Hermes 3: Instruction-Tuning Dataset for MLX-LM

Hermes 3 is a dataset converted for direct use with the MLX-LM and MLX-LM-LoRA frameworks. The dataset was uploaded by the mlx-community author and was last updated on July 29, 2025. Its specific content and size are not detailed in the provided metadata.

Use Cases

• Fine-tuning language models using the MLX-LM-LoRA framework based on the provided example command.
• Training language models using the standard MLX-LM framework based on the provided example command.
• Instruction-tuning models with sequences up to 4096 tokens in length, as suggested by the example parameter.

Strengths

• Specifically formatted for compatibility with the MLX-LM and MLX-LM-LoRA frameworks, reducing preprocessing effort.
• Last updated on July 29, 2025, indicating recent maintenance.

Limitations

• Description metadata is limited; actual data quality, content, and size require manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale training.

Provenance

Source

mlx-community

Freshness

Last updated 2025-07-29 18:07:44