Spaces:
Build error
Build error
Commit History
-
checkpoint 2 - we are in a very good position right now, audio transcription is working, gradio page looks good, local llms are working, returning the consultation, the json that will be sent to the mcp client, and a debugging panel
5e4e457 -
improved seperation of concerns best practice in the code, added print statements for better understanding of what code is doing
9d2bec8 -
app.py wrapper for Gradio waa very bad idea, regoranizing project for clarity, utils folder will be very import of separation of concerns
3b5fe24 -
t-shooting poor performance
d364129
-
checkpoint 1
996be5a -
Revert "1. Structure responses with both diagnoses and follow-up questions"
This reverts commit f21d279f5305b3325d8c12144d6035d652b3f769.3a7b207 -
1. Structure responses with both diagnoses and follow-up questions
2. Maintain context between interactions 3. Use Mistral model (either local or API) when available 4. Format responses in a more interactive way5c7ca93 -
see a progress counter when submit is it
b348126 -
submit button wasn't working
3102246 -
these integrations could potentially multiply winnings
11e9f13 -
no need for gradio live
8a3861b -
solved ffmpeg-related error
4b8e08c -
1. Removed global pipeline initialization to prevent warm-up crashes
2. Added lazy loading of ASR pipeline 3. Fixed input processing with proper features and attention mask 4. Improved error handling 5. Proper audio preprocessing with correct sampling rate9ae574b -
FutureWarning: The input name `inputs` is deprecated. Please make sure to use `input_features` instead.
1f0d5ee -
preventing segmentation fault
7254b3e -
1. Added memory management
2. Limited input text length 3. Limited response length 4. Added batch size control 5. Added thread limiting75020f1 -
1. Added text input processing function
2. Added submit button event handler 3. Updated clear button to clear text input 4. Maintained existing voice functionality 5. Used same response formatting21d0c24 -
attention mask is not set
ec82b9a -
audio processing pipeliine
54fa492 -
1. Removed unsafe `.get()` calls
2. Added proper type checking 3. Better handling of different result formats 4. More robust error handling 5. Cleaner string handling00bcf43 -
taking out type
95321db -
enable cross-origin requests
65e09c8 -
should resolve font-related console errors
112a6dc -
microphone should now transcribe directly to text wihtout needing intermediate audio file processing
ede3b41 -
popped the bubble
5758d49 -
bug fix
e4c649c -
Add footer with social links
6099043 -
- Create the models directory on demand
- Fallback to tmp if permission issues - Keep models out of version control - Handle both local and HF Space environments - Prevent accidental model commitsf3e3012 -
fixed problems
fba5119 -
1. Added a professional header and instructions
2. Implemented user-friendly response formatting 3. Added advanced settings panel for API keys and model configuration 4. Improved chat interface with bubbles and better formatting 5. Added optional text-to-speech capability (commented out but ready to use) 6. Added clear chat button 7. Improved overall layout and themedaa3242 -
enhanced logging
a487eb3
-
should resolve the `ModuleNotFoundError` in the Hugging Face Space while maintaining the local development setup
f9e956a -
1. More structured query format focusing on ICD-10 codes
2. Better parsing of diagnoses from response 3. Weighted confidence scores 4. Specific error handling 5. Focus on clinical differentiationec7356b -
1. Modified the query prompt to explicitly ask for differentiating questions
2. Started with lower confidence (0.6) to reflect initial uncertainty 3. Used the query response to guide specific follow-up questions 4. Removed generic timing questions 5. Added structure to parse multiple potential diagnosesb31399c -
1. Removed hardcoded "cough" question
2. Added generic follow-up about symptom timing 3. Included transcript in error responses for debugging 4. Maintained JSON structure but made content dynamiccc35b42 -
better visibility into the startup process
8cb83ca -
added import
bb5eec7 -
1. Only download the specific `.gguf` file we need
2. Use less memory during download 3. Avoid the "Killed" error from system memory exhaustion6afe259 -
doesn't seem to be running
db0b943 -
setup for local llm
f6efb83 -
use llamacpp integration instead
984d858 -
no problems, local llm setup
255428d -
feat: Add local LLM support with ctransformers
c024962 -
works
932f0ad -
problems resolved
d6d864a -
the api has changed in recent version of gradio
2901e96 -
include MCP-specific configurations
d74007c -
interactive questioning enhancement
ed04336 -
getting local to work to prevent burining up my credits
0ef172c -
neaten
3f1fdcf