Abdur Raheem Ali

icon indicating a website link abdur-raheem.com

icon company Independent (formerly Microsoft) icon location London, UK icon location counterfactual oracle predicting subspaces of the residual stream

Results 1 issues of Abdur Raheem Ali

[Bug Report] convert_llama_weights fails if I already quantized the weights to 4 bits

2 comment

https://github.com/neelnanda-io/TransformerLens/blob/ce82675a8e89b6d5e6229a89620c843c794f3b04/transformer_lens/loading_from_pretrained.py#L1395C7-L1395C69 runpod/pytorch:2.1.0-py3.10-cuda11.8.0-devel-ubuntu22.04 !pip install torch einops transformer_lens plotly circuitsvis numpy transformers==4.30.0 sentencepiece scikit-learn jaxtyping pytest bitsandbytes accelerate import gc import os import torch import numpy as np import einops import...