@Doragrynn Essaie de rendre sur blender cycles ou c4d octane, rien qu'une simple scene. Vois combien de temps tu mets pour rendre une seule frame 😅 trou de cul qui parles d'optimisation, tu sais pas de quoi tu parles le mieux c'est de te taire
@sharbel Iteration budget reached (60/60) everytime I use Hermes agent, was told that it was autonomous
and self improving, but constantly hitting tool calls limits, am I missing something?
@Lndsp12@spiritbuun I checked yestarday, the last llama.cpp version used in LM Studio was released a week ago, There is no option for nightly builds, so it may take some time
@matthieurouif@huggingface@julien_c@ClementDelangue OpenAI: It takes billions to train a model.
Photoroom: It takes billions of flux images to train a model
You did great, well played for that, but build a proper dataset and be humble, don't compare your model with openai this sounds stupid, especially when your dataset is synth
@LordVXrp Ne pariez jamais ce que vous n'êtes pas prêts a perdre, le marché n'a pas d'émotions vous pouvez tout perdre.
Apprenez comment stocker et échanger vos actifs avent de faire n'importe quoi avec votre argent
@str0ngl4nd@aakashgupta ppl are lazy, you can set guards:
Session Initialization/Stop loading history on every message
Model Routing/Use cheap models (or even local models) for simple tasks/summarise context
Move your heartbeat checks to a free local LLM
Prevent runaway automation from burning tokens