RWKV-ASR: Speech Recognition Data for Frozen RWKV Language Models

An exploratory experiment to enable frozen pretrained RWKV language models to accept speech modality input. The dataset, created by author 'echodict', is hosted on Hugging Face and was last updated on 2026-04-01. It follows the SLAM_ASR approach to bridge the gap between text-trained LLMs and speech recognition tasks.

Use Cases

• Training speech recognition adapters based on the described SLAM_ASR methodology.
• Benchmarking frozen language model performance on audio input tasks.
• Researching neural audio codec integration for text-to-speech modality bridging.

Strengths

• Focuses on a specific, emerging research area: adapting frozen RWKV models for speech.
• Builds upon a documented methodology (SLAM_ASR) mentioned in the description.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and sample data are unknown, which may limit suitability assessment.

Provenance

Source

Hugging Face, uploaded by author 'echodict'.

Collection Method

Created as an exploratory experiment, following the SLAM_ASR idea.

Time Range

null

Freshness

Last updated 2026-04-01 03:37:59; freshness should be verified.

Geography

null