DeepSeek LLM is an advanced language model available in both 7 billion and 67 billion parameters. Both a chat and base variation are available.

• Superior General Capabilities: DeepSeek LLM 67B Base outperforms Llama2 70B Base in areas such as reasoning, coding, math, and Chinese comprehension.

• Proficient in Coding and Math: DeepSeek LLM 67B Chat exhibits outstanding performance in coding (using the HumanEval benchmark) and mathematics (using the GSM8K benchmark).

References

GitHub

HuggingFace