AI model inference training

Reflection Llama-3.1 70B is currently the world's top open source large language model (LLM). It is trained using a new technology called Reflection-Tuning, which enables the model to detect errors in its reasoning and make corrections. The model was trained on synthetic data, generated by Glaive. Glaive is an excellent tool for users who are training models. The model uses the standard Llama 3.1 chat format, with special tags to differentiate between the model's internal thinking and the final answer, improving the user experience.

MiniCPM3-4B is the third generation of the MiniCPM series, surpassing Phi-3.5-mini-Instruct and GPT-3.5-Turbo-0125 in overall performance and on par with many recent 7B to 9B models. Compared with the previous two generations, MiniCPM3-4B has greater versatility and supports function calls and code interpreters, making it more widely applicable to various scenarios. In addition, MiniCPM3-4B has a 32k context window, and with LLMxMapReduce technology, it can theoretically handle unlimited contexts without requiring a lot of memory.

C4AI Command R+ 08-2024 is a 104B parameter large-scale research release model with highly advanced capabilities including Retrieval Augmentation Generation (RAG) and tool usage to automate complex tasks. The model supports training in 23 languages ​​and evaluation in 10 languages. It is optimized for multiple use cases, including reasoning, summarization, and question answering.

Mistral NeMo is a 12B model built by Mistral AI in partnership with NVIDIA, with a large context window of 128k tokens. It leads the way in reasoning, world knowledge, and coding accuracy. The model is designed for global multilingual applications and supports multiple languages ​​including English, French, German, Spanish, Italian, Portuguese, Chinese, Japanese, Korean, Arabic and Hindi. Mistral NeMo also uses a new tokenizer, Tekken, to improve text and source code compression efficiency. Additionally, the model has been fine-tuned with instructions to improve its ability to follow precise instructions, reason, handle multiple rounds of conversations, and generate code.

The Llama-3-Groq-8B-Tool-Use model is an 8B parameter causal language model specially designed for advanced tool usage and function calling tasks. The model has an optimized transformer architecture and is trained on the Llama 3 8B base model with full fine-tuning and direct preference optimization (DPO). It excels at tasks involving API interaction, structured data manipulation, and complex tool usage. However, users should be aware that this model may produce inaccurate or biased content in some cases, and users will need to implement appropriate security measures for their specific use cases.

Mistral Small is a new model optimized for low latency and cost, superior to Mixtral 8x7B with lower latency and a superior intermediate solution between open weight products and flagship models. Mistral Small features the same innovations as Mistral Large, including RAG-enablement and function calls. We have simplified endpoint provisioning, offering open weight endpoints at competitive pricing, and introduced new optimized model endpoints mistral-small-2402 and mistral-large-2402.