fix tutorial visualize

This commit is contained in:
TommyWang 2022-11-21 03:21:59 +00:00
parent 62e31a69ff
commit 798312d72d
5 changed files with 7 additions and 6 deletions

View File

@ -34,7 +34,7 @@ Then use the config to add a delta model to the backbone model
delta_model = AutoDeltaModel.from_config(delta_config, backbone_model=backbone_model)
# now visualize the modified backbone_model
from opendelta import Visualization
from bigmodelvis import Visualization
Visualizaiton(backbone_model).structure_graph()
```

View File

@ -15,7 +15,7 @@ Delta tuning's core change in the structure of the base model is to decorate (mo
We should first know the name of the feedforward layer in the BART model by visualization. <img src="../imgs/hint-icon-2.jpg" height="30px"> *For more about visualization, see [Visualization](visualization).*
```python
from opendelta import Visualization
from bigmodelvis import Visualization
Visualization(model).structure_graph()
```

View File

@ -30,12 +30,12 @@ name: raw_print
The original presentation of models is **not tailored for repeated structures, big models, or parameters-centric tasks**.
## Using visualization from opendelta.
## Using visualization from bigmodelvis.
First let's visualize all the parameters in the bert model. As we can see, structure inside a bert model, and the all the paramters location of the model are neatly represented in tree structure. (See [color scheme](color_schema) for the colors)
```python
from opendelta import Visualization
from bigmodelvis import Visualization
model_vis = Visualization(backbone_model)
model_vis.structure_graph()
```

View File

@ -35,7 +35,7 @@ print(root.name_b[0].name_a)
We can visualize the model (For details, see [visualization](visualization))
```python
from opendelta import Visualization
from bigmodelvis import Visualization
Visualization(root).structure_graph()
```

View File

@ -28,6 +28,7 @@ from model_center.dataset import DistributedDataLoader
import opendelta as od
from opendelta import LoraModel, AdapterModel, CompacterModel, LowRankAdapterModel, BitFitModel, ParallelAdapterModel
from opendelta.utils.inspect import inspect_optimizer_statistics
from bigmodelvis import Visualization
print("before modify")
class BertModel(torch.nn.Module):
@ -56,7 +57,7 @@ def get_model(args):
"WiC" : 2,
}
model = BertModel(args, num_types[args.dataset_name])
# od.Visualization(model).structure_graph()
Visualization(model).structure_graph()
if args.delta_type == "lora":
delta_model = LoraModel(backbone_model=model, modified_modules=['project_q', 'project_k'], backend='bmt')