hide sgpt for now

app.py

@@ -13,9 +13,9 @@ pinecone.init(
     environment=os.environ.get('PINECONE_ENV', '')
 )
 
-model = AutoModel.from_pretrained('monsoon-nlp/gpt-nyc')
-tokenizer = AutoTokenizer.from_pretrained('monsoon-nlp/gpt-nyc')
-zos = np.zeros(4096-1024).tolist()
+# model = AutoModel.from_pretrained('monsoon-nlp/gpt-nyc')
+# tokenizer = AutoTokenizer.from_pretrained('monsoon-nlp/gpt-nyc')
+# zos = np.zeros(4096-1024).tolist()
 
 def list_me(matches):
     result = ''
@@ -43,38 +43,39 @@ def query(question):
     )
 
     # SGPT search
-    batch_tokens = tokenizer(
-        [question],
-        padding=True,
-        truncation=True,
-        return_tensors="pt"
-    )
-    with torch.no_grad():
-        last_hidden_state = model(**batch_tokens, output_hidden_states=True, return_dict=True).last_hidden_state
-    weights = (
-        torch.arange(start=1, end=last_hidden_state.shape[1] + 1)
-        .unsqueeze(0)
-        .unsqueeze(-1)
-        .expand(last_hidden_state.size())
-        .float().to(last_hidden_state.device)
-    )
-    input_mask_expanded = (
-        batch_tokens["attention_mask"]
-        .unsqueeze(-1)
-        .expand(last_hidden_state.size())
-        .float()
-    )
-    sum_embeddings = torch.sum(last_hidden_state * input_mask_expanded * weights, dim=1)
-    sum_mask = torch.sum(input_mask_expanded * weights, dim=1)
-    embeddings = sum_embeddings / sum_mask
-    closest_sgpt = index.query(
-        top_k=2,
-        include_metadata=True,
-        namespace="mini",
-        vector=embeddings[0].tolist() + zos,
-    )
+    # batch_tokens = tokenizer(
+    #     [question],
+    #     padding=True,
+    #     truncation=True,
+    #     return_tensors="pt"
+    # )
+    # with torch.no_grad():
+    #     last_hidden_state = model(**batch_tokens, output_hidden_states=True, return_dict=True).last_hidden_state
+    # weights = (
+    #     torch.arange(start=1, end=last_hidden_state.shape[1] + 1)
+    #     .unsqueeze(0)
+    #     .unsqueeze(-1)
+    #     .expand(last_hidden_state.size())
+    #     .float().to(last_hidden_state.device)
+    # )
+    # input_mask_expanded = (
+    #     batch_tokens["attention_mask"]
+    #     .unsqueeze(-1)
+    #     .expand(last_hidden_state.size())
+    #     .float()
+    # )
+    # sum_embeddings = torch.sum(last_hidden_state * input_mask_expanded * weights, dim=1)
+    # sum_mask = torch.sum(input_mask_expanded * weights, dim=1)
+    # embeddings = sum_embeddings / sum_mask
+    # closest_sgpt = index.query(
+    #     top_k=2,
+    #     include_metadata=True,
+    #     namespace="mini",
+    #     vector=embeddings[0].tolist() + zos,
+    # )
 
-    return '<h3>Cohere</h3><ul>' + list_me(closest['matches']) + '</ul><h3>SGPT</h3><ul>' + list_me(closest_sgpt['matches']) + '</ul>'
+    return '<h3>Cohere</h3><ul>' + list_me(closest['matches']) + '</ul>'
+    #'<h3>SGPT</h3><ul>' + list_me(closest_sgpt['matches']) + '</ul>'
 
 
 iface = gr.Interface(