We lifted 16 functions out of this paper's own repositories and ran 10 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 |
|---|---|---|
| tensorflow/compression | canonical | 0 of 1 |
| google/codex | pwc_unofficial | 10 of 15 |
| Function | Status | Where it lives |
|---|---|---|
| autocorrelate | Ran | google/codex/codex/ems/fourier.py code served (permissive licence) · get_code("2b166835e4864877") |
| logsum_expbig_minus_expsmall | Ran | google/codex/codex/ems/continuous.py code served (permissive licence) · get_code("18b1ac950cad22c7") |
| lowpass | Ran | google/codex/codex/loss/wasserstein.py code served (permissive licence) · get_code("92ec5a215026fc33") |
| matrix_init | Ran | google/codex/codex/ems/deep_factorized.py code served (permissive licence) · get_code("9ccc5edc054d63f7") |
| quantize_distribution | Ran | google/codex/codex/ecs/range_ans.py code served (permissive licence) · get_code("69650369b53c555f") |
| safe_sqrt_jvp | Ran | google/codex/codex/loss/wasserstein.py code served (permissive licence) · get_code("92acca2b1a8712d6") |
| soft_round | Ran | google/codex/codex/ops/quantization.py code served (permissive licence) · get_code("9f7e2f9dec296743") |
| soft_round_conditional_mean | Ran | google/codex/codex/ops/quantization.py code served (permissive licence) · get_code("1d9e9b22846b022b") |
| soft_round_inverse | Ran | google/codex/codex/ops/quantization.py code served (permissive licence) · get_code("7577b798076f2f68") |
| verysoftplus | Ran | google/codex/codex/ops/activation.py code served (permissive licence) · get_code("ce43c2b2c164374a") |
| lower_limit | Not yet run | google/codex/codex/ops/gradient.py code served (permissive licence) · get_code("280b32e777201abd") |
| periodic_prob | Not yet run | google/codex/codex/ems/fourier.py code served (permissive licence) · get_code("5eefb49b52323703") |
| safe_sqrt | Not yet run | google/codex/codex/loss/wasserstein.py code served (permissive licence) · get_code("485c2547f2922965") |
| source_dataset | Not yet run | tensorflow/compression/models/toy_sources/compression_model.py code served (permissive licence) · get_code("ebb36a320973720b") |
| upper_limit | Not yet run | google/codex/codex/ops/gradient.py code served (permissive licence) · get_code("c6b3bdaa7c690524") |
| upper_limit_fwd | Not yet run | google/codex/codex/ops/gradient.py code served (permissive licence) · get_code("22e32871c286e462") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Neural compression is the application of neural networks and other machine learning methods to data compression. Recent advances in statistical machine learning have opened up new possibilities for data compression, allowing compression algorithms to be learned end-to-end from data using powerful generative models such as normalizing flows, variational autoencoders, diffusion probabilistic models, and generative adversarial networks. The present article aims to introduce this field of research to a broader machine learning audience by reviewing the necessary background in information theory (e.g., entropy coding, rate-distortion theory) and computer vision (e.g., image quality assessment, perceptual metrics), and providing a curated guide through the essential ideas and methods in the literature thus far.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2202.06533")
get_code_for_paper("2202.06533")
have("2202.06533")
Connect an agent — have() is free.