EnthrallingEmil

3 ms·

1. 1.▲
   by EnthrallingEmil 6mo ago

   Also 3090. using Q4_XL, reduced max context size, with 100k prompt length, I get 2520 tk/s for prompt processing, 68 token/s generation: llama-server \ --model /mnt/ubuntu/models/llama-cpp-qwen/Qwe
2. 2.▲
   by EnthrallingEmil 6mo ago

   Gemini had a partial outage on their status page and then once it was over after 10 days, it became a single day partial outage. For me, it was nearly two weeks of 95% failure rates.
3. 3.▲
   by EnthrallingEmil 6mo ago

   And again just now... Again nothing on status. Maybe they sould vibe code some probers and automatic testing.