Instruction-tuned LLM model distillation
22 repos
Fine-tuned and distilled variants of large language models optimized for instruction-following and text generation tasks. This cluster centers on creating smaller, more efficient models through distillation techniques applied to base architectures like Mistral and Yi, with a focus on maintaining instruction-following capability across various model sizes (7B to 34B parameters). Repositories here contain model weights, training approaches, and implementations compatible with standard text generation endpoints.
distillation ·2,946
endpoints_compatible ·2,780
transformers ·2,780
text-generation ·2,780
instruct ·2,767
finetune ·2,759
en ·2,759
text-generation-inference ·2,756
gpt4 ·2,718
pytorch ·2,451
