Kefuro-30B
Weights are not public. The code is open (Apache-2.0). Everything needed to run Kefuro on top of the...
Weights are not public. The code is open (Apache-2.0). Everything needed to run Kefuro on top of the...
Un modèle de langage (LLM) français construit de zéro — tokenizer, architecture, pré-entraînement, fine-tuning chat, jusqu’à l’infrastructure d’entraînement...
> NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 is the full-precision reference release of NVIDIA’s Nemotron 3.5 Lightning model, a hybrid Mixture-of-Experts architecture interleaving...
This model was converted to GGUF format from nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 using llama.cpp via the ggml.ai’s GGUF-my-repo space. Refer to...
Luciole-8B-Instruct-1.1-FP8 is a FP8-quantized version of Luciole-8B-Instruct-1.1 in the transformers / compressed-tensors (https://github.com/neuralmagic/compressed-tensors) format, quantized with LLM Compressor...
> ⚠️ Preliminary results — 1 seed, variance not controlled. Every number below is a single training run...
Взяв то, что хорошо работало с линией Crimson_Dawn, Effervescence-27B переносит ту же двухфазную методологию на гораздо большую базу,...
The post-training data has a cutoff date of November 28, 2025. The pre-training data has a cutoff date...
Villanova-2B-2603 is a fully open, multilingual instruction-tuned Large Language Model developed by Villanova.AI. Part of the Villanova project,...
Trinity-Large-Preview is a 398B-parameter sparse Mixture-of-Experts (MoE) model with approximately 13B active parameters per token. It is the...