datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
audio-spectogram-transformer-checkpointlisten-to-reason-checkpoints
LISTEN-to-Reason — checkpoints
Graph + retrieval index + prototypes for LISTEN-to-Reason:
a frozen text LLM answers audio questions from a serialized multimodal knowledge graph, never
hearing the clip and never being fine-tuned.
No audio is redistributed here — only CLAP embeddings, graph structure, and the reference
captions. See the attribution table for the licence that follows those captions.
git clone https://github.com/poonehmousavi/listen-to-reason && cd listen-to-reason… See the full description on the dataset page: https://huggingface.co/datasets/poonehmousavi/listen-to-reason-checkpoints.checkpoint_similarity_scores_with_audiocheckpoints-young-male-co-spanishwpp_pav_transcrito_wav2vec2-portuguese-wpp-checkpoint-480wav2vec2-20pct-20250624-122057-checkpoint-18000ORIGINAL_wav2vec2-portuguese-wpp-checkpoint-480dac_inference_1B_TBD-LLaMA-DAC-Denoiser-checkpoint-70001B-DAC-SE2_1B_np_UPSAMPLE_checkpoint_op_v3TBD-LLaMA-DAC-Denoiser-checkpoint-12200_correct_codebookTBD-LLaMA-DAC-Denoiser-packet-loss-fine_tuned-checkpoint-1000-checkpoint-1000_correct_codebook1B_natural_noise_fine_tuned-checkpoint-2600_correct_codebookTBD-LLaMA-DAC-Denoiser-checkpoint-7800_correct_codebook_disco