andrewdalpino commited on
Commit
0122e21
·
verified ·
1 Parent(s): 273ecd0

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +9 -3
README.md CHANGED
@@ -19,6 +19,10 @@ An Evolutionary-scale Model (ESM) for protein function calling from amino acid s
19
 
20
  **Note**: This model specilizes on the `celluar component` subgraph of the gene ontology.
21
 
 
 
 
 
22
  ## Model Specs
23
 
24
  **Embedding Dimensions**: 640
@@ -44,11 +48,13 @@ model = EsmForSequenceClassification.from_pretrained(model_name)
44
 
45
  model.eval()
46
 
47
- sequnce = "MCNAWYISVDFEKNREDKSKCIHTRRNSGPKLLEHVMYEVLRDWYCLEGENVYMMGKKWQMPMCSLH"
 
 
48
 
49
  out = tokenizer(
50
  sequence,
51
- max_length=args.context_length,
52
  truncation=True,
53
  )
54
 
@@ -61,7 +67,7 @@ with torch.no_grad():
61
 
62
  probabilities = torch.sigmoid(outputs.logits.squeeze(0))
63
 
64
- probabilities, indices = torch.topk(probabilities, args.top_k)
65
 
66
  probabilities = probabilities.tolist()
67
 
 
19
 
20
  **Note**: This model specilizes on the `celluar component` subgraph of the gene ontology.
21
 
22
+ ## Code Repository
23
+
24
+ https://github.com/andrewdalpino/esm2-function-classifier
25
+
26
  ## Model Specs
27
 
28
  **Embedding Dimensions**: 640
 
48
 
49
  model.eval()
50
 
51
+ sequence = "MCNAWYISVDFEKNREDKSKCIHTRRNSGPKLLEHVMYEVLRDWYCLEGENVYMMGKKWQMPMCSLH"
52
+
53
+ top_k = 10
54
 
55
  out = tokenizer(
56
  sequence,
57
+ max_length=1026,
58
  truncation=True,
59
  )
60
 
 
67
 
68
  probabilities = torch.sigmoid(outputs.logits.squeeze(0))
69
 
70
+ probabilities, indices = torch.topk(probabilities, top_k)
71
 
72
  probabilities = probabilities.tolist()
73