# Mlx Lm

> You can use use the`mlx-lm`package to fine-tune an LLM with low rank

## Pages

- [Fine-Tuning with LoRA or QLoRA](lora.md): You can use use the`mlx-lm`package to fine-tune an LLM with low rank
- [Api_Benchmark](api-benchmark.md): def setup_arg_parser():
- [Api_Cache_Prompt](api-cache-prompt.md): def setup_arg_parser():
- [Api_Chat](api-chat.md): def setup_arg_parser():
- [Api_Convert](api-convert.md): def mixed_quant_predicate_builder(
- [Api_Evaluate](api-evaluate.md): """
- [Api_Fuse](api-fuse.md): def parse_arguments() -> argparse.Namespace:
- [Api_Generate](api-generate.md): def str2bool(string):
- [Api_Gguf](api-gguf.md): class TokenType(IntEnum):
- [Api_Lora](api-lora.md): """^(?:
- [Api_Manage](api-manage.md): def tabulate(rows: List[List[Union[str, int]]], headers: List[str]) -> str:
- [Api_Perplexity](api-perplexity.md): """
- [Api_Sample_Utils](api-sample-utils.md): def make_sampler(
- [Api_Server](api-server.md): def get_system_fingerprint():
- [Api_Tokenizer_Utils](api-tokenizer-utils.md): class StreamingDetokenizer:
- [Api_Upload](api-upload.md): def main():
- [Api_Utils](api-utils.md): def _unpack_awq_weights(qweight: mx.array) -> mx.array: