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 farm version #936

Merged
merged 7 commits into from
Apr 1, 2021
Merged

Update farm version #936

Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
d51ff39
Update farm version
Timoeller Mar 31, 2021
3d26984
Add new DPR loading, fix dpr param name
Timoeller Apr 1, 2021
a4f4014
Add latest docstring and tutorial changes
github-actions[bot] Apr 1, 2021
4eb20ea
Fix param in test
Timoeller Apr 1, 2021
790f7bf
Merge branch 'farm_v0.7.1' of github.com:deepset-ai/haystack into far…
Timoeller Apr 1, 2021
09262a1
Add QA model confidence as answer probability, fix prams in test
Timoeller Apr 1, 2021
e1560df
Fix dpr loading single model
Timoeller Apr 1, 2021
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
Prev Previous commit
Fix dpr loading single model
  • Loading branch information
@Timoeller
Timoeller committed Apr 1, 2021
commit e1560dfcb9128f9a861600e1bce021daeb6422c4
22 changes: 12 additions & 10 deletions haystack/retriever/dense.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,23 +154,25 @@ def __init__(self,
embed_title=embed_title,
num_hard_negatives=0,
num_positives=1)
prediction_head = TextSimilarityHead(similarity_function=similarity_function)
self.model = BiAdaptiveModel(
language_model1=self.query_encoder,
language_model2=self.passage_encoder,
prediction_heads=[prediction_head],
embeds_dropout_prob=0.1,
lm1_output_types=["per_sequence"],
lm2_output_types=["per_sequence"],
device=self.device,
)
else:
self.processor = TextSimilarityProcessor.load_from_dir(single_model_path)
self.processor.max_seq_len_passage = max_seq_len_passage
self.processor.max_seq_len_query = max_seq_len_query
self.processor.embed_title = embed_title
self.processor.num_hard_negatives = 0
self.processor.num_positives = 1 # during indexing of documents only one embedding is created
self.model = BiAdaptiveModel.load(single_model_path, device=self.device)

prediction_head = TextSimilarityHead(similarity_function=similarity_function)
self.model = BiAdaptiveModel(
language_model1=self.query_encoder,
language_model2=self.passage_encoder,
prediction_heads=[prediction_head],
embeds_dropout_prob=0.1,
lm1_output_types=["per_sequence"],
lm2_output_types=["per_sequence"],
device=self.device,
)
self.model.connect_heads_with_processor(self.processor.tasks, require_labels=False)

def retrieve(self, query: str, filters: dict = None, top_k: Optional[int] = None, index: str = None) -> List[Document]:
Expand Down