X-LANCE/SLAM-LLM

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

1066 stars · 118 forks observed · 2026-10-03

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

SLAM-LLM is a deep learning toolkit for training custom multimodal large language models focused on speech, language, audio, and music processing. It provides training recipes, high-performance checkpoints, and support for large-scale industrial training with DeepSpeed and multi-GPU distributed inference.

Use cases

• train a multimodal LLM for speech recognition
• fine-tune an LLM on audio data
• build a voice interaction assistant with controllable timbre
• train ASR models on 100,000 hours of audio
• run distributed multi-GPU inference for speech models
• adapt an LLM to music or audio understanding tasks

When to choose

• you need to train or fine-tune speech/audio/music multimodal LLMs
• you want reproducible recipes and pretrained checkpoints for speech-LLM research
• you need large-scale industrial ASR/ST training with DeepSpeed

When to avoid

• you only need inference of off-the-shelf speech APIs without training
• you work outside Linux/CUDA environments
• you need a no-code or production SaaS speech solution

Facets

framework · maturity active

machine-learning deep-learning llm-training audio-processing speech-recognition nlp speech-processing large-language-models machine-learning deep-learning python multimodal-llm speech music peft deepspeed asr pytorch audio linux gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="X-LANCE/SLAM-LLM")