Update app.py (#2)

- Update app.py (f20f8a73f1dfc56dfe1d3b6a56118e3fda0001e4)
- Add queue (f8a810646edd593f006aa61cc6f5915ea15c0928)
- Attempt adding seed + make optimization optional (f0e697bf3a2896eb938cfa08c0469cb16759892a)

app.py

@@ -185,9 +185,9 @@ def find_best_parameters(eval_data, model, tokenizer, max_length=85):
 
 
 
-def run_model(inputs, tokenizer, model, num_beams=2, num_beam_groups=2, temperature=0.5, num_return_sequences=1, max_length=85):
+def run_model(inputs, tokenizer, model, num_beams=2, num_beam_groups=2, temperature=0.5, num_return_sequences=1, max_length=85, seed=42069):
     all_outputs = []
-    torch.manual_seed(42069)
+    torch.manual_seed(seed)
     for input_text in inputs:
         model_inputs = tokenizer([input_text], max_length=512, padding=True, truncation=True)
         input_ids = torch.tensor(model_inputs['input_ids']).to(device)
@@ -232,7 +232,7 @@ def run_model(inputs, tokenizer, model, num_beams=2, num_beam_groups=2, temperat
 
 
 @spaces.GPU
-def gen(content, temperature_qg=0.5, temperature_qa=0.75, num_return_sequences_qg=1, num_return_sequences_qa=1, max_length=85):
+def gen(content, temperature_qg=0.5, temperature_qa=0.75, num_return_sequences_qg=1, num_return_sequences_qa=1, max_length=85, seed=42069, optimize_questions=False):
     inputs = [
         f'context: {content}'
     ]
@@ -244,21 +244,24 @@ def gen(content, temperature_qg=0.5, temperature_qa=0.75, num_return_sequences_q
         num_beam_groups=num_return_sequences_qg, 
         temperature=temperature_qg, 
         num_return_sequences=num_return_sequences_qg, 
-        max_length=max_length
+        max_length=max_length,
+        seed=seed
     )
 
-    q_params = find_best_parameters(list(chain.from_iterable(question)), qg_model, tokenizer, max_length=max_length)
-
-    question = run_model(
-        inputs, 
-        tokenizer, 
-        qg_model, 
-        num_beams=q_params[0],
-        num_beam_groups=q_params[1],
-        temperature=temperature_qg, 
-        num_return_sequences=num_return_sequences_qg, 
-        max_length=max_length
-    )
+    if optimize_questions:
+        q_params = find_best_parameters(list(chain.from_iterable(question)), qg_model, tokenizer, max_length=max_length)
+
+        question = run_model(
+            inputs, 
+            tokenizer, 
+            qg_model, 
+            num_beams=q_params[0],
+            num_beam_groups=q_params[1],
+            temperature=temperature_qg, 
+            num_return_sequences=num_return_sequences_qg, 
+            max_length=max_length,
+            seed=seed
+        )
 
     inputs = list(chain.from_iterable([
         [f'question: {q} context: {content}' for q in q_set] for q_set in question
@@ -271,7 +274,8 @@ def gen(content, temperature_qg=0.5, temperature_qa=0.75, num_return_sequences_q
         num_beam_groups=num_return_sequences_qa,
         temperature=temperature_qa,
         num_return_sequences=num_return_sequences_qa,
-        max_length=max_length
+        max_length=max_length,
+        seed=seed
     )
 
     questions = list(chain.from_iterable(question))
@@ -305,6 +309,30 @@ def create_file_download(qnas):
 
 
 with gr.Blocks(css='.hidden_input {display: none;}') as demo:
+    with gr.Row(equal_height=True):
+        gr.Markdown(
+        """
+        # QA-Generator
+        A combination of fine-tuned flan-T5(-small) models chained into sequence
+        to generate:
+        
+        A) a versatile set of questions
+        B) an accurate set of matching answers
+        
+        according to a given piece of text content.
+
+        The idea is simple:
+        
+        1. Add your content
+        2. Select the amount of questions you want to generate
+        2.2 (optional) Select the amount of answers you want to generate per goven question
+        3. Press generate
+        4. ???
+        5. Profit
+
+        If you're satisfied with the generated data set, you can export it as TSV
+        to edit or import it into your favourite tool.
+        """)
     with gr.Row(equal_height=True):
         with gr.Group("Content"):
             content = gr.Textbox(label='Content', lines=15, placeholder='Enter text here', max_lines=10_000)
@@ -314,6 +342,8 @@ with gr.Blocks(css='.hidden_input {display: none;}') as demo:
             max_length = gr.Number(label='Max Length', value=85, minimum=1, step=1, maximum=512)
             num_return_sequences_qg = gr.Number(label='Number Questions', value=max_questions, minimum=1, step=1, maximum=max(max_questions, max_elem_value))
             num_return_sequences_qa = gr.Number(label="Number Answers", value=max_answers, minimum=1, step=1, maximum=max(max_questions, max_elem_value))
+            seed = gr.Number(label="seed", value=42069)
+            optimize_questions = gr.Checkbox(label="Optimize questions?", value=False)
 
     with gr.Row():
         gen_btn = gr.Button("Generate")
@@ -321,14 +351,14 @@ with gr.Blocks(css='.hidden_input {display: none;}') as demo:
     @gr.render(
         inputs=[
             content, temperature_qg, temperature_qa, num_return_sequences_qg, num_return_sequences_qa,
-            max_length
+            max_length, seed, optimize_questions
         ],
         triggers=[gen_btn.click]
     )
     def render_results(content, temperature_qg, temperature_qa, num_return_sequences_qg, num_return_sequences_qa, max_length):
         qnas = gen(
             content, temperature_qg, temperature_qa, num_return_sequences_qg, num_return_sequences_qa,
-            max_length
+            max_length, seed, optimize_questions
         )
         df = gr.Dataframe(
             value=[u.values() for u in qnas],
@@ -342,4 +372,5 @@ with gr.Blocks(css='.hidden_input {display: none;}') as demo:
 
 
 
+demo.queue()
 demo.launch()