mradermacher

/

model_requests

like 126

Edens-Fall-L3.3-70b-0.3b, Edens-Fall-L3.3-70b-0.3c are booth high-priority queued both for static and imatrix quants! :D
EdensFall-L3.3-70b-0.1a, Eden-L3.3-70b-0.1, Liliths-Whisper-L3.3-70b-0.1 and Apocrypha-L3.3-70b-0.3 are all medium priority queued both for static and imatrix quants as they are all interesting models worth quantizing in my opinion. Don't hesitate to request any new model you create.

You can check for progress at http://hf.tst.eu/status.html
Or regularly check the model summary pages for quants to appear under:

• https://hf.tst.eu/model#Edens-Fall-L3.3-70b-0.3b-GGUF
• https://hf.tst.eu/model#Edens-Fall-L3.3-70b-0.3c-GGUF
• https://hf.tst.eu/model#EdensFall-L3.3-70b-0.1a-GGUF
• https://hf.tst.eu/model#Eden-L3.3-70b-0.1-GGUF
• https://hf.tst.eu/model#Liliths-Whisper-L3.3-70b-0.1-GGUF
• https://hf.tst.eu/model#Apocrypha-L3.3-70b-0.3-GGUF

For some reason rain took 3 of the priority 0 ones and budget softlocked itself. Maybe specifying worker back,rain,kaos,leia,marco,rich1,nico2 disabled some protections against this. I only did so to ensure they don't get queued on nico1 which beside marco would usually take the 70B models.

In any case I did llmc nuke Apocrypha-L3.3-70b-0.3 && llmc add force 0 si https://huggingface.co/BruhzWater/Apocrypha-L3.3-70b-0.3 worker back,kaos,leia,marco,rich1,nico2 && llmc push to hopefully fix this.

Wow that allocated it to rich1 which usually doesn't take 70B models anymore so manually listing workers indeed must disable the usually limits. That's interesting. In any case rain is still softlocked so llmc nuke Liliths-Whisper-L3.3-70b-0.1 && llmc add force 0 si https://huggingface.co/BruhzWater/Liliths-Whisper-L3.3-70b-0.1 worker marco && llmc push and llmc nuke Eden-L3.3-70b-0.1 && llmc add force 0 si https://huggingface.co/BruhzWater/Eden-L3.3-70b-0.1 worker marco && llmc push. Nice so that fixed the situation. I keep learning something new every day.

Edens-Fall-L3.3-70b-0.3b, Edens-Fall-L3.3-70b-0.3c are booth high-priority queued both for static and imatrix quants! :D
EdensFall-L3.3-70b-0.1a, Eden-L3.3-70b-0.1, Liliths-Whisper-L3.3-70b-0.1 and Apocrypha-L3.3-70b-0.3 are all medium priority queued both for static and imatrix quants as they are all interesting models worth quantizing in my opinion. Don't hesitate to request any new model you create.

You can check for progress at http://hf.tst.eu/status.html
Or regularly check the model summary pages for quants to appear under:

• https://hf.tst.eu/model#Edens-Fall-L3.3-70b-0.3b-GGUF
• https://hf.tst.eu/model#Edens-Fall-L3.3-70b-0.3c-GGUF
• https://hf.tst.eu/model#EdensFall-L3.3-70b-0.1a-GGUF
• https://hf.tst.eu/model#Eden-L3.3-70b-0.1-GGUF
• https://hf.tst.eu/model#Liliths-Whisper-L3.3-70b-0.1-GGUF
• https://hf.tst.eu/model#Apocrypha-L3.3-70b-0.3-GGUF

Good looks! Big thanks!

Maybe specifying worker

Yes, llmjob trusts you know what you are doing, so when you tell it to schedule a model on rain, it might assign it to it. Same is true when you "force" a model or set other settings (such as quants). In all cases, it will do as told when explicitly told to do so, whether that causes disaster or not.