A more compliant, RP-oriented version of MTSAIR/multiversemodel, fine-tuned on carefully selected datasets. It’s smart, adept at following the desired markdown format and adhering to the provided character card. The first message of the character card significantly influences its writing style. Pulsar7B pairs well with guidance from CFG Scale and works effectively with PLists + Ali:Chat character cards. Pulsar7B was fine-tuned on the following datasets: — grimulkan/theory-of-mind — grimulkan/physical-reasoning — ResplendentAI/LunaAlpaca — unalignment/toxic-dpo-v0.2 — kira/math-dpo — athirdpath/DPOPairs-Roleplay-Alpaca-NSFW-v1-SHUFFLED Thanks to mradermacher, static GGUF quants are available here. Pulsar7B works well with Alpaca, it’s not a picky model when it comes to formatting/preset. Mistral should be compatible too. The custom chat template from MTSAIR/multiverse_model also performs well:
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: rmdhirr
Теги: mistral, text-generation-inference, unsloth, trl, dpo, uncensored, roleplay, fine-tune
Лайков: 3 | Загрузок: 26
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.