João Alberto Fay

[João Alberto Fay @fay_joao]

@ciruai Can I run it in a Rtx 4060 (8gb VRAM)?

[João Alberto Fay @fay_joao]

@Haddad_Fernando O pior ministro da fazenda da história do país nao consegue calar a boca e só fala merda.

[João Alberto Fay @fay_joao]

@Teknium @iamlukethedev Opa, top demais ter esse reconhecimento

[João Alberto Fay @fay_joao]

@iamlukethedev Mandou muito bem, Luke

[João Alberto Fay @fay_joao]

@sudoingX I thought I would run on the Strix Halo easily, apparently not even on the DGX.

[João Alberto Fay @fay_joao]

@iamlukethedev Estamos aí usando no máximo que dá dentro de uma rtx com 8gb e frontier model nas subs. Comprei uma strix halo, esperando chegar, aí sim vai ficar legal.

[João Alberto Fay @fay_joao]

@glauber_doc Com isso a maioria hj se forma sem saber palpar um pulso distal, manejar TVP, diferenciar doença venosa de arterial, etc

[João Alberto Fay @fay_joao]

@ItsmeAjayKV I was running on an rtx 4060 (8gb) the qwen 3.6 35 A3b but after testing the Ornith 1.0 35b I preferred it. The ternary Bonsai found weak compared to these two. But I mean user experience, not benchmarks. I have not yet tested Gemma or Laguna.

[João Alberto Fay @fay_joao]

@NousResearch @blocks I have several profiles running separately one feeding the other via kanban and dashboard. Could this be a better integration than Kanban? Can anyone explain to me if there is a difference?

[João Alberto Fay @fay_joao]

@ModelScope2022 @KwaiAICoder No vision even if I try to run qwen's mmproj with it?

[João Alberto Fay @fay_joao]

@erenciracioglu @sudoingX It worked. Thanks for the suggestion. Another question, already abusing your goodwill: There's a MTP version?

[João Alberto Fay @fay_joao]

@erenciracioglu @sudoingX Does it have a mmproj or another vision tool? I use it with some frequency on my hermes setup

[João Alberto Fay @fay_joao]

@iotcoi That's nice. For what kind of work you use each model?

[João Alberto Fay @fay_joao]

@witcheer This post came at a great time. I am waiting for a GMKTec X2 to arrive and I will need to take my Hermes there. Good to know that I won't have to start from scratch on the new machine.

[João Alberto Fay @fay_joao]

@GabiiAH11 @0xSero Ran it in a 4060 laptop 8gb VRAM: 96k context, with mmproj and got about 27tks decoding

[João Alberto Fay @fay_joao]

@clarapachrc Que criança imbecil

[João Alberto Fay @fay_joao]

@CesinhaRonaldo @DamaDeFerroTV Gosto muito dele desde a rádio, mas é outro vendido tb

[João Alberto Fay @fay_joao]

@LeBunkerBtc A Europa acabou hahahah

[João Alberto Fay @fay_joao]

@SandroNascimm Jornalista falando merda. Nada de novo no bostil