YuanGongND / ltu

Code, Dataset, and Pretrained Models for Audio and Speech Large Language Model "Listen, Think, and Understand".

YuanGongND/ltu Issues

why pad_or_trim use 1000 rather than 3000 when transcribe_audio?
Updated a month ago
Where use whisper decoder?
Updated a month ago
Why loss is always 0?
Updated 2 months ago
Stage training sh scripts for low resource
Updated 2 months ago1
where to download whisper model?
Updated 2 months ago2
Error when run finetune_toy_low_resource.sh
Updated 2 months ago
train_scripts
Updated 2 months ago3
Modifications to the llama model
Updated 2 months ago1
Question: Half Float Inference?
Updated 2 months ago1
About the audio-text pair of AudioSet dataset.
Updated 2 months ago1
Question: LLaMA-7B LLM
Updated 3 months ago2
Issue with Loading 13B Model: Size Mismatch Error
Updated 3 months ago4
Running Issue about Low-Resource Training for LTU-AS
Updated 3 months ago8
Question：Why are the prompts for training and inference for audio event classification are different?
Updated 4 months ago2
OpenAQA Dataset's audio files
Updated 4 months ago2
question on cutoff_len
Updated 4 months ago1
Eval code error
Updated 4 months ago4
LICENSE of AQA datasets and checkpoints
Updated 4 months ago
Eval_metrics
Closed 4 months ago4
Question about Finetune exp
Updated 5 months ago4
Question about vicuna version
Closed 5 months ago2
Maximux Length for LTU-AS Audio Input
Updated 5 months ago1
Batch Inference Support
Closed 6 months ago
CPU local inference is not working.
Updated 6 months ago
Question about model loading in inference
Updated 6 months ago2
Issue while loading openaqa_5.6M.json
Updated 7 months ago4
Question about LTU-AS Downstream Tasks
Updated 7 months ago3
Question about Multi-GPU Training
Updated 7 months ago1
Question about LTU-AS base model
Updated 7 months ago1
LTU_AS ASR Task
Updated 7 months ago4
Missing Checkpoints
Closed 7 months ago8
OpenAQA Dataset Access
Updated 7 months ago2
extract_whisper_feature.py
Updated 7 months ago3
Missing Tokenize Audio Info during Fine-tuning/Training
Updated 7 months ago1
How to process audio that exceeds 10 seconds in length
Updated 7 months ago5
no evaluation script for open-set problem
Updated 7 months ago1
whisper-at on cuda:1
Updated 7 months ago1
Requirements for more pretrained weights
Closed 7 months ago5
Model Parallelization
Updated 8 months ago5
vicuna_ltu model file missing
Updated 8 months ago1
Which model is 7B (Default) and which is 13B (Beta)?
Updated 8 months ago12
Question about the Realism of Simulated Acoustic Event Combinations in Data Generation
Updated 8 months ago2
About the experimental results of the paper LTU-AS
Updated 10 months ago1
API access to Gradio demos broken?
Closed a year ago9
Questions about data construction
Updated a year ago1