|
pulsatrix 1.1.0
|
Sampling from a masked diffusion language model (DIFF-2): start from a sequence with masked positions and fill them over a number of steps, recording the whole trajectory so it can be explained afterwards (DIFF-3). DPLM's remasking sampler (DIFF-4) also takes positions back. More...
#include <cstdint>#include <functional>#include <optional>#include <vector>#include "pulsatrix/causal_lm.hpp"#include "pulsatrix/encoder_lm.hpp"

Go to the source code of this file.
Classes | |
| struct | pulsatrix::MaskedDiffusionConfig |
| Sampling settings. The defaults fill greedily by confidence. More... | |
| struct | pulsatrix::MaskedDiffusionWrite |
| One position filled at one step. More... | |
| struct | pulsatrix::MaskedDiffusionStep |
| One step that filled at least one position (under Remasking, any step). More... | |
| struct | pulsatrix::MaskedDiffusionResult |
| A sampled sequence and how it was produced. More... | |
Namespaces | |
| namespace | pulsatrix |
Typedefs | |
| using | pulsatrix::DenoiserFn = std::function< std::vector< float >(const std::vector< int64_t > &tokens)> |
The model as the sampler sees it: the current sequence (masked positions included) in, logits for every position out, L * vocab values, row-major. MakeDenoiser builds one for an EncoderLM or a bidirectional CausalLM. | |
Enumerations | |
| enum class | pulsatrix::UnmaskingSchedule { pulsatrix::Confidence , pulsatrix::Ancestral , pulsatrix::Remasking } |
| How the sampler chooses which masked positions to fill at each step. More... | |
Functions | |
| MaskedDiffusionResult | pulsatrix::SampleMaskedDiffusion (const DenoiserFn &model, const std::vector< int64_t > &tokens, const MaskedDiffusionConfig &config) |
Fills every masked position of tokens over the configured steps. | |
| DenoiserFn | pulsatrix::MakeDenoiser (EncoderLM &model, DeviceBackend *backend) |
A DenoiserFn for an EncoderLM: one forward pass over the sequence, (1, L) to logits (1, L, vocab). The model isn't owned and must outlive the function. | |
| DenoiserFn | pulsatrix::MakeDenoiser (CausalLM &model, DeviceBackend *backend) |
A DenoiserFn for a CausalLM loaded with bidirectional attention, such as an A2D diffusion model (HfModelConfig::causal false). The model must outlive the function. | |
Sampling from a masked diffusion language model (DIFF-2): start from a sequence with masked positions and fill them over a number of steps, recording the whole trajectory so it can be explained afterwards (DIFF-3). DPLM's remasking sampler (DIFF-4) also takes positions back.