sij/khoj
1
0
Fork 0
mirror of https://github.com/khoj-ai/khoj.git synced 2025-02-17 08:04:21 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Use torch.tensor as torch.Tensor cannot create tensor on MPS device

Browse source 
- `torch.Tensor' is apparently a legacy tensor constructor
- Using that to create tensor on MPS devices throws error:
  RuntimeError: legacy constructor expects device type: cpu but device type: mps was passed
- `torch.tensor' can handle creating tensors on Mac GPU (MPS) fine
This commit is contained in:
Debanjum Singh Solanky 2023-01-09 12:49:11 -03:00
parent 9def3f8c6f
commit eace7c6215
1 changed files with 1 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
src/search_type/text_search.py
View file

@ -69,7 +69,7 @@ def compute_embeddings(entries_with_ids: list[tuple[int, Entry]], bi_encoder, em
if new_entries:
new_embeddings = bi_encoder.encode(new_entries, convert_to_tensor=True, device=state.device, show_progress_bar=True)
existing_entry_ids = [id for id, _ in entries_with_ids if id != -1]
existing_embeddings = torch.index_select(corpus_embeddings, 0, torch.tensor(existing_entry_ids, device=state.device)) if existing_entry_ids else torch.Tensor(device=state.device)
existing_embeddings = torch.index_select(corpus_embeddings, 0, torch.tensor(existing_entry_ids, device=state.device)) if existing_entry_ids else torch.tensor([], device=state.device)
corpus_embeddings = torch.cat([existing_embeddings, new_embeddings], dim=0)
# Else compute the corpus embeddings from scratch
else: