Add usage (#2)
Browse files- Add usage (3bc41d6cd3ce19a24f080d497dc2bbb6cc109ef4)
README.md
CHANGED
@@ -41,7 +41,50 @@ We evaluate the benefits of pretraining DNA FM 7B by conducting a comprehensive
|
|
41 |
TODO (@Caleb), we will need to see what results we want to put here.
|
42 |
|
43 |
## How to Use
|
44 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
45 |
|
46 |
|
47 |
## Citation
|
|
|
41 |
TODO (@Caleb), we will need to see what results we want to put here.
|
42 |
|
43 |
## How to Use
|
44 |
+
### Build any downstream models from this backbone
|
45 |
+
#### Embedding
|
46 |
+
```python
|
47 |
+
from genbio_finetune.tasks import Embed
|
48 |
+
model = Embed.from_config({"model.backbone": "dnafm"})
|
49 |
+
collated_batch = model.collate({"sequences": ["ACGT", "ACGT"]})
|
50 |
+
embedding = model(collated_batch)
|
51 |
+
print(embedding.shape)
|
52 |
+
print(embedding)
|
53 |
+
```
|
54 |
+
#### Sequence Level Classification
|
55 |
+
```python
|
56 |
+
import torch
|
57 |
+
from genbio_finetune.tasks import SequenceClassification
|
58 |
+
model = SequenceClassification.from_config({"model.backbone": "dnafm", "model.n_classes": 2})
|
59 |
+
collated_batch = model.collate({"sequences": ["ACGT", "ACGT"]})
|
60 |
+
logits = model(collated_batch)
|
61 |
+
print(logits)
|
62 |
+
print(torch.argmax(logits, dim=-1))
|
63 |
+
```
|
64 |
+
#### Token Level Classification
|
65 |
+
```python
|
66 |
+
import torch
|
67 |
+
from genbio_finetune.tasks import TokenClassification
|
68 |
+
model = TokenClassification.from_config({"model.backbone": "dnafm", "model.n_classes": 3})
|
69 |
+
collated_batch = model.collate({"sequences": ["ACGT", "ACGT"]})
|
70 |
+
logits = model(collated_batch)
|
71 |
+
print(logits)
|
72 |
+
print(torch.argmax(logits, dim=-1))
|
73 |
+
```
|
74 |
+
#### Regression
|
75 |
+
```python
|
76 |
+
from genbio_finetune.tasks import SequenceRegression
|
77 |
+
model = SequenceRegression.from_config({"model.backbone": "dnafm"})
|
78 |
+
collated_batch = model.collate({"sequences": ["ACGT", "ACGT"]})
|
79 |
+
logits = model(collated_batch)
|
80 |
+
print(logits)
|
81 |
+
```
|
82 |
+
#### Or use our one-liner CLI to finetune or evaluate any of the above!
|
83 |
+
```
|
84 |
+
gbft fit --model SequenceClassification --model.backbone dnafm --data SequenceClassification --data.path <hf_or_local_path_to_your_dataset>
|
85 |
+
gbft test --model SequenceClassification --model.backbone dnafm --data SequenceClassification --data.path <hf_or_local_path_to_your_dataset>
|
86 |
+
```
|
87 |
+
For more information, visit: [Model Generator](https://github.com/genbio-ai/test)
|
88 |
|
89 |
|
90 |
## Citation
|