### Install TRIBE v2 with Plotting

Source: https://github.com/facebookresearch/tribev2/blob/main/tribe_demo.ipynb

Installs the TRIBE v2 library, including plotting utilities, from its GitHub repository. Restart your environment after installation.

```bash
!uv pip install "tribev2[plotting] @ git+https://github.com/facebookresearch/tribev2.git"
```

--------------------------------

### Install TRIBE v2 with Training Dependencies

Source: https://github.com/facebookresearch/tribev2/blob/main/README.md

Install TRIBE v2 with all dependencies needed for training, including PyTorch Lightning and Weights & Biases.

```bash
pip install -e ".[training]"
```

--------------------------------

### Install TRIBE v2 (Inference Only)

Source: https://github.com/facebookresearch/tribev2/blob/main/README.md

Install TRIBE v2 with basic dependencies for inference purposes.

```bash
pip install -e .
```

--------------------------------

### Install TRIBE v2 with Brain Visualization

Source: https://github.com/facebookresearch/tribev2/blob/main/README.md

Install TRIBE v2 with additional dependencies required for brain visualization features.

```bash
pip install -e ".[plotting]"
```

--------------------------------

### Run Local Training Test

Source: https://github.com/facebookresearch/tribev2/blob/main/README.md

Execute a quick local test run of the training pipeline to verify setup and basic functionality.

```bash
python -m tribev2.grids.test_run
```

--------------------------------

### Prepare Text Input for Prediction

Source: https://github.com/facebookresearch/tribev2/blob/main/tribe_demo.ipynb

Converts text to speech using gTTS, then transcribes it to get word-level timings. This is necessary for predicting brain responses to text. The resulting events dataframe is used as input for the model.

```python
text = """
To be or not to be, that is the question.
Whether 'tis nobler in the mind to suffer
The slings and arrows of outrageous fortune,
Or to take arms against a sea of troubles
And by opposing end them. To die, to sleep,
No more; and by a sleep to say we end
The heartache and the thousand natural shocks
"""

text_path = CACHE_FOLDER / "shakespeare.txt"
text_path.write_text(text)

df = model.get_events_dataframe(text_path=text_path)
display(df.head(8)[["type", "start", "duration", "filepath", "text", "context"]])
```

--------------------------------

### Download Video and Create Events Dataframe

Source: https://github.com/facebookresearch/tribev2/blob/main/tribe_demo.ipynb

Downloads a sample video and processes it to create an events dataframe. This involves extracting audio, transcribing speech, and adding context to words. Ensure CACHE_FOLDER is defined.

```python
video_path = CACHE_FOLDER / "sample_video.mp4"
url = "https://download.blender.org/durian/trailer/sintel_trailer-480p.mp4"
download_file(url, video_path)
df = model.get_events_dataframe(video_path=video_path)
display(df.head(8)[["type", "start", "duration", "filepath", "text", "context"]])
```

--------------------------------

### Set Environment Variables for Training

Source: https://github.com/facebookresearch/tribev2/blob/main/README.md

Configure essential environment variables for data and output paths before running training scripts. Alternatively, these can be set directly in 'tribev2/grids/defaults.py'.

```bash
export DATAPATH="/path/to/studies"
export SAVEPATH="/path/to/output"
```

--------------------------------

### Run Grid Search on Slurm

Source: https://github.com/facebookresearch/tribev2/blob/main/README.md

Initiate a grid search for training models on a Slurm cluster. Use separate commands for cortical and subcortical training.

```bash
python -m tribev2.grids.run_cortical
python -m tribev2.grids.run_subcortical
```

--------------------------------

### Load TRIBE v2 Model and Plotter

Source: https://github.com/facebookresearch/tribev2/blob/main/tribe_demo.ipynb

Loads a pretrained TRIBE v2 model from HuggingFace Hub and initializes a PlotBrain object for 3D brain surface visualization. The model checkpoint is downloaded on the first run.

```python
from tribev2.demo_utils import TribeModel, download_file
from tribev2.plotting import PlotBrain
from pathlib import Path

CACHE_FOLDER = Path("./cache")

model = TribeModel.from_pretrained(
    "facebook/tribev2",
    cache_folder=CACHE_FOLDER,
)
plotter = PlotBrain(mesh="fsaverage5")
```

--------------------------------

### Load Pretrained Model and Predict Brain Responses

Source: https://github.com/facebookresearch/tribev2/blob/main/README.md

Load a pretrained TRIBE v2 model from HuggingFace and predict brain responses to a video. Ensure the 'cache_folder' is set to your desired directory. Predictions are for the 'average' subject and use the 'fsaverage5' cortical mesh.

```python
from tribev2 import TribeModel

model = TribeModel.from_pretrained("facebook/tribev2", cache_folder="./cache")

df = model.get_events_dataframe(video_path="path/to/video.mp4")
preds, segments = model.predict(events=df)
print(preds.shape)  # (n_timesteps, n_vertices)
```

--------------------------------

### Visualize Visual Stimulus Predictions

Source: https://github.com/facebookresearch/tribev2/blob/main/tribe_demo.ipynb

Plots predicted fMRI activity for visual stimuli on the fsaverage5 cortical mesh. Predictions are offset to compensate for hemodynamic lag. Use this to visualize model output for visual inputs.

```python
n_timesteps = 15
fig = plotter.plot_timesteps(preds[:n_timesteps], segments=segments[:n_timesteps], cmap="fire", norm_percentile=99, vmin=.6, alpha_cmap=(0, .2), show_stimuli=True)
```

--------------------------------

### Run TRIBE v2 Prediction Model

Source: https://github.com/facebookresearch/tribev2/blob/main/tribe_demo.ipynb

Feeds the events dataframe into the TRIBE v2 model to predict brain activity. This step requires access to the Llama-3.2 model. The output contains predictions and corresponding time segments.

```python
preds, segments = model.predict(events=df)
print(f"Predictions shape: {preds.shape}  (n_timesteps, n_vertices)")
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.