Felipe97/llama-cpp-compiled

1# Diffusion Text Generation2 3This directory contains implementations for Diffusion LLMs (DLLMs)4 5More Info:6- https://github.com/ggml-org/llama.cpp/pull/146447- https://github.com/ggml-org/llama.cpp/pull/147718 9## Parameters10The diffusion CLI supports various parameters to control the generation process:11 12### Core Diffusion Parameters13- `--diffusion-steps`: Number of diffusion steps (default: 256)14- `--diffusion-algorithm`: Algorithm for token selection15  - `0`: DIFFUSION_ALGORITHM_ORIGIN - Token will be generated in a purely random order from https://arxiv.org/abs/2107.03006.16  - `1`: DIFFUSION_ALGORITHM_ENTROPY_BASED - Entropy-based selection17  - `2`: DIFFUSION_ALGORITHM_MARGIN_BASED - Margin-based selection18  - `3`: DIFFUSION_ALGORITHM_RANDOM - Random selection19  - `4`: DIFFUSION_ALGORITHM_CONFIDENCE_BASED - Confidence-based selection (default)20  - More documentation here https://github.com/DreamLM/Dream21- `--diffusion-visual`: Enable live visualization during generation22 23### Scheduling Parameters24Choose one of the following scheduling methods:25 26**Timestep-based scheduling:**27- `--diffusion-eps`: Epsilon value for timestep scheduling (e.g., 0.001)28 29**Block-based scheduling:**30- `--diffusion-block-length`: Block size for block-based scheduling (e.g., 32)31 32### Sampling Parameters33- `--temp`: Temperature for sampling (0.0 = greedy/deterministic, higher = more random)34- `--top-k`: Top-k filtering for sampling35- `--top-p`: Top-p (nucleus) filtering for sampling36- `--seed`: Random seed for reproducibility37 38### Model Parameters39- `-m`: Path to the GGUF model file40- `-p`: Input prompt text41- `-ub`: Maximum sequence length (ubatch size)42- `-c`: Context size43- `-b`: Batch size44 45### Examples46#### Dream architecture:47```48llama-diffusion-cli -m dream7b.gguf -p "write code to train MNIST in pytorch" -ub 512 --diffusion-eps 0.001 --diffusion-algorithm 3 --diffusion-steps 256 --diffusion-visual49```50 51#### LLaDA architecture:52```53llama-diffusion-cli -m llada-8b.gguf -p "write code to train MNIST in pytorch" -ub 512 --diffusion-block-length 32 --diffusion-steps 256 --diffusion-visual54```55 56#### RND1 architecture:57```58llama-diffusion-cli -m RND1-Base-0910.gguf -p "write code to train MNIST in pytorch" -ub 512 --diffusion-algorithm 1 --diffusion-steps 256 --diffusion-visual --temp 0.5 --diffusion-eps 0.00159```60