Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Denoising Diffusion Probabilistic Models
Authors: Jonathan Ho, Ajay Jain, Pieter Abbeel
NeurIPS 2020 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | On the unconditional CIFAR10 dataset, we obtain an Inception score of 9.46 and a state-of-the-art FID score of 3.17. On 256x256 LSUN, we obtain sample quality similar to Progressive GAN. |
| Researcher Affiliation | Academia | Jonathan Ho UC Berkeley EMAIL Ajay Jain UC Berkeley EMAIL Pieter Abbeel UC Berkeley EMAIL |
| Pseudocode | Yes | Algorithm 1 Training and Algorithm 2 Sampling are presented in the paper. |
| Open Source Code | Yes | Our implementation is available at https://github.com/hojonathanho/diffusion. |
| Open Datasets | Yes | On the unconditional CIFAR10 dataset, we obtain an Inception score of 9.46 and a state-of-the-art FID score of 3.17. On 256x256 LSUN, we obtain sample quality similar to Progressive GAN. |
| Dataset Splits | No | The paper mentions 'train' and 'test' in Table 1 (NLL Test (Train)) and 'training set' and 'test set' in Section 4.1 but does not provide specific split percentages, sample counts, or a detailed splitting methodology for reproducibility. |
| Hardware Specification | No | Google s Tensor Flow Research Cloud (TFRC) provided Cloud TPUs. This mentions the type of hardware (Cloud TPUs) but lacks specific details such as the model or version of the TPUs, or any CPU/GPU specs or memory amounts. |
| Software Dependencies | No | The paper mentions software components and methods like 'U-Net backbone', 'Pixel CNN++', 'group normalization', and 'Adam', but it does not provide specific version numbers for any libraries, frameworks, or solvers used. |
| Experiment Setup | Yes | We set T = 1000 for all experiments so that the number of neural network evaluations needed during sampling matches previous work [53, 55]. We set the forward process variances to constants increasing linearly from β1 = 10 4 to βT = 0.02. Algorithm 1 displays the complete training procedure with this simplified objective. |