File size: 3,139 Bytes
7ecf721
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
942af0b
7ecf721
942af0b
 
 
 
7ecf721
 
 
 
942af0b
7ecf721
 
 
 
 
942af0b
 
 
 
 
7ecf721
 
 
942af0b
 
 
 
 
7ecf721
942af0b
7ecf721
942af0b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7ecf721
942af0b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7ecf721
942af0b
 
 
 
 
 
 
 
7ecf721
 
 
 
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
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
---
license: mit
tags:
  - chemistry
  - quantum-chemistry
  - dft
  - hamiltonian
  - equivariant
  - flow-matching
datasets:
  - MD17
language:
  - en
pipeline_tag: other
---

# QHFlow2 — MD17 Pre-trained Checkpoints

Pre-trained checkpoints for **QHFlow2** on the **MD17** dataset (DFT Hamiltonian prediction).

> **Paper:** [High-order Equivariant Flow Matching for Density Functional Theory Hamiltonian Prediction](https://arxiv.org/abs/2602.16897)
> **Authors:** Seongsu Kim, Nayoung Kim, Dongwoo Kim, Sungsoo Ahn (KAIST)
> **Venue:** NeurIPS 2025
> **Code:** [github.com/seongsukim-ml/QHFlow2](https://github.com/seongsukim-ml/QHFlow2)

## Model Variants

| Size | hidden_size | num_gnn_layers | Checkpoint Size |
|------|-------------|----------------|-----------------|
| small_v2 | 64 | 3 | 313 MB |
| middle | 128 | 3 | 975 MB |

## Molecules

| Molecule | Atoms | Formula |
|----------|-------|---------|
| ethanol | 9 | C₂H₆O |
| malondialdehyde | 9 | C₃H₄O₂ |
| uracil | 12 | C₄H₄N₂O₂ |

## File Structure

```
{molecule}/
  QHFlow_so2_v5_1_{size}_b10-{molecule}/
    weights-epoch=79-val_loss=0.0000000.ckpt
```

## Quick Start

### 1. Install QHFlow2

```bash
git clone https://github.com/seongsukim-ml/QHFlow2.git
cd QHFlow2
pip install -e ".[fairchem]"
```

### 2. Download Checkpoints

```bash
# Download a single checkpoint
huggingface-cli download ksusu/QHFlow2-MD17 \
  "ethanol/QHFlow_so2_v5_1_small_v2_b10-ethanol/weights-epoch=79-val_loss=0.0000000.ckpt" \
  --local-dir ckpt/md17

# Download all checkpoints (~3.9 GB)
huggingface-cli download ksusu/QHFlow2-MD17 --local-dir ckpt/md17
```

Or in Python:

```python
from huggingface_hub import hf_hub_download

path = hf_hub_download(
    repo_id="ksusu/QHFlow2-MD17",
    filename="ethanol/QHFlow_so2_v5_1_small_v2_b10-ethanol/weights-epoch=79-val_loss=0.0000000.ckpt",
)
```

### 3. Download Dataset

Download from [Google Drive](https://drive.google.com/drive/folders/1d3HTu0H7gdg54kirWBqN24x-s1QW6OKV?usp=sharing) and place under `dataset/`:

```
QHFlow2/dataset/
  ethanol/
  malondialdehyde/
  uracil/
```

### 4. Run Prediction

```bash
python -m qhflow2.experiment.train_md17 \
  mode=predict \
  dataset=ethanol \
  ckpt=ckpt/md17/ethanol/QHFlow_so2_v5_1_small_v2_b10-ethanol/weights-epoch=79-val_loss=0.0000000.ckpt
```

### 5. Python API

```python
import torch
from qhflow2.models import get_model, get_default_model_args

args = get_default_model_args("md17")
args["version"] = "QHFlow_so2_v5_1"
args["hidden_size"] = 64
args["num_gnn_layers"] = 3
model = get_model(args)

ckpt = torch.load("weights-epoch=79-val_loss=0.0000000.ckpt", map_location="cpu")
state_dict = {k.replace("model.", ""): v for k, v in ckpt["state_dict"].items()}
model.load_state_dict(state_dict, strict=False)
model.eval()
```

## Citation

```bibtex
@inproceedings{kim2025high,
  title={High-order Equivariant Flow Matching for Density Functional Theory Hamiltonian Prediction},
  author={Kim, Seongsu and Kim, Nayoung and Kim, Dongwoo and Ahn, Sungsoo},
  booktitle={Advances in Neural Information Processing Systems},
  year={2025}
}
```

## License

MIT