add the mean and fix the output

Dockerfile

@@ -2,7 +2,7 @@ FROM bentoml/model-server:0.11.0-py38
 MAINTAINER ersilia
 
 RUN pip install rdkit
-#RUN pip install numpy==1.23.1
+RUN pip install numpy==1.23.1
 RUN pip install fastai==1.0.61
 RUN pip install scikit-learn
 RUN pip install spacy

metadata.json

@@ -10,7 +10,7 @@
     "Input Shape": "Single",
     "Output": ["Other value"],
     "Output Type": ["Float"],
-    "Output Shape": "Flexible List",
+    "Output Shape": "Matrix",
     "Interpretation": "Embedding vectors of each smiles are obtained, represented in a matrix, where each row is a vector of embedding of each smiles character, with a dimension of 400. The pretrained model is loaded using the fastai library",
     "Tag": [
         "Descriptor",

model/framework/code/main.py

@@ -69,7 +69,8 @@ def my_model(smiles_list):
     for smile in smiles_list:
         smile_tokenizer=tokenizer.tokenizer(smile)
         indices = [qsar_db.vocab.itos.index(token) for token in smile_tokenizer]
-        embes=embs_v1[indices][:, :100].numpy()
+        embes=embs_v1[indices][:, :].numpy()
+        embes = np.mean(embes, axis=0)
         list_embeddings.append(embes)
 
     return list_embeddings
@@ -91,8 +92,8 @@ def my_model(smiles_list):
 # write output in a .csv file
 with open(output_file, "w") as f:
     writer = csv.writer(f)
-    writer.writerow(["embeddings"])  # header
+    writer.writerow(["emb-{0}".format(i) for i in range(400)])  # header
     for o in outputs:
-        writer.writerow([o])
+        writer.writerow(list(o))