Contributions de Pitpitt


Rechercher des contributionsaffichermasquer
⧼contribs-top⧽
⧼contribs-date⧽
(les plus récentes | les plus anciennes) Voir ( | ) (20 | 50 | 100 | 250 | 500)

4 janvier 2024

2 janvier 2024

1 janvier 2024

30 décembre 2023

  • 10:4330 décembre 2023 à 10:43 diff hist +1 111‎ N Emu Video ‎ Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' Emu Video''' == Anglais == ''' Emu Video ''' Emu Video, a text-to-video generation model that factorizes the generation into two steps: first generating an image conditioned on the text, and then generating a video conditioned on the text and the generated image. We identify critical design decisions--adjusted noise schedules for diffusion, and multi-stage training--that enable us to direct... »
  • 10:4230 décembre 2023 à 10:42 diff hist +981‎ N Segment Anything ‎ Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Segment Anything'''  Segment Anything (SA) project: a new task, model, and dataset for image segmentation. Using our efficient model in a data collection loop, we built the largest segmentation dataset to date (by far), with over 1 billion masks on 11M licensed and privacy respecting images. The model is designed and trained to be promptable, so it can transfe... »
  • 10:3930 décembre 2023 à 10:39 diff hist +605‎ N Mistral 7B ‎ Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' Mistral 7B''' == Anglais == ''' Mistral 7B''' The Mistral 7B paper introduces a compact yet powerful language model that, despite its relatively modest size of 7 billion tokens, outperforms its larger counterparts, such as the 13B Llama 2 model, in various benchmarks. (Next to the two-times larger Qwen 14B, Mistral 7B was also the base model used in the winning solutions of this year's Ne... »
  • 10:3830 décembre 2023 à 10:38 diff hist +1 836‎ N Optimisation directe des préférences ‎ Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' XXXXXXXXX ''' == Anglais == ''' Direct Preference Optimization''' While large-scale unsupervised language models (LMs) learn broad world knowledge and some reasoning skills, achieving precise control of their behavior is difficult due to the completely unsupervised nature of their training. Existing methods for gaining such steerability collect human labels of the relative quality of model g... »
  • 10:3730 décembre 2023 à 10:37 diff hist +824‎ N Adaptation par modèle auxiliaire quantifiée ‎ Page créée avec « ==en construction== == Définition == XXXXXXXXX == Français == ''' QLoRA ''' == Anglais == ''' QLoRA''' QLoRA stands for quantized LoRA (low-rank adaptation). The standard LoRA method modifies a pretrained LLM by adding low-rank matrices to the weights of the model's layers. These matrices are smaller and, therefore, require fewer resources to update during finetuning. In QLoRA, these low-rank matrices are quantized, meaning their numerical precision is... »
  • 10:3630 décembre 2023 à 10:36 diff hist +554‎ N LlaMA 2 ‎ Page créée avec « ==en construction== == Définition == XXXXXXXXX voir LlaMA == Français == ''' LlaMA 2''' == Anglais == ''' LlaMA 2''' what differentiates the Llama 2 suite from many other LLMs is that the models come as standard pretrained models and chat models that have been finetuned via reinforcement learning with human feedback (RLHF, the method used to create ChatGPT) to follow human instructions similar to ChatGPT — RLHF-finetuned models are still rare. <sm... »

29 décembre 2023

28 décembre 2023

(les plus récentes | les plus anciennes) Voir ( | ) (20 | 50 | 100 | 250 | 500)
Récupérée de « https://datafranca.org/wiki/Spécial:Contributions/Pitpitt »