Ecosyste.ms: Issues
An open API service for providing issue and pull request metadata for open source projects.
GitHub / lucidrains/audiolm-pytorch issues and pull requests
#282 - Description of wandb logs
Issue -
State: open - Opened by anuragkumar95 5 days ago
#281 - CUDA error: an illegal memory access was encountered
Issue -
State: open - Opened by bulieme 7 days ago
#280 - Does SoundStream support streaming inference? If so, can you provide the relevant code?
Issue -
State: open - Opened by dengcunqin 10 days ago
#279 - SoundStream training hangs using accelerate launch and the use_finite_scalar_quantizer=True setting
Issue -
State: closed - Opened by ThomasLWang 23 days ago
- 23 comments
#278 - How to use multiple GPUs for training and inference?
Issue -
State: open - Opened by Oyiyi 3 months ago
#277 - what's your loss rate?
Issue -
State: open - Opened by Oyiyi 3 months ago
#276 - Trying to overfit SounsStream
Issue -
State: open - Opened by hishammadcor 4 months ago
#275 - Why is Encodec only encoding 1 frame?
Issue -
State: closed - Opened by sivannavis 6 months ago
- 1 comment
#274 - checkpoint
Issue -
State: open - Opened by why414 8 months ago
#273 - Classifier for detecting synthetic speech
Issue -
State: open - Opened by Ashigarg123 8 months ago
#272 - Model cascade training
Issue -
State: open - Opened by a897456 9 months ago
#271 - Audiolm as an embedder model?
Issue -
State: open - Opened by Darel13712 9 months ago
#270 - Soundstream training using birdsongs. Any guidance appreciated!
Issue -
State: open - Opened by haydensflee 10 months ago
#269 - About get_embeds function
Issue -
State: open - Opened by jihoojung0106 10 months ago
- 1 comment
#268 - Why not use the output of Attention in Transformer?
Issue -
State: closed - Opened by jihoojung0106 10 months ago
#267 - AssertionError: File Not Found: data/hyp.scratch.yaml
Issue -
State: closed - Opened by zrshello 10 months ago
- 1 comment
#266 - skip the eos when adding offset to avoid overlapping
Pull Request -
State: closed - Opened by biendltb 10 months ago
- 5 comments
#265 - fix wrong tensor assignment of the output of attention
Pull Request -
State: closed - Opened by biendltb 10 months ago
- 1 comment
#264 - Training dataset
Issue -
State: open - Opened by hahust191806 11 months ago
#263 - Missing softmax after Linear layer
Issue -
State: closed - Opened by biendltb 11 months ago
- 1 comment
#262 - Cannot retrieve dependency version for gateloop-transformer>=0.5.2, possible regression?
Issue -
State: closed - Opened by afreemanio 11 months ago
- 1 comment
#261 - Removal of the last token id from fine_token_ids in FineTransformerWrapper.forward()
Issue -
State: closed - Opened by biendltb 11 months ago
- 1 comment
#260 - Fix #259
Pull Request -
State: closed - Opened by orrp 11 months ago
- 1 comment
#259 - `data_max_length_seconds` causes typecheck error in `CoarseTransformerTrainer`
Issue -
State: closed - Opened by orrp 11 months ago
#258 - `use_wandb_tracking` was not stored in most Trainers when it is `False`
Pull Request -
State: closed - Opened by orrp 12 months ago
- 1 comment
#257 - Added wandb tracking to SemanticTransformerTrainer, CoarseTransformerTrainer, and FineTransformerTrainer
Pull Request -
State: closed - Opened by LukasNel 12 months ago
- 1 comment
#256 - Soundstream discriminator clip_grad_norm - some params are not clipped.
Issue -
State: closed - Opened by avihu111 12 months ago
- 3 comments
#255 - Gradient Issue when Finetuning
Issue -
State: closed - Opened by tysonjordan about 1 year ago
#254 - Error in exporting soundstream to onnx
Issue -
State: open - Opened by kalradivyanshu about 1 year ago
- 14 comments
#253 - Only noise as a result
Issue -
State: open - Opened by mpastewski about 1 year ago
- 4 comments
#251 - Update RVQ projection layers during training
Pull Request -
State: closed - Opened by ilya16 about 1 year ago
- 1 comment
#249 - Question: Random semantic embedding in SemanticTransformer?
Issue -
State: open - Opened by stg1205 about 1 year ago
- 1 comment
#248 - bugfix - swap codec variable for course wrapper
Pull Request -
State: closed - Opened by rgxb2807 about 1 year ago
- 1 comment
#247 - I very thanks for your work. But when i train the soundstream model, why does it need a pre-trained Encodec and then error?
Issue -
State: closed - Opened by DingWeiPeng about 1 year ago
#246 - IndexError Using Encodec and setting return_coarse_generated_wave=True
Issue -
State: closed - Opened by rgxb2807 about 1 year ago
- 5 comments
#244 - Fixed typo in README.md
Pull Request -
State: closed - Opened by y4umeng about 1 year ago
#243 - Bugfix - Fixing validation dataset variable on FineTransformerTrainer
Pull Request -
State: closed - Opened by rgxb2807 about 1 year ago
- 1 comment
#242 - Question: Any way to specify validation dataset for SemanticTransformer, CoarseTransformer and FineTransformer?
Issue -
State: closed - Opened by rgxb2807 about 1 year ago
- 2 comments
#241 - Question: Checkpoint of the model
Issue -
State: open - Opened by fernandals about 1 year ago
- 1 comment
#240 - Question: Are there any work arounds for using DeepSpeed for multi-gpu training
Issue -
State: open - Opened by rgxb2807 about 1 year ago
- 4 comments
#239 - Question: How to load pythorch format as HubertWithKmeans?
Issue -
State: open - Opened by Selectorrr about 1 year ago
#238 - Question on discrepancy between original data and reconstructed data sizes
Issue -
State: open - Opened by tysonjordan about 1 year ago
- 1 comment
#236 - Bug in generation when generating with Encodec
Issue -
State: closed - Opened by FrancescoVV about 1 year ago
- 7 comments
#235 - have trouble to generate semantic tokens using the demo code
Issue -
State: closed - Opened by dwangF0 about 1 year ago
- 2 comments
#234 - multi-gpu training not working with accelerate
Issue -
State: closed - Opened by FrancescoVV about 1 year ago
- 13 comments
#233 - Does VALL-E follow the same semantic/coarse hierarchical structure as AudioLM?
Issue -
State: closed - Opened by williamluer about 1 year ago
- 3 comments
#231 - Likely beartype package breakage
Issue -
State: closed - Opened by rsxdalv about 1 year ago
- 1 comment
#230 - Question about 'attention bias not supported for flash attention'
Issue -
State: open - Opened by amitaie about 1 year ago
- 2 comments
#229 - Error when running 3rd cell in demo ipynb
Issue -
State: closed - Opened by uday18git about 1 year ago
#228 - Code Refactoring
Pull Request -
State: closed - Opened by tosemml about 1 year ago
- 1 comment
#227 - pretrained soundstream weights?
Issue -
State: open - Opened by muazhuda about 1 year ago
- 1 comment
#226 - Dependency error
Issue -
State: closed - Opened by amrzv about 1 year ago
- 4 comments
#225 - bandwidth params not work!
Issue -
State: closed - Opened by wotulong over 1 year ago
- 3 comments
#224 - Inconsistent samples for multiple targets in SoundDataset
Issue -
State: closed - Opened by ilya16 over 1 year ago
- 2 comments
#223 - Average validation loss across grad_accum_every
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 1 comment
#222 - Dataloader save
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 6 comments
#221 - Soundstream Training Goes From Great to Horrible
Issue -
State: open - Opened by adamfils over 1 year ago
- 11 comments
#219 - hi. If i want to train text to Chinese speech audiolm , what confuse me is that pretrained hubert model is English-style, Does it affect my Chinese version training? Or i have to re-train hubert with my own large chinese dataset ? Appreciate !!!! @lucidrains
Issue -
State: closed - Opened by hyhzl over 1 year ago
- 4 comments
#218 - Is resampling needed when using EnCodec?
Issue -
State: open - Opened by m-pana over 1 year ago
- 7 comments
#216 - yolov5
Issue -
State: open - Opened by fangwei888 over 1 year ago
- 1 comment
#215 - AssertionError: only one Trainer can be instantiated at a time for training
Issue -
State: open - Opened by tiansiyuan over 1 year ago
- 2 comments
#214 - Separate transformer and trainer checkpoint load logic
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 1 comment
#213 - How to run the inference?
Issue -
State: closed - Opened by LianaN over 1 year ago
- 6 comments
#212 - TypeError: cannot unpack non-iterable NoneType object
Issue -
State: closed - Opened by tiansiyuan over 1 year ago
- 16 comments
#211 - When running the example code, I get an error where the trainer says to be instantiated twice:
Issue -
State: closed - Opened by LukasNel over 1 year ago
- 6 comments
#210 - accelerate's wait_for_everyone hangs on the final step of training coarse/fine transformer
Issue -
State: closed - Opened by LWprogramming over 1 year ago
- 4 comments
#209 - Accelerate failing on multi-gpu rng synchronization
Issue -
State: closed - Opened by LWprogramming over 1 year ago
- 15 comments
#208 - Question about length of data in training \ generating
Issue -
State: closed - Opened by amitaie over 1 year ago
#207 - Implement accelerate support for semantic/coarse/fine transformers
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 2 comments
#206 - Change torch.no_grad() to torch.inference_mode()
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 1 comment
#205 - Max length fix
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 1 comment
#204 - Questions about training Soundstream: poor intelligibility and gradients explosion after 10k steps. (sr=16k, B=96)
Issue -
State: open - Opened by Makiyuyuko over 1 year ago
- 1 comment
#203 - OpenBLAS/OpenMP Loop error message
Issue -
State: closed - Opened by LWprogramming over 1 year ago
- 8 comments
#201 - generation form of the inference
Issue -
State: closed - Opened by Hit1ron over 1 year ago
- 4 comments
#200 - Eos handling
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 3 comments
#199 - Audio generation failing at FineTransformer
Issue -
State: closed - Opened by LWprogramming over 1 year ago
- 15 comments
#198 - pin sklearn exactly to 0.24.0
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 7 comments
#197 - 'SemanticTransformerWrapper' object has no attribute 'embed_text'
Issue -
State: closed - Opened by jinyuli over 1 year ago
- 4 comments
#196 - A problem with EncodecWrapper()
Issue -
State: closed - Opened by Leezp99 over 1 year ago
- 3 comments
#195 - Bug fix in encodec.py
Pull Request -
State: closed - Opened by yang1fan2 over 1 year ago
- 1 comment
#194 - `MultiScaleDiscriminator` differs from paper
Issue -
State: closed - Opened by haydenshively over 1 year ago
- 3 comments
#193 - Improve ComplexConv2d FSDP compatibility
Pull Request -
State: closed - Opened by haydenshively over 1 year ago
- 3 comments
#192 - Question about the generate
Issue -
State: open - Opened by asr-pub over 1 year ago
#191 - Poor audio quality
Issue -
State: closed - Opened by cpdu over 1 year ago
- 1 comment
#190 - a small question about the loss function
Issue -
State: closed - Opened by PB20000090 over 1 year ago
- 1 comment
#189 - Fix off-by-one error in train step update
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 1 comment
#188 - Fix path type
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
#187 - Load from correct self.steps to resume training
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 1 comment
#186 - question about the semantic process
Issue -
State: closed - Opened by asr-pub over 1 year ago
- 3 comments
#185 - Loss about CoarseTransformerWrapper
Issue -
State: closed - Opened by asr-pub over 1 year ago
- 2 comments
#184 - Something wrong when i use the “soundstream” repo
Issue -
State: open - Opened by wangyuxuan11 over 1 year ago
- 1 comment
#183 - training data
Issue -
State: open - Opened by linlongrd over 1 year ago
#182 - can not install audiolm-pytorch
Issue -
State: closed - Opened by linlongrd over 1 year ago
- 4 comments
#181 - /path/to/audio/files
Issue -
State: open - Opened by linlongrd over 1 year ago
#179 - more demo needed
Issue -
State: open - Opened by fire-keeper over 1 year ago
#177 - Use a pretrained model as a discriminator (and for feature maps)
Issue -
State: closed - Opened by turian over 1 year ago
- 5 comments
#176 - BIG NEWS!!!!!! Encodec License change MIT License!!!
Issue -
State: closed - Opened by fd873630 over 1 year ago
#175 - update readme encodec mit license
Pull Request -
State: closed - Opened by LWprogramming over 1 year ago
- 1 comment
#174 - ComplexSTFTDiscriminator diverges from paper
Issue -
State: closed - Opened by eagomez2 over 1 year ago
- 5 comments
#169 - Potentially useful pretrained models
Issue -
State: closed - Opened by gkucsko over 1 year ago
- 2 comments