this projector is so fucking high res and expensive there’s literally two of them
callie
@callie.on-her.computer
663 followers 878 following 4.8K posts
check out my other website computers.sex/~callie/
check out my other website computers.sex/~callie/
![Terminal screenshot showing repeated model performance logs, with the right side of each line cut off:
```text
usage: 0.09, accept len: 3.23, accept rate: 0.32, cuda graph: True, gen throughput (token/s): 261.34, #queue-req:
[2026-08-16 06:16:46] Decode batch, #running-req: 2, #full token: 72155, full token usage: 0.08, mamba num: 6, mamba...
usage: 0.09, accept len: 3.55, accept rate: 0.36, cuda graph: True, gen throughput (token/s): 287.69, #queue-req:
[2026-08-16 06:16:47] Decode batch, #running-req: 2, #full token: 72394, full token usage: 0.08, mamba num: 6, mamba...
usage: 0.09, accept len: 2.99, accept rate: 0.28, cuda graph: True, gen throughput (token/s): 241.97, #queue-req:
[2026-08-16 06:16:48] Decode batch, #running-req: 2, #full token: 72582, full token usage: 0.08, mamba num: 6, mamba...
usage: 0.09, accept len: 2.36, accept rate: 0.19, cuda graph: True, gen throughput (token/s): 191.29, #queue-req:
[2026-08-16 06:16:49] Decode batch, #running-req: 2, #full token: 72816, full token usage: 0.08, mamba num: 6, mamba...
usage: 0.09, accept len: 2.92, accept rate: 0.28, cuda graph: True, gen throughput (token/s): 236.59, #queue-req:
[2026-08-16 06:16:50] Decode batch, #running-req: 2, #full token: 73040, full token usage: 0.08, mamba num: 6, mamba...
usage: 0.09, accept len: 2.77, accept rate: 0.25, cuda graph: True, gen throughput (token/s): 224.54, #queue-req:
[2026-08-16 06:16:51] Decode batch, #running-req: 2, #full token: 73344, full token usage: 0.08, mamba num: 6, mamba...
usage: 0.09, accept len: 3.86, accept rate: 0.41, cuda graph: True, gen throughput (token/s): 312.52, #queue-req:
[2026-08-16 06:16:52] Decode batch, #running-req: 2, #full token: 73576, full token usage: 0.08, mamba num: 6, mamba...
usage: 0.09, accept len: 2.84, accept rate: 0.26, cuda graph: True, gen throughput (token/s): 229.49, #queue-req:
[2026-08-16 06:16:53] Decode batch, #running-req: 2, #full token: 73889, full token usage: 0.08, mamba num: 6, mamba...
usage: 0.09, acc](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:wl4wyug27klcee5peb3xkeut/bafkreihbankjz64ds7zq3fzdtmr2tqjfqripcf2zxvezajr7hdm3dm5sia)
