AMD ROCm GPU Kobold AI running and benchmarking LLAMA 8B. 13B, 8x7 MoE Mixtral and 70B LLMs

Save
8 months ago
16
Technology AI KoboldAI Kobold LLMs LLM Llama Llama2 Rocm Gpu AMD

KoboldC runs 8B LLAMA3, outdated 13B Hypermantis GGML. 8x7 MoE Mistral/Mixtral and a 70B LLAMA3 based model after being asked why Open source and decentralized alternatives like Newpipe or Ody Blockchain or hardened mail like Mailfence and alias protection with a RX 7800XT AMD GPU without the older CLblast mode and with ROCm. Trident Z5 Neo 6000 MhZ overclocked RAM and a Ryzen 5 7600X Zen 4 AM5 CPU.

Mirrored on Odysee. 29th july 2024

Revolt (Discord FOSS)rvlt.gg/8h3PzaWt
Bluesky (Twitter FOSS)bsky.app/profile/newjersey14
https://odysee.com/$/invite/7Bi1qdNoJ968zfe3bQe5Lo7KjrsRRJL3
https://dlive.io?ref=asleepinglaffey
Kick-I use the same name

Loading comments...