-
This issues describes the high level directions that "create LLM Engine V2". We want the design to be as transparent as possible and created this issue to track progress and solicit feedback.
Goal…
-
The CollimatorOutputFactor extracted from the plan.Trial file for the MU calculation is incorrect in the plan.Trial file and will result in wrong MU for beams with mixed models (e.g. motorized wedges …
-
the models/pac/jiggle/base_jiggle_0.mdl models are handy for making certain pac3s such as tails by popularity, but what if the jiggle models could act as a beam where a bone could represent an endpoin…
-
When simulating beam emission and attenuation, cherab requires data specifying how the beam interacts with *all* the species in the plasma, and falls over when it can't find it. In some ways this is s…
-
# Title
Integration of open-source AI models in Galaxy (web-based scientific data analysis platform)
# Description
[Galaxy (Europe)](https://usegalaxy.eu/) is a web-based, open-source data anal…
-
https://github.com/mozilla/firefox-translations-training/blob/03a2ddaa3f7d9c9af3a236bb2dbb94db36c16373/pipeline/translate/translate.sh#L22
When performing backtranslation, we want slightly differen…
-
https://git.astron.nl/RD/EveryBeam
currently, the software manually downloads beam models. Tammo-jan indicated astron wants to rewrite the code, and have some beam models without downloading. WHen …
-
- [ ] September: Finish implementing GPT2 Inference w/ float32, clang.
- [x] Solidify aIR System
- [x] Make aIR type-safe
- [x] Optimize AIR (FastGraph). Simplify GPT2 < 1.0s
- [x] E…
-
For example
`dials.index lots_of_things.expt lots_of_things.refl beam.fix=all detector.fix=all joint_index=false`
should maintain the same topology of shared models on output as exists on input …
-
I try inference my T5 model with C++ runtime used Paged KV at the commit `b777bd64750abf30ca7eda48e8b6ba3c5174aafd`. Its result is normal when inference with single input text, but with multiple input…