B, it is recommended to have at least 85-100GB of memory.
$-/run
266.4K
Huggingface
falcon-rw-1b
falcon-rw-1b
The falcon-rw-1b model is a text generation model specifically designed for the application of simulating natural language responses in conversational agents. It is trained to generate contextually relevant and coherent responses to text prompts, making it a valuable tool for building chatbots, virtual assistants, and other dialogue systems.
$-/run
18.5K
Huggingface
falcon-rw-7b
$-/run
2.7K
Huggingface