Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Update datagenerators to support tokenizing for LMs #91

Merged
merged 2 commits into from
Jan 24, 2024
Merged

Update datagenerators to support tokenizing for LMs #91

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
ee75da6
Update datagenerators to support tokenizing for LMs
atticusg Jan 24, 2024
c7b44b6
Bugfix for causal model
atticusg Jan 24, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
38 changes: 28 additions & 10 deletions pyvene/data_generators/causal_model.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -275,7 +275,7 @@ def check_path(total_setting):

return check_path

def inputToTensor(self, setting):
def input_to_tensor(self, setting):
result = []
for input in self.inputs:
temp = torch.tensor(setting[input]).float()
Expand All @@ -284,7 +284,7 @@ def inputToTensor(self, setting):
result.append(temp)
return torch.cat(result)

def outputToTensor(self, setting):
def output_to_tensor(self, setting):
result = []
for output in self.outputs:
temp = torch.tensor(float(setting[output]))
Expand All @@ -293,16 +293,28 @@ def outputToTensor(self, setting):
result.append(temp)
return torch.cat(result)

def generate_factual_dataset(self, size, sampler=None, filter=None, device="cpu"):
def generate_factual_dataset(
self,
size,
sampler=None,
filter=None,
device="cpu",
inputFunction=None,
outputFunction=None
):
if inputFunction is None:
inputFunction = self.input_to_tensor
if outputFunction is None:
outputFunction = self.output_to_tensor
if sampler is None:
sampler = self.sample_input
X, y = [], []
count = 0
while count < size:
input = sampler()
if filter is None or filter(input):
X.append(self.inputToTensor(input))
y.append(self.outputToTensor(self.run_forward(input)))
X.append(inputFunction(input))
y.append(outputFunction(self.run_forward(input)))
count += 1
return torch.stack(X).to(device), torch.stack(y).to(device)

Expand All @@ -315,6 +327,8 @@ def generate_counterfactual_dataset(
intervention_sampler=None,
filter=None,
device="cpu",
inputFunction=None,
outputFunction=None
):
maxlength = len(
[
Expand All @@ -323,6 +337,10 @@ def generate_counterfactual_dataset(
if var not in self.inputs and var not in self.outputs
]
)
if inputFunction is None:
inputFunction = self.input_to_tensor
if outputFunction is None:
outputFunction = self.output_to_tensor
if sampler is None:
sampler = self.sample_input
if intervention_sampler is None:
Expand All @@ -341,17 +359,17 @@ def generate_counterfactual_dataset(
if var not in intervention:
continue
source = sampler()
sources.append(self.inputToTensor(source))
sources.append(inputFunction(source))
source_dic[var] = source
for _ in range(maxlength - len(sources)):
sources.append(torch.zeros(self.inputToTensor(sampler()).shape))
example["labels"] = self.outputToTensor(
sources.append(torch.zeros(self.input_to_tensor(sampler()).shape))
example["labels"] = outputFunction(
self.run_interchange(base, source_dic)
).to(device)
example["base_labels"] = self.outputToTensor(
example["base_labels"] = outputFunction(
self.run_forward(base)
).to(device)
example["input_ids"] = self.inputToTensor(base).to(device)
example["input_ids"] = inputFunction(base).to(device)
example["source_input_ids"] = torch.stack(sources).to(device)
example["intervention_id"] = torch.tensor(
[intervention_id(intervention)]
Expand Down