Search results
bitnet.cpp is the official inference framework for 1-bit LLMs (e.g., BitNet b1.58). It offers a suite of optimized kernels, that support fast and lossless inference of 1.58-bit models on CPU (with NPU and GPU support coming next).
README.md. bitnet.cpp is the official inference framework for 1-bit LLMs (e.g., BitNet b1.58). It offers a suite of optimized kernels, that support fast and lossless inference of 1.58-bit models on CPU (with NPU and GPU support coming next).
This is a reproduction of the BitNet b1.58 paper. The models are trained with RedPajama dataset for 100B tokens. The hypers, as well as two-stage LR and weight decay, are implemented as suggested in their following paper. All models are open-source in the repo.
18 wrz 2024 · BitNet is a special transformers architecture that represents each parameter with only three values: (-1, 0, 1), offering a extreme quantization of just 1.58 ( l o g 2 (3) log_2(3) l o g 2 (3)) bits per parameter. However, it requires to train a model from scratch.
29 mar 2024 · Here is the commands to run the evaluation: pip install lm-eval==0.3.0. python eval_ppl.py --hf_path 1bitLLM/bitnet_b1_58-3B --seqlen 2048. python eval_task.py --hf_path 1bitLLM/bitnet_b1_58-3B \ --batch_size 1 \ --tasks \ --output_path result.json \ --num_fewshot 0 \ --ctx_size 2048.
27 lut 2024 · Recent research, such as BitNet, is paving the way for a new era of 1-bit Large Language Models (LLMs). In this work, we introduce a 1-bit LLM variant, namely BitNet b1.58, in which every single parameter (or weight) of the LLM is ternary {-1, 0, 1}.
11 mar 2024 · When 2 numbers are quantized with the same bit count, the computational cost of floating point operations on the numbers reduces almost by the same factor of number of bits reduced (In Theory). This allows us the increase speed and reduce the ram consumption of ML models.
