File size: 1,423 Bytes
87d8758 72efd6d 2d0bcfc 87d8758 de4850a 72efd6d de4850a 72efd6d eda2644 72efd6d eda2644 72efd6d eda2644 72efd6d eda2644 2d0bcfc |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 |
---
license: cc-by-nc-4.0
datasets:
- clane9/NSD-Flat
---
# Model card for `boldgpt_small_patch10.kmq`
A Vision Transformer (ViT) model trained on BOLD activation maps from [NSD-Flat](https://huggingface.co/datasets/clane9/NSD-Flat). Patches were quantized to discrete tokens using k-means (`KMeansTokenizer`). The training objective was to auto-regressively predict the next patch with shuffled patch order and cross-entropy loss.
## Dependencies
- [boldGPT](https://github.com/clane9/boldGPT)
## Usage
```python
from boldgpt.data import ActivityTransform
from boldgpt.models import create_model
from datasets import load_dataset
model = create_model("boldgpt_small_patch10.kmq", pretrained=True)
dataset = load_dataset("clane9/NSD-Flat", split="train")
dataset.set_format("torch")
transform = ActivityTransform()
batch = dataset[:1]
batch["activity"] = transform(batch["activity"])
# output: (B, N + 1, K) predicted next token logits
output, state = model(batch)
```
## Reproducing
- Training command:
```bash
torchrun --standalone --nproc_per_node=4 \
scripts/train_gpt.py --out_dir results \
--model boldgpt_small \
--ps 10 --vs 1024 --vocab_state checkpoints/ps-10_vs-1024_vss-4000_seed-42/tok_state.pt \
--shuffle --epochs 1000 --bs 512 \
--workers 0 --amp --compile --wandb
```
- Commit: `f9720ca52d6fa6b3eb47a34cf95f8e18a8683e4c`
|