SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2301.12503 · ICML · 2023

AudioLDM: Text-to-Audio Generation with Latent Diffusion Models

Danilo Mandic, Zehua Chen, Haohe Liu, Yi Yuan, Xinhao Mei, Xubo Liu, Wenwu Wang, Mark Plumbley, Vae Encoder

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 21 functions out of this paper's own repositories and ran 15 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
haoheliu/AudioLDM canonical 5 of 9
haoheliu/audioldm_eval pwc_unofficial 9 of 11
copy not recorded — 1 of 1
FunctionStatusWhere it lives
DiagonalGaussianDistribution Ran haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("2cda99b6def85873")
clip_bce Ran haoheliu/audioldm_eval/audioldm_eval/feature_extractors/panns/losses.py
code served (permissive licence) · get_code("97d45ba74474fdfc")
disabled_train Ran this paper's copy was not recorded; identical code first harvested from zhouc20/LatentGraphDiffusion
pointer only · get_code("4cb732f513d69dfd")
extract_into_tensor Ran haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("8971042e12016a60")
get_loss_func Ran haoheliu/audioldm_eval/audioldm_eval/feature_extractors/panns/losses.py
code served (permissive licence) · get_code("e424ea90cdf3fd97")
get_mel_from_wav Ran haoheliu/audioldm_eval/audioldm_eval/audio/tools.py
code served (permissive licence) · get_code("af0a67f05ca8144d")
griffin_lim Ran haoheliu/audioldm_eval/audioldm_eval/audio/audio_processing.py
code served (permissive licence) · get_code("9f9ec26d4cdfdf7d")
interpolate_bilinear_2d_like_tensorflow1x Ran haoheliu/audioldm_eval/audioldm_eval/feature_extractors/inception3.py
code served (permissive licence) · get_code("21e50186ab33dd24")
load_json Ran haoheliu/audioldm_eval/audioldm_eval/audio/tools.py
code served (permissive licence) · get_code("777e661635ced528")
load_module2model Ran haoheliu/audioldm_eval/audioldm_eval/feature_extractors/melception_audioset.py
code served (permissive licence) · get_code("ffc354ebaecde324")
load_pickle Ran haoheliu/audioldm_eval/audioldm_eval/audio/tools.py
code served (permissive licence) · get_code("78222542af18c6c5")
make_ddim_sampling_parameters Ran haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("3ee640131c9d4362")
make_ddim_timesteps Ran haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("0ea4e960ea54514c")
noise_like Ran haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("2d7200ac60d9a4bd")
pad_short_audio Ran haoheliu/audioldm_eval/audioldm_eval/datasets/load_mel.py
code served (permissive licence) · get_code("e0e3ae45a19356fa")
DDIMSampler Not yet run haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("332deacf272549e4")
DDPM Not yet run haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("b7521b6c520cc691")
DiffusionWrapper Not yet run haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("bb9a12b66a9694f1")
LatentDiffusion Not yet run haoheliu/AudioLDM/audioldm/ldm.py
pointer only (licence: NOASSERTION) · get_code("1d0e58f8ba043419")
dynamic_range_compression Not yet run haoheliu/audioldm_eval/audioldm_eval/audio/audio_processing.py
code served (permissive licence) · get_code("0b80ec09e3ba47ae")
load_npy_data Not yet run haoheliu/audioldm_eval/audioldm_eval/datasets/load_mel.py
code served (permissive licence) · get_code("98e449b3a6b45350")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Text-to-audio (TTA) systems have recently gained attention for their ability to synthesize general audio based on text descriptions. However, previous studies in TTA have limited generation quality with high computational costs. In this study, we propose AudioLDM, a TTA system that is built on a latent space to learn continuous audio representations from contrastive language-audio pretraining (CLAP) embeddings. The pretrained CLAP models enable us to train LDMs with audio embeddings while providing text embeddings as the condition during sampling. By learning the latent representations of audio signals without modelling the cross-modal relationship, Au-dioLDM improves both generation quality and computational efficiency. Trained on AudioCaps with a single GPU, AudioLDM achieves stateof-the-art TTA performance compared to other open-sourced systems, measured by both objective and subjective metrics. AudioLDM is also the first TTA system that enables various textguided audio manipulations (e.g., style transfer) in a zero-shot fashion. Our implementation and demos are available at https://audioldm. github.io.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2301.12503")
get_code_for_paper("2301.12503")
have("2301.12503")

Connect an agent — have() is free.