This model LiquidAI/LFM2.5-24B-A2B-MLX-6bit was converted to MLX format from LiquidAI/LFM2.5-24B-A2B using mlx-lm version 0.32.0. Quantized to 6 bits per weight (group size 64, affine mode). MoE routing gates kept unquantized in float32. For base model details, capabilities, and recommended generation settings, see the original model card. Requires mlx-lm 0.32.0 or later. Earlier versions load the model but route the experts incorrectly. LFM2.5 uses a ChatML-like format. See the Chat Template documentation for details. Example: Because LFM2.5-24B-A2B is a reasoning model, assistant turns contain an explicit chain of thought before the final answer. tokenizer.applychattemplate() formats messages automatically. LFM2.5 supports function calling. Pass tools to tokenizer.applychattemplate(…, tools=…). See the Tool Use documentation for the full guide. LFM Open License v1.0 — see the LICENSE file or the original model card for terms.
Модальности:
Генерация текста
Области применения:
Диалог / чат
Задача: Генерация текста
Автор: LiquidAI
Теги: mlx, lfm2_moe, liquid, lfm2.5, edge, moe, conversational, en
Лайков: 3 | Загрузок: 4,195
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
