Spaces:
Sleeping
Sleeping
Commit
·
2a25399
1
Parent(s):
b68eb49
Show iteration count and time used
Browse files- app.py +2 -0
- chill.py +6 -1
- gradio_cached_examples/14/log.csv +18 -12
app.py
CHANGED
|
@@ -67,6 +67,8 @@ def chill_out(text):
|
|
| 67 |
<li>Faithfulness score: {result['faithfulness_score']:.0%}</li>
|
| 68 |
<li>Spicy score: {result['spicy_score']:.0%}</li>
|
| 69 |
<li>Overall score: {result['overall_score']:.0%}</li>
|
|
|
|
|
|
|
| 70 |
<li>LLM requests made: {result['request_count']}</li>
|
| 71 |
</ul>
|
| 72 |
</div>
|
|
|
|
| 67 |
<li>Faithfulness score: {result['faithfulness_score']:.0%}</li>
|
| 68 |
<li>Spicy score: {result['spicy_score']:.0%}</li>
|
| 69 |
<li>Overall score: {result['overall_score']:.0%}</li>
|
| 70 |
+
<li>Iterations: {result['iteration_count'] / result['max_allowed_iterations']:.0%} {result['iteration_count']} of max allowed iterations: {result['max_allowed_iterations']}</li>
|
| 71 |
+
<li>Time used: {result['time_used']:.2f} seconds</li>
|
| 72 |
<li>LLM requests made: {result['request_count']}</li>
|
| 73 |
</ul>
|
| 74 |
</div>
|
chill.py
CHANGED
|
@@ -42,7 +42,7 @@ global suggestions
|
|
| 42 |
suggestions = []
|
| 43 |
last_edit = ""
|
| 44 |
request_count = 0
|
| 45 |
-
start_time =
|
| 46 |
|
| 47 |
|
| 48 |
def improve_text_attempt():
|
|
@@ -125,6 +125,7 @@ def improvement_loop(input_text):
|
|
| 125 |
global suggestions
|
| 126 |
global request_count
|
| 127 |
global start_time
|
|
|
|
| 128 |
suggestions = []
|
| 129 |
last_edit = ""
|
| 130 |
request_count = 0
|
|
@@ -133,6 +134,7 @@ def improvement_loop(input_text):
|
|
| 133 |
original_text = input_text
|
| 134 |
|
| 135 |
for iteration in range(1, max_iterations + 1):
|
|
|
|
| 136 |
try:
|
| 137 |
if iteration % 2 == 1:
|
| 138 |
last_edit = improve_text_attempt()
|
|
@@ -154,6 +156,9 @@ def improvement_loop(input_text):
|
|
| 154 |
print("ValueError:", e)
|
| 155 |
continue
|
| 156 |
assert len(suggestions) > 0
|
|
|
|
|
|
|
|
|
|
| 157 |
return suggestions[0]
|
| 158 |
|
| 159 |
|
|
|
|
| 42 |
suggestions = []
|
| 43 |
last_edit = ""
|
| 44 |
request_count = 0
|
| 45 |
+
start_time = None
|
| 46 |
|
| 47 |
|
| 48 |
def improve_text_attempt():
|
|
|
|
| 125 |
global suggestions
|
| 126 |
global request_count
|
| 127 |
global start_time
|
| 128 |
+
iteration_count = 0
|
| 129 |
suggestions = []
|
| 130 |
last_edit = ""
|
| 131 |
request_count = 0
|
|
|
|
| 134 |
original_text = input_text
|
| 135 |
|
| 136 |
for iteration in range(1, max_iterations + 1):
|
| 137 |
+
iteration_count = iteration
|
| 138 |
try:
|
| 139 |
if iteration % 2 == 1:
|
| 140 |
last_edit = improve_text_attempt()
|
|
|
|
| 156 |
print("ValueError:", e)
|
| 157 |
continue
|
| 158 |
assert len(suggestions) > 0
|
| 159 |
+
suggestions[0]["iteration_count"] = iteration_count
|
| 160 |
+
suggestions[0]["max_allowed_iterations"] = max_iterations
|
| 161 |
+
suggestions[0]["time_used"] = time_used
|
| 162 |
return suggestions[0]
|
| 163 |
|
| 164 |
|
gradio_cached_examples/14/log.csv
CHANGED
|
@@ -2,26 +2,32 @@ output,flag,username,timestamp
|
|
| 2 |
"
|
| 3 |
<div>
|
| 4 |
<h4>Edited text:</h4>
|
| 5 |
-
<p>
|
| 6 |
<h4>Details:</h4>
|
| 7 |
<ul>
|
| 8 |
-
<li>Critique:
|
| 9 |
-
<li>Faithfulness
|
| 10 |
-
<li>Spicy
|
| 11 |
-
<li>Overall
|
|
|
|
|
|
|
|
|
|
| 12 |
</ul>
|
| 13 |
</div>
|
| 14 |
-
",,,2024-02-29
|
| 15 |
"
|
| 16 |
<div>
|
| 17 |
<h4>Edited text:</h4>
|
| 18 |
-
<p>
|
| 19 |
<h4>Details:</h4>
|
| 20 |
<ul>
|
| 21 |
-
<li>Critique: This critique is
|
| 22 |
-
<li>Faithfulness
|
| 23 |
-
<li>Spicy
|
| 24 |
-
<li>Overall
|
|
|
|
|
|
|
|
|
|
| 25 |
</ul>
|
| 26 |
</div>
|
| 27 |
-
",,,2024-02-29
|
|
|
|
| 2 |
"
|
| 3 |
<div>
|
| 4 |
<h4>Edited text:</h4>
|
| 5 |
+
<p>It seems we're moving a bit slower than anticipated. I'm concerned we might not meet our shipping deadline.</p>
|
| 6 |
<h4>Details:</h4>
|
| 7 |
<ul>
|
| 8 |
+
<li>Critique: This is well-phrased and less inflammatory than the original. However, it might be more direct and still polite to say: 'We're behind schedule. I'm concerned we might not make our shipping deadline.'</li>
|
| 9 |
+
<li>Faithfulness score: 95%</li>
|
| 10 |
+
<li>Spicy score: 10%</li>
|
| 11 |
+
<li>Overall score: 97%</li>
|
| 12 |
+
<li>Iterations: 33% 2 of max allowed iterations: 6</li>
|
| 13 |
+
<li>Time used: 2.24 seconds</li>
|
| 14 |
+
<li>LLM requests made: 4</li>
|
| 15 |
</ul>
|
| 16 |
</div>
|
| 17 |
+
",,,2024-02-29 17:35:32.665693
|
| 18 |
"
|
| 19 |
<div>
|
| 20 |
<h4>Edited text:</h4>
|
| 21 |
+
<p>It seems your understanding of a balanced diet differs from mine, as it appears to include biscuits quite often.</p>
|
| 22 |
<h4>Details:</h4>
|
| 23 |
<ul>
|
| 24 |
+
<li>Critique: This critique is more subtle and less confrontational, which is good for preventing inflaming discussions. However, it could be more concise and closer to the original semantic intent. A possible revision could be: 'Your concept of a balanced diet appears to involve biscuits more often than usual.'</li>
|
| 25 |
+
<li>Faithfulness score: 95%</li>
|
| 26 |
+
<li>Spicy score: 10%</li>
|
| 27 |
+
<li>Overall score: 97%</li>
|
| 28 |
+
<li>Iterations: 33% 2 of max allowed iterations: 6</li>
|
| 29 |
+
<li>Time used: 2.36 seconds</li>
|
| 30 |
+
<li>LLM requests made: 4</li>
|
| 31 |
</ul>
|
| 32 |
</div>
|
| 33 |
+
",,,2024-02-29 17:35:35.029163
|