We lifted 11 functions out of this paper's own repositories and ran 2 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.
| Repository | Role | Ran |
|---|---|---|
| soroushmehr/sampleRNN_ICLR2017 | canonical | 2 of 3 |
| deepsound-project/samplernn-pytorch | pwc_unofficial | 0 of 5 |
| dada-bots/dadabots_sampleRNN | pwc_unofficial | 0 of 3 |
| Function | Status | Where it lives |
|---|---|---|
| linear2mu | Ran | soroushmehr/sampleRNN_ICLR2017/datasets/dataset.py code served (permissive licence) · get_code("bf123cbc1a5c7c22") |
| mu2linear | Ran | soroushmehr/sampleRNN_ICLR2017/datasets/dataset.py code served (permissive licence) · get_code("bfa157a2f01ee323") |
| find_dataset | Not yet run | soroushmehr/sampleRNN_ICLR2017/datasets/dataset.py code served (permissive licence) · get_code("b66f9b2d11a10ef2") |
| find_dataset | Not yet run | dada-bots/dadabots_sampleRNN/datasets/dataset.py code served (permissive licence) · get_code("483324e37147887a") |
| gradient_clipping | Not yet run | deepsound-project/samplernn-pytorch/optim.py code served (permissive licence) · get_code("76a3848fdc07a751") |
| linear2mu | Not yet run | dada-bots/dadabots_sampleRNN/datasets/dataset.py code served (permissive licence) · get_code("bcb1aa73abd49bb2") |
| linear_dequantize | Not yet run | deepsound-project/samplernn-pytorch/utils.py code served (permissive licence) · get_code("10c1a66056242af7") |
| linear_quantize | Not yet run | deepsound-project/samplernn-pytorch/utils.py code served (permissive licence) · get_code("1d698a1b8ebdffef") |
| mu2linear | Not yet run | dada-bots/dadabots_sampleRNN/datasets/dataset.py code served (permissive licence) · get_code("08d55cf94660e24e") |
| q_zero | Not yet run | deepsound-project/samplernn-pytorch/utils.py code served (permissive licence) · get_code("75e69c94dae9a379") |
| sequence_nll_loss_bits | Not yet run | deepsound-project/samplernn-pytorch/nn.py code served (permissive licence) · get_code("c05eb4083a6e42d3") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
In this paper we propose a novel model for unconditional audio generation based on generating one audio sample at a time. We show that our model, which profits from combining memory-less modules, namely autoregressive multilayer perceptrons, and stateful recurrent neural networks in a hierarchical structure is able to capture underlying sources of variations in the temporal sequences over very long time spans, on three datasets of different nature. Human evaluation on the generated samples indicate that our model is preferred over competing models. We also show how each component of the model contributes to the exhibited performance.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("1612.07837")
get_code_for_paper("1612.07837")
have("1612.07837")
Connect an agent — have() is free.