Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -497,7 +497,7 @@ If any part of the information cannot be verified from this source, clearly stat
|
|
497 |
source_response = ""
|
498 |
for message in client.chat_completion(
|
499 |
messages=[{"role": "user", "content": prompt}],
|
500 |
-
max_tokens=
|
501 |
temperature=temperature,
|
502 |
stream=True,
|
503 |
):
|
@@ -510,6 +510,28 @@ If any part of the information cannot be verified from this source, clearly stat
|
|
510 |
# Generate an overall summary after processing all sources
|
511 |
overall_prompt = f"""Based on the summaries you've generated for each source: '{accumulated_response}', provide a concise overall summary that addresses the user's query: '{query}'
|
512 |
Highlight any conflicting information or gaps in the available data."""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
513 |
|
514 |
def get_response_from_pdf(query, model, selected_docs, num_calls=3, temperature=0.2):
|
515 |
logging.info(f"Entering get_response_from_pdf with query: {query}, model: {model}, selected_docs: {selected_docs}")
|
|
|
497 |
source_response = ""
|
498 |
for message in client.chat_completion(
|
499 |
messages=[{"role": "user", "content": prompt}],
|
500 |
+
max_tokens=10000,
|
501 |
temperature=temperature,
|
502 |
stream=True,
|
503 |
):
|
|
|
510 |
# Generate an overall summary after processing all sources
|
511 |
overall_prompt = f"""Based on the summaries you've generated for each source: '{accumulated_response}', provide a concise overall summary that addresses the user's query: '{query}'
|
512 |
Highlight any conflicting information or gaps in the available data."""
|
513 |
+
|
514 |
+
if model == "@cf/meta/llama-3.1-8b-instruct":
|
515 |
+
# Use Cloudflare API for overall summary
|
516 |
+
overall_response = ""
|
517 |
+
for response in get_response_from_cloudflare(prompt="", context="", query=overall_prompt, num_calls=1, temperature=temperature, search_type="web"):
|
518 |
+
overall_response += response
|
519 |
+
accumulated_response += f"Overall Summary:\n\n{overall_response}\n\n"
|
520 |
+
yield accumulated_response, ""
|
521 |
+
else:
|
522 |
+
# Use Hugging Face API for overall summary
|
523 |
+
overall_summary = ""
|
524 |
+
for message in client.chat_completion(
|
525 |
+
messages=[{"role": "user", "content": overall_prompt}],
|
526 |
+
max_tokens=10000,
|
527 |
+
temperature=temperature,
|
528 |
+
stream=True,
|
529 |
+
):
|
530 |
+
if message.choices and message.choices[0].delta and message.choices[0].delta.content:
|
531 |
+
chunk = message.choices[0].delta.content
|
532 |
+
overall_summary += chunk
|
533 |
+
accumulated_response += f"Overall Summary:\n\n{overall_summary}\n\n"
|
534 |
+
yield accumulated_response, ""
|
535 |
|
536 |
def get_response_from_pdf(query, model, selected_docs, num_calls=3, temperature=0.2):
|
537 |
logging.info(f"Entering get_response_from_pdf with query: {query}, model: {model}, selected_docs: {selected_docs}")
|