Flow Matching MLP Model

Flow Matching model with MLP vector field

Model Details

Model Type: flow_matching_mlp
Training Epochs: 20000
Batch Size: 2000
Learning Rate: 0.001

Architecture

Hidden Layers: [100, 100, 100, 100]
Data Dimensionality: 2
Standard Deviation: 1.0
Scale: 10.0

Usage

from huggingface_hub import hf_hub_download
import torch
from model_config import FlowMatchingConfig

# Download model and config
model_path = hf_hub_download("derekwong/flow-matching-mlp", "pytorch_model.bin")
config_path = hf_hub_download("derekwong/flow-matching-mlp", "config.json")

# Load config
import json
with open(config_path) as f:
    config_dict = json.load(f)
config = FlowMatchingConfig.from_dict(config_dict)

# Load model (you'll need the MLPVectorField class)
model = MLPVectorField(dim=config.dim, hiddens=config.hiddens)
model.load_state_dict(torch.load(model_path))
model.eval()

Training

This model was trained using Flow Matching on 2D Gaussian mixture data.