LC Workflow Enablement Group (WEG) and LC DOERs are happy to announce that we recently upgraded our on-prem LC Large Language Model hosted via LLamaMe to the Llama 3.3 70B model in the CZ and SCF. As part of this upgrade we also increased the maximum context length to 64k in the CZ and 100k in the SCF.
Large Language Models hosted in LC
LLM Details
| LC Network Zone | Large Language Model | Max Content Length | Infrastructure GPU |
|---|---|---|---|
| Collaboration Zone (CZ) | Meta-Llama-3.3-70B-Instruct | 64000 | 2 A100 80GB RAM |
| Secure Compute Facility (SCF) | Meta-Llama-3.3-70B-Instruct | 110000 | 4 H100 80GB RAM |
See the LLamaMe documentation for more details: https://hpc.llnl.gov/services/cloud-services/ai-ml-services/llamame-llm…