a module to visualise attention layer activations from transformer based models from huggingface
Project description
attention-visualiser
a module to visualise attention layer activations from transformer based models from huggingface
installation
pip install git+https://codeberg.org/rashomon/attention-visualiser
usage
from attention_visualiser import AttentionVisualiser
from transformers import AutoModel, AutoTokenizer
# visualising activations from gpt
model_name = "openai-community/openai-gpt"
model = AutoModel.from_pretrained(model_name)
model.eval()
tokenizer = AutoTokenizer.from_pretrained(model_name)
text = "Look on my Works, ye Mighty, and despair!"
encoded_inputs = tokenizer.encode_plus(text, truncation=True, return_tensors="pt")
visualiser = AttentionVisualiser(model, tokenizer)
# visualise from the first attn layer
visualiser.visualise_attn_layer(0, encoded_inputs)
local dev
# env setup
uv sync
source .venv/bin/activate
# tests
uv run pytest
# tests with coverage
uv run pytest --cov --cov-report=xml
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
attention_visualiser-0.1.0.tar.gz
(110.8 kB
view details)
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file attention_visualiser-0.1.0.tar.gz.
File metadata
- Download URL: attention_visualiser-0.1.0.tar.gz
- Upload date:
- Size: 110.8 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: uv/0.9.10 {"installer":{"name":"uv","version":"0.9.10"},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
55c18ea775641df1bb771c0988c9bdb505ff98c66448ada172edf8e910c8ac9c
|
|
| MD5 |
917d94a031c6c58ec18423303bf8d578
|
|
| BLAKE2b-256 |
26d56fa55fdfefc2955ee3eaa41c5660b7fbc2e61b5c166e193bff4cf805f331
|
File details
Details for the file attention_visualiser-0.1.0-py3-none-any.whl.
File metadata
- Download URL: attention_visualiser-0.1.0-py3-none-any.whl
- Upload date:
- Size: 2.3 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: uv/0.9.10 {"installer":{"name":"uv","version":"0.9.10"},"python":null,"implementation":{"name":null,"version":null},"distro":{"name":"Ubuntu","version":"24.04","id":"noble","libc":null},"system":{"name":null,"release":null},"cpu":null,"openssl_version":null,"setuptools_version":null,"rustc_version":null,"ci":null}
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
b1aaaac3117e98e65f331eaf1631a7a1f10febf185066dcf9eb627dbf75f0bbd
|
|
| MD5 |
b26960e54c6cd612b744e69842150671
|
|
| BLAKE2b-256 |
81bb08c572022dc3e08433eb2e54f60834cc422918b95d363a46bd85dbb526e0
|
